maint: update copyright dates

Problem reported by Dan Jacobsen (bug#79959).
* src/diff.c (option_help_msgid): Tweak.

.gitignore

@@ -1,94 +1,110 @@
 *.a
 *.exe
 *.o
+*.orig
+*.patch
+*.rej
 *~
-.deps
+.Tpo
+.deps/
 .dirstamp
 .gdb-history
+/*.diff
 /ABOUT-NLS
 /GNUmakefile
 /INSTALL
 /README-release
 /aclocal.m4
-/build-aux
+/autom4te.cache/
+/build-aux/
+/confdefs*
 /config.cache
 /config.log
 /config.status
 /configure
+/conftest*
 /diffutils-*.tar.xz
-/doc/.gitignore
-/doc/diffutils.aux
-/doc/diffutils.cp
-/doc/diffutils.cps
-/doc/diffutils.info
-/doc/diffutils.log
-/doc/diffutils.toc
-/doc/stamp-vti
-/doc/version.texi
-/gnulib-tests
-/gnulib-tests/.gitignore
-/gnulib-tests/locale/
-/gnulib-tests/uniwidth/
-/lib/.gitignore
-/lib/alloca.h
-/lib/c++defs.h
-/lib/charset.alias
-/lib/config.h
-/lib/config.hin
-/lib/configmake.h
-/lib/fcntl.h
-/lib/iconv*.h
-/lib/inttypes.h
-/lib/langinfo.h
-/lib/locale.h
-/lib/ref-add.sed
-/lib/ref-del.sed
-/lib/signal.h
-/lib/stamp-h1
-/lib/stdio.h
-/lib/stdlib.h
-/lib/string.h
-/lib/strings.h
-/lib/sys/
-/lib/time.h
-/lib/unistd.h
-/lib/unistr
-/lib/unistr.h
-/lib/unitypes.h
-/lib/uniwidth
-/lib/uniwidth.h
-/lib/unused-parameter.h
-/lib/warn-on-use.h
-/lib/wchar.h
-/lib/wctype.h
-/m4/.gitignore
-/m4/gnulib-cache.m4
+/gnulib-tests/
+/m4/
 /maint.mk
-/po/*.pot
-/po/*.sed
-/po/.gitignore
-/po/.reference
-/po/LINGUAS
-/po/Makefile.in.in
-/po/Makevars
-/po/Makevars.template
-/po/POTFILES
-/po/Rules-quot
-/po/stamp-po
-ABOUT-NLS
 ChangeLog
 ID
 Makefile
 Makefile.in
 TAGS
-autom4te.cache
-build-aux
-gnulib-tests/*.log
+doc/.gitignore
+doc/diffutils.aux
+doc/diffutils.cp
+doc/diffutils.cps
+doc/diffutils.info
+doc/diffutils.log
+doc/diffutils.toc
+doc/stamp-vti
+doc/version.texi
+lib/*/
+lib/.gitignore
+lib/alloca.h
 lib/arg-nonnull.h
-lib/link-warning.h
+lib/c++defs.h
+lib/c-file-type.c
+lib/config.h
+lib/config.hin
+lib/ctype.h
+lib/dirent.h
+lib/errno.h
+lib/error.h
+lib/fcntl.h
+lib/file-type.h
+lib/fnmatch.h
+lib/getopt-cdefs.h
+lib/getopt.h
+lib/iconv*.h
+lib/inttypes.h
+lib/langinfo.h
+lib/limits.h
+lib/locale.h
+lib/signal.h
+lib/sigsegv.h
+lib/stamp-h1
+lib/stdarg.h
+lib/stdbit.h
+lib/stdckdint.h
+lib/stddef.h
+lib/stdint.h
+lib/stdio.h
+lib/stdlib.h
+# For some reason 'bootstrap' doesn't add /stdopen.[ch] to lib/.gitignore.
+lib/stdopen.[ch]
+lib/string.h
+lib/strings.h
+lib/time.h
+lib/uchar.h
+lib/unicase.h
+lib/unictype.h
+lib/uninorm.h
+lib/unistd.h
+lib/unistr.h
+lib/unistr/
+lib/unitypes.h
+lib/uniwidth.h
+lib/uniwidth/
+lib/warn-on-use.h
+lib/wchar.h
+lib/wctype.h
 man/*.1
 po/*.gmo
 po/*.po
+po/*.pot
+po/*.sed
+po/.gitignore
+po/.reference
+po/LINGUAS
+po/Makefile.in.in
+po/Makevars
+po/Makevars.template
+po/POTFILES
+po/Rules-quot
+po/stamp-po
 src/cmp
 src/diff
 src/diff3
@@ -97,3 +113,5 @@ src/sdiff
 src/version.[ch]
 tests/*.log
 tests/*.trs
+/tests/init.sh
+vc-dwim-log-*

.gitmodules

@@ -1,3 +1,3 @@
 [submodule "gnulib"]
 	path = gnulib
-	url = git://git.sv.gnu.org/gnulib.git
+	url = https://git.savannah.gnu.org/git/gnulib

.prev-version

@@ -1 +1 @@
-3.5
+3.12

AUTHORS

@@ -33,7 +33,7 @@ Patrick D'Cruze
 Eli Zaretskii
 
 
-Copyright (C) 2001, 2006, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+Copyright (C) 2001, 2006, 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
 This file is part of GNU diffutils.
 

ChangeLog-2008

@@ -4265,7 +4265,7 @@ Thu Nov  3 16:30:24 1988  Randall Smith  (randy at gluteus.ai.mit.edu)
 
 	-----
 
-	Copyright (C) 1988-1994, 1997-2002, 2004, 2006, 2009-2013, 2015-2017
+	Copyright (C) 1988-1994, 1997-2002, 2004, 2006, 2009-2013, 2015-2026
 	Free Software Foundation, Inc.
 
 	Copying and distribution of this file, with or without

HACKING

@@ -3,14 +3,6 @@ Diffutils Contribution Guidelines
 
 Prerequisites
 =============
-You will need the "git" version control tools.
-On Fedora-based systems, do "yum install git".
-On Debian-based ones install the "git-core" package.
-Then run "git --version".  If that says it's older than
-version 1.4.4, then you'd do well to get a newer version.
-At worst, just download the latest stable release from
-http://git.or.cz/ and build from source.
-
 For details on building the programs in this package, see
 the file, README-hacking.
 
@@ -20,14 +12,11 @@ Use the latest upstream sources
 Base any changes you make on the latest upstream sources.
 You can get a copy of the latest with this command:
 
-    git clone git://git.sv.gnu.org/diffutils
+    git clone https://git.savannah.gnu.org/git/diffutils
 
-That downloads the entire repository, including revision control history
-dating back to 1991.  The repository (the part you download, and which
-resides in diffutils/.git) currently weighs in at about 9MB.  So you don't
-want to download it more often than necessary.  Once downloaded, you
-can get incremental updates by running one of these commands from
-inside your new diffutils/ directory:
+That downloads the entire repository, including revision control history.
+Once downloaded, you can get incremental updates by running one of
+these commands from inside your new diffutils/ directory:
 
 If you have made *no* changes:
     git pull
@@ -94,7 +83,7 @@ Make your changes on a private "topic" branch
 =============================================
 So you checked out diffutils like this:
 
-  git clone git://git.sv.gnu.org/diffutils
+  git clone https://git.savannah.gnu.org/git/diffutils
 
 Now, cd into the diffutils/ directory and run:
 
@@ -221,7 +210,7 @@ keep the maximum line length at 72 or smaller, so that the generated
 ChangeLog lines, each with its leading TAB, will not exceed 80 columns.
 As for the ChangeLog-style content, please follow these guidelines:
 
-  http://www.gnu.org/software/guile/changelogs/guile-changelogs_3.html
+  https://www.gnu.org/prep/standards/html_node/Change-Logs.html
 
 Try to make the summary line fit one of the following forms:
 
@@ -438,18 +427,18 @@ The forms to choose from are in gnulib's doc/Copyright/ directory.
 If you want to assign a single change, you should use the file,
 doc/Copyright/request-assign.changes:
 
-    http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=doc/Copyright/request-assign.changes;hb=HEAD
+    https://www.gnu.org/software/gnulib/Copyright/request-assign.changes
 
 If you would like to assign past and future contributions to a project,
 you'd use doc/Copyright/request-assign.future:
 
-    http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=doc/Copyright/request-assign.future;hb=HEAD
+    https://www.gnu.org/software/gnulib/Copyright/request-assign.future
 
 You may make assignments for up to four projects at a time.
 
 In case you're wondering why we bother with all of this, read this:
 
-    http://www.gnu.org/licenses/why-assign.html
+    https://www.gnu.org/licenses/why-assign.html
 
 
 Run "make syntax-check", or even "make distcheck"
@@ -483,13 +472,11 @@ Do not add any more trailing blanks anywhere.  While "make syntax-check"
 will alert you if you slip up, it's better to nip any problem in the
 bud, as you're typing.  A good way to help you adapt to this rule is
 to configure your editor to highlight any offending characters in the
-files you edit.  If you use Emacs, customize its font-lock mode (FIXME:
-provide more detail) or try one of its whitespace packages.  This appears
-to be the one that will end up in emacs 23:
+files you edit.  If you use Emacs, customize its font-lock mode
+or use its WhiteSpace mode:
 
-    http://www.emacswiki.org/emacs/WhiteSpace
+    https://www.emacswiki.org/emacs/WhiteSpace
 
-[that page says its version also works with emacs 21 and 22]
 If you use vim, add this to ~/.vimrc:
 
     let c_space_errors=1
@@ -581,7 +568,7 @@ Then just open the index.html file (in the generated lcov-html directory)
 in your favorite web browser.
 
 ========================================================================
-Copyright (C) 2009-2013, 2015-2017 Free Software Foundation, Inc.
+Copyright (C) 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
 Permission is granted to copy, distribute and/or modify this document
 under the terms of the GNU Free Documentation License, Version 1.3 or

Makefile.am

@@ -1,6 +1,6 @@
 # Main Automakefile for GNU diffutils.
 
-# Copyright (C) 2001-2002, 2004, 2006, 2009-2013, 2015-2017 Free Software
+# Copyright (C) 2001-2002, 2004, 2006, 2009-2013, 2015-2026 Free Software
 # Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify

NEWS

@@ -1,5 +1,149 @@
 GNU diffutils NEWS                                    -*- outline -*-
 
+* Noteworthy changes in release ?.? (????-??-??) [?]
+
+
+* Noteworthy changes in release 3.12 (2025-04-08) [stable]
+
+** Bug fixes
+
+  diff -r no longer merely summarizes when comparing an empty regular
+  file to a nonempty regular file.
+  [bug#76452 introduced in 3.11]
+
+  diff -y no longer crashes when given nontrivial differences.
+  [bug#76613 introduced in 3.11]
+
+
+* Noteworthy changes in release 3.11 (2025-02-02) [stable]
+
+** Improvements
+
+  Programs now quote file names more consistently in diagnostics.
+  For example; "cmp 'none of' /etc/passwd" now might output
+  "cmp: EOF on ‘none of’ which is empty" instead of outputting
+  "cmp: EOF on none of which is empty".  In diagnostic messages
+  that traditionally omit quotes and where backward compatibility
+  seems to be important, programs continue to omit quotes unless
+  a file name contains shell metacharacters, in which case programs
+  use shell quoting.  For example, although diff continues to output
+  "Only in a: b" as before for most file names, it now outputs
+  "Only in 'a: b': 'c: d'" instead of "Only in a: b: c: d" because the
+  file names 'a: b' and 'c: d' contain spaces.  For compatibility
+  with previous practice, diff -c and -u headers continue to quote for
+  C rather than for the shell.
+
+  diff now outputs more information when symbolic links differ, e.g.,
+  "Symbolic links ‘d/f’ -> ‘a’ and ‘e/f’ -> ‘b’ differ", not just
+  "Symbolic links d/f and e/f differ".  Special files too, e.g.,
+  "Character special files ‘d/f’ (1, 3) and ‘e/f’ (5, 0) differ", not
+  "File d/f is a character special file while file e/f is a character
+  special file".
+
+  diff's --ignore-case (-i) and --ignore-file-name-case options now
+  support multi-byte characters.  For example, they treat Greek
+  capital Δ like small δ when input uses UTF-8.
+
+  diff now supports multi-byte characters when treating white space.
+  In options like --expand-tabs (-t), --ignore-space-change (-b) and
+  --ignore-tab-expansion (-E), diff now recognizes non-ASCII space
+  characters and counts columns for non-ASCII characters.
+
+** Bug fixes
+
+  cmp -bl no longer omits "M-" from bytes with the high bit set in
+  single-byte locales like en_US.iso8859-1.  This fix causes the
+  behavior to be locale independent, and to be the same as the
+  longstanding behavior in the C locale and in locales using UTF-8.
+  [bug introduced in 2.9]
+
+  cmp -i N and -n N no longer fail merely because N is enormous.
+  [bug present since "the beginning"]
+
+  cmp -s no longer mishandles /proc files, for which the Linux kernel
+  reports a zero size even when nonempty.  For example, the following
+  shell command now outputs nothing, as it should:
+    cp /proc/cmdline t; cmp -s /proc/cmdline t || echo files differ
+  [bug present since "the beginning"]
+
+  diff -E no longer mishandles some input lines containing '\a', '\b',
+  '\f', '\r', '\v', or '\0'.
+  [bug present since 2.8]
+
+  diff -ly no longer mishandles non-ASCII input.
+  [bug#64461 introduced in 2.9]
+
+  diff - A/B now works correctly when standard input is a directory,
+  by reading a file named B in that directory.
+  [bug present since "the beginning"]
+
+  diff no longer suffers from race conditions in some cases
+  when comparing files in a mutating file system.
+  [bug present since "the beginning"]
+
+** Release
+
+  distribute gzip-compressed tarballs once again
+
+
+* Noteworthy changes in release 3.10 (2023-05-21) [stable]
+
+** Bug fixes
+
+  cmp/diff can again work with file dates past Y2K38
+  [bug introduced in 3.9]
+
+  diff -D no longer fails to output #ifndef lines.
+  [bug#61193 introduced in 3.9]
+
+
+* Noteworthy changes in release 3.9 (2023-01-15) [stable]
+
+** Bug fixes
+
+  diff -c and -u no longer output incorrect timezones in headers
+  on platforms like Solaris where struct tm lacks tm_gmtoff.
+  [bug#51228 introduced in 3.4]
+
+
+* Noteworthy changes in release 3.8 (2021-08-01) [stable]
+
+** Incompatible changes
+
+  diff no longer treats a closed stdin as representing an absent file
+  in usage like 'diff --new-file - foo <&-'.  This feature was rarely
+  if ever used and was not portable to POSIX platforms that reopen
+  stdin on exec, such as SELinux if the process underwent an AT_SECURE
+  transition, or HP-UX even if not setuid.
+  [bug#33965 introduced in 2.8]
+
+** Bug fixes
+
+  diff and related programs no longer get confused if stdin, stdout,
+  or stderr are closed.  Previously, they sometimes opened files into
+  file descriptors 0, 1, or 2 and then mistakenly did I/O with them
+  that was intended for stdin, stdout, or stderr.
+  [bug#33965 present since "the beginning"]
+
+  cmp, diff and sdiff no longer treat negative command-line
+  option-arguments as if they were large positive numbers.
+  [bug#35256 introduced in 2.8]
+
+
+* Noteworthy changes in release 3.7 (2018-12-31) [stable]
+
+** Bug fixes
+
+  diff --strip-trailing-cr with a single CR byte in one input file
+  would provoke an uninitialized memory read, e.g.,
+    diff -a --strip-trailing-cr <(printf '\r') <(echo a)
+  [bug introduced in 2.8 with addition of the --strip-trailing-cr option]
+
+** Improvements
+
+  diff --color now produces output compatible with less -R.
+
+
 * Noteworthy changes in release 3.6 (2017-05-21) [stable]
 
 ** New features
@@ -309,12 +453,12 @@ User-visible changes in version 2.4:
       printed according to the printf specification.
       E.g. '%5df' prints the number of the first line in the
       group in the old file using the "%5d" format.
-	e: line number just before the group in old file; equals f - 1
-	f: first line number in group in the old file
-	l: last line number in group in the old file
-	m: line number just after the group in old file; equals l + 1
-	n: number of lines in group in the old file; equals l - f + 1
-	E, F, L, M, N: likewise, for lines in the new file
+        e: line number just before the group in old file; equals f - 1
+        f: first line number in group in the old file
+        l: last line number in group in the old file
+        m: line number just after the group in old file; equals l + 1
+        n: number of lines in group in the old file; equals l - f + 1
+        E, F, L, M, N: likewise, for lines in the new file
   %(A=B?T:E)
       If A equals B then T else E.  A and B are each either a decimal
       constant or a single letter interpreted as above.  T and E are
@@ -404,7 +548,7 @@ User-visible changes in version 2.0:
 
 
 
-Copyright (C) 1993-1994, 1998, 2001-2002, 2004, 2006, 2009-2013, 2015-2017 Free
+Copyright (C) 1993-1994, 1998, 2001-2002, 2004, 2006, 2009-2013, 2015-2026 Free
 Software Foundation, Inc.
 
 This file is part of GNU Diffutils.

README

@@ -30,19 +30,12 @@ installation.  If you have an older version of libiconv, please
 upgrade to the latest one; see <ftp://ftp.gnu.org/gnu/libiconv/>.  If
 the problem seems isolated to diffutils, though, please report a bug.
 
-This program requires a Standard C compiler (C89 or later).  If you
+This program requires a Standard C compiler (C99 or later).  If you
 have a nonstandard compiler, please install GCC first.
 
 If you make changes to the source code, you may need appropriate
-versions of GNU build tools to regenerate the intermediate files.  The
-following versions were used to generate the intermediate files in
-this distribution:
-
-* Autoconf 2.59   <ftp://ftp.gnu.org/gnu/autoconf/autoconf-2.59.tar.gz>
-* Automake 1.8.3  <ftp://ftp.gnu.org/gnu/automake/automake-1.8.3.tar.gz>
-* gettext 0.14.1  <ftp://ftp.gnu.org/gnu/gettext/gettext-0.14.1.tar.gz>
-* help2man 1.33   <ftp://ftp.gnu.org/gnu/help2man/help2man-1.33.1.tar.gz>
-* Texinfo 4.7     <ftp://ftp.gnu.org/gnu/texinfo/texinfo-4.7.tar.gz>
+versions of GNU build tools to regenerate the intermediate files.
+See README-prereq for details.
 
 For any copyright year range specified as YYYY-ZZZZ in this package
 note that the range specifies every single year in that closed interval.
@@ -51,7 +44,7 @@ Please report bugs to <bug-diffutils@gnu.org>.
 
 -----
 
-Copyright (C) 1992, 1998, 2001-2002, 2004, 2009-2013, 2015-2017 Free Software
+Copyright (C) 1992, 1998, 2001-2002, 2004, 2009-2013, 2015-2026 Free Software
 Foundation, Inc.
 
 This file is part of GNU Diffutils.

README-hacking

@@ -1,58 +1,63 @@
--*- outline -*-
+Building from a Git repository				-*- outline -*-
 
 These notes intend to help people working on the checked-out sources.
 These requirements do not apply when building from a distribution tarball.
-See also HACKING for more detailed contribution guidelines.
+If this package has a file HACKING, please also read that file for
+more detailed contribution guidelines.
 
 * Requirements
 
-We've opted to keep only the highest-level sources in the GIT repository.
-This eases our maintenance burden, (fewer merges etc.), but imposes more
+We've opted to keep only the highest-level sources in the Git repository.
+This eases our maintenance burden (fewer merges etc.), but imposes more
 requirements on anyone wishing to build from the just-checked-out sources.
-Note the requirements to build the released archive are much less and
-are just the requirements of the standard ./configure && make procedure.
+(The requirements to build from a release are much less and are just
+the requirements of the standard './configure && make' procedure.)
 Specific development tools and versions will be checked for and listed by
 the bootstrap script.  See README-prereq for specific notes on obtaining
 these prerequisite tools.
 
-Valgrind <http://valgrind.org/> is also highly recommended, if
-Valgrind supports your architecture.
+Valgrind <https://valgrind.org/> is also highly recommended, if
+Valgrind supports your architecture.  See also README-valgrind
+(if present).
 
 While building from a just-cloned source tree may require installing a
-few prerequisites, later, a plain 'git pull && make' should be sufficient.
+few prerequisites, later, a plain 'git pull && make' typically suffices.
 
-* First GIT checkout
+* First Git checkout
 
 You can get a copy of the source repository like this:
 
-        $ git clone git://git.sv.gnu.org/diffutils
-        $ cd diffutils
+        $ git clone https://git.savannah.gnu.org/git/<packagename>
+        $ cd <packagename>
 
-As an optional step, if you already have a copy of the gnulib git
-repository on your hard drive, then you can use it as a reference to
-reduce download time and disk space requirements:
+where '<packagename>' stands for 'coreutils' or whatever other package
+you are building.
 
-        $ export GNULIB_SRCDIR=/path/to/gnulib
-
-The next step is to get and check other files needed to build,
-which are extracted from other source packages:
-
-        $ ./bootstrap
-
-To use the most-recent gnulib (as opposed to the gnulib version that
+To use the most-recent Gnulib (as opposed to the Gnulib version that
 the package last synchronized to), do this next:
 
         $ git submodule foreach git pull origin master
         $ git commit -m 'build: update gnulib submodule to latest' gnulib
 
+As an optional step, if you already have a copy of the Gnulib Git
+repository, then you can use it as a reference to reduce download
+time and file system space requirements:
+
+        $ export GNULIB_SRCDIR=/path/to/gnulib
+
+The next step is to get and check other files needed to build,
+which are extracted from other source packages [1]:
+
+        $ ./bootstrap
+
 And there you are!  Just
 
-        $ ./configure --quiet #[--enable-gcc-warnings] [*]
+        $ ./configure --quiet #[--disable-gcc-warnings] [2]
         $ make
         $ make check
 
 At this point, there should be no difference between your local copy,
-and the GIT master copy:
+and the Git master copy:
 
         $ git diff
 
@@ -60,27 +65,35 @@ should output no difference.
 
 Enjoy!
 
-[*] The --enable-gcc-warnings option is useful only with glibc
-and with a very recent version of gcc.  You'll probably also have
-to use recent system headers.  If you configure with this option,
-and spot a problem, please be sure to send the report to the bug
-reporting address of this package, and not to that of gnulib, even
-if the problem seems to originate in a gnulib-provided file.
+[1] You can also bootstrap in two phases: first, './bootstrap --pull'
+fetches auxiliary files from the network, and then './bootstrap --gen'
+generates files by using local tools without accessing the net.
+Although './bootstrap' by default does both phases, auditable builds
+can do them separately.
+
+[2] By default GCC warnings are enabled when building from Git.
+If you get warnings with recent GCC and Glibc with default
+configure-time options, please report the warnings to the bug
+reporting address of this package instead of to bug-gnulib,
+even if the problem seems to originate in a Gnulib-provided file.
+If you get warnings with other configurations, you can run
+'./configure --disable-gcc-warnings' or 'make WERROR_CFLAGS='
+to build quietly or verbosely, respectively.
+-----
 
 * Submitting patches
 
 If you develop a fix or a new feature, please send it to the
 appropriate bug-reporting address as reported by the --help option of
 each program.  One way to do this is to use vc-dwim
-<http://www.gnu.org/software/vc-dwim/>), as follows.
+<https://www.gnu.org/software/vc-dwim/>), as follows.
 
-  Run the command "vc-dwim --help", copy its definition of the
-  "git-changelog-symlink-init" function into your shell, and then run
-  this function at the top-level directory of the package.
+  Run the command "vc-dwim --initialize" from the top-level directory
+  of this package's git-cloned hierarchy.
 
   Edit the (empty) ChangeLog file that this command creates, creating a
   properly-formatted entry according to the GNU coding standards
-  <http://www.gnu.org/prep/standards/html_node/Change-Logs.html>.
+  <https://www.gnu.org/prep/standards/html_node/Change-Logs.html>.
 
   Make your changes.
 
@@ -92,9 +105,21 @@ each program.  One way to do this is to use vc-dwim
   Run the command "git format-patch --stdout -1", and email its output
   in, using the output's subject line.
 
+* Continuous integration
+
+There are two continuous integrations that regularly build and test
+diffutils:
+- On a Linux/glibc system only:
+  https://gitlab.com/gnu-diffutils/ci-distcheck/pipelines
+  https://gitlab.com/gnu-diffutils/ci-distcheck/-/jobs?scope=finished
+  This one will catch only the most blatant mistakes.
+- On many platforms:
+  https://github.com/gnu-diffutils/ci-check/actions
+  This one catches platform-specific bugs.
+
 -----
 
-Copyright (C) 2002-2007, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+Copyright (C) 2002-2026 Free Software Foundation, Inc.
 
 This program is free software: you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
@@ -107,4 +132,4 @@ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License
-along with this program.  If not, see <http://www.gnu.org/licenses/>.
+along with this program.  If not, see <https://www.gnu.org/licenses/>.

README-prereq

@@ -0,0 +1,39 @@
+This gives some notes on obtaining the tools required for development.
+These tools can be used by the 'bootstrap' and 'configure' scripts,
+as well as by 'make'.  They include:
+
+- Autoconf   <https://www.gnu.org/software/autoconf/>
+- Automake   <https://www.gnu.org/software/automake/>
+- Gettext    <https://www.gnu.org/software/gettext/>
+- Git        <https://git-scm.com/>
+- Gzip       <https://www.gnu.org/software/gzip/>
+- Help2man   <https://www.gnu.org/software/help2man/>
+- M4         <https://www.gnu.org/software/m4/>
+- Make       <https://www.gnu.org/software/make/>
+- Perl       <https://www.cpan.org/>
+- Tar        <https://www.gnu.org/software/tar/>
+- Texinfo    <https://www.gnu.org/software/texinfo/>
+- Wget       <http://www.gnu.org/software/wget/>
+- XZ Utils   <https://tukaani.org/xz/>
+
+It is generally better to use official packages for your system.
+If a package is not officially available you can build it from source
+and install it into a directory that you can then use to build this
+package.  If some packages are available but are too old, install the
+too-old versions first as they may be needed to build newer versions.
+
+Here is an example of how to build a program from source.  This
+example is for Autoconf; a similar approach should work for the other
+developer prerequisites.  This example assumes Autoconf 2.72; it
+should be OK to use a later version of Autoconf, if available.
+
+  prefix=$HOME/prefix   # (or wherever else you choose)
+  export PATH=$prefix/bin:$PATH
+  wget https://ftp.gnu.org/pub/gnu/autoconf/autoconf-2.72.tar.gz
+  gzip -d <autoconf-2.72.tar.gz | tar xf -
+  cd autoconf-2.72
+  ./configure --prefix=$prefix
+  make install
+
+Once the prerequisites are installed, you can build this package as
+described in README-hacking.

TODO

@@ -1,7 +1,15 @@
+Add --git option to generate output compatible with 'git diff -p'.
+This would behave like 'diff -p', except that it would also generate
+the extended headers 'old mode', 'new mode', 'deleted file mode', and
+'new file mode', and it would quote file names with unusual characters.
+GNU patch already parses this format.
+
 Add --include option (opposite of --exclude).
 
+Add SEEK_DATA/SEEK_HOLE sparse file optimization to cmp, diff -q, etc.
+
 Look into sdiff improvement here:
-http://www.pkix.net/~chuck/sdiff2.diff
+https://web.archive.org/web/20090106164131/http://www.pkix.net/~chuck/sdiff2.diff
 
 Propagate stderr from subprocess so that diff3 does
 a better job of explaining _why_:

bootstrap.conf

@@ -1,6 +1,6 @@
 # Bootstrap configuration.
 
-# Copyright (C) 2006-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2006-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -15,14 +15,52 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
+avoided_gnulib_modules='
+  --avoid=hard-locale
+  --avoid=localename-unsafe
+  --avoid=localename-unsafe-limited
+  --avoid=lock-tests
+  --avoid=mbuiter
+  --avoid=mbuiterf
+  --avoid=setlocale
+'
+# These tests depend on localcharset, which we avoid.
+avoided_gnulib_tests='
+  --avoid=c32width-tests
+  --avoid=quotearg-simple-tests
+  --avoid=regex-tests
+  --avoid=wcwidth-tests
+'
+# These test whether all bytes are valid in the C locale,
+# which we don't care about.
+avoided_gnulib_tests=$avoided_gnulib_tests'
+  --avoid=btoc32-tests
+  --avoid=btowc-tests
+  --avoid=mbrtoc32-tests
+  --avoid=mbrtowc-tests
+  --avoid=mbsrtoc32s-tests
+  --avoid=mbsrtowcs-tests
+'
 
 # gnulib modules used by this package.
 gnulib_modules='
 announce-gen
 argmatch
+assert-h
+attribute
 binary-io
+bool
+builtin-expect
+c-ctype
+c-file-type
 c-stack
+c32isprint
+c32isspace
+c32tolower
+c32width
+careadlinkat
 config-h
+d-type
 diffseq
 dirname
 do-release-commit-and-tag
@@ -31,99 +69,151 @@ error
 exclude
 exitfail
 extensions
+extern-inline
 fcntl
-fdl
-file-type
+fdopendir
 filenamecat
+flexmember
 fnmatch-gnu
-getopt
+fopen-gnu
+fstatat
+getopt-gnu
 gettext-h
-gettime
 git-version-gen
 gitlog-to-changelog
 gnu-make
 gnu-web-doc-update
 gnumakefile
 gnupload
-hard-locale
-inttostr
-inttypes
-isblank
+ialloc
+idx
+intprops
+inttypes-h
 largefile
 lstat
 maintainer-makefile
 manywarnings
-mbrtowc
+mbscasecmp
+mcel-prefer
+mempcpy
+minmax
 mkstemp
 mktime
+nstrftime-limited
+nullptr
+openat
+pclose
+perl
+popen
 progname
-propername
+propername-lite
+quote
+raise
 rawmemchr
+readdir
+readlinkat
 readme-release
 regex
+same-inode
 sh-quote
-signal
+sigaction
+signal-h
+sigprocmask
 stat
 stat-macros
+stat-size
 stat-time
-stdint
+stdc_bit_width
+stdckdint-h
+stdcountof-h
+stdint-h
+stpcpy
 strcase
-strftime
 strptime
-strtoumax
-sys_wait
+strtoimax
 system-quote
-unistd
+sys_types-h
+sys_wait-h
+timespec
+timespec_get
+time_rz
+unistd-h
 unlocked-io
 update-copyright
-vararrays
 verify
 version-etc
 version-etc-fsf
-wcwidth
 xalloc
 xfreopen
-xreadlink
-xstrtoumax
-xvasprintf
+xmalloca
+xstdopen
+xstrtoimax
+year2038
 '
 
 # Additional xgettext options to use.  Use "\\\newline" to break lines.
 XGETTEXT_OPTIONS=$XGETTEXT_OPTIONS'\\\
+ --flag=asnprintf:3:c-format\\\
+ --flag=asprintf:2:c-format\\\
  --from-code=UTF-8\\\
- --flag=asprintf:2:c-format --flag=vasprintf:2:c-format\\\
- --flag=asnprintf:3:c-format --flag=vasnprintf:3:c-format\\\
- --flag=message:1:c-format --flag=message5:1:c-format\\\
+ --flag=message:1:c-format\\\
  --flag=try_help:1:c-format\\\
+ --flag=vasprintf:2:c-format\\\
+ --flag=vasnprintf:3:c-format\\\
 '
 
-gnulib_tool_option_extras="--tests-base=gnulib-tests --with-tests
-  --avoid=localename
-  --avoid=lock
-  --symlink
-  --makefile-name=gnulib.mk
+gnulib_tool_option_extras="--tests-base=gnulib-tests
+ --local-dir=gl
+ --with-tests
+ --symlink
+ --makefile-name=gnulib.mk
+ $avoided_gnulib_modules
+ $avoided_gnulib_tests
 "
 
 # Build prerequisites
 buildreq="\
-autoconf   2.61
-automake   1.12.2
-autopoint  -
+autoconf   2.64
+automake   1.14
+autopoint  0.19.2
 gettext    0.19.2
 git        1.4.4
 gperf      -
 gzip       -
 help2man   -
+m4         -
 makeinfo   -
 perl       5.5
-rsync      -
+texi2pdf   6.1
 tar        -
+wget       -
+xz         -
 "
 
-bootstrap_post_import_hook()
+bootstrap_post_import_hook ()
 {
   # Automake requires that ChangeLog exist.
   touch ChangeLog || exit 1
+
+  # Copy tests/init.sh from Gnulib.
+  $gnulib_tool --copy-file tests/init.sh
+
+  # Copy pkg-config's pkg.m4 so that our downstream users don't need to.
+  local ac_dir=`aclocal --print-ac-dir`
+  test -s "$ac_dir/dirlist" && ac_dir=$ac_dir:`tr '\n' : < "$ac_dir/dirlist"`
+  oIFS=$IFS
+  IFS=:
+  for dir in \
+    $ACLOCAL_PATH $ac_dir /usr/share/aclocal ''
+  do
+    IFS=$oIFS
+    if test -n "$dir" && test -r "$dir/pkg.m4"; then
+      cp "$dir/pkg.m4" m4/pkg.m4
+      return
+    fi
+  done
+  IFS=$oIFS
+  die 'Cannot find pkg.m4; perhaps you need to install pkg-config'
 }
 
 bootstrap_epilogue()

cfg.mk

@@ -1,5 +1,5 @@
 # Customize maint.mk                           -*- makefile -*-
-# Copyright (C) 2003-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2003-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -16,7 +16,7 @@
 
 # Use the direct link.  This is guaranteed to work immediately, while
 # it can take a while for the faster mirror links to become usable.
-url_dir_list = http://ftp.gnu.org/gnu/$(PACKAGE)
+url_dir_list = https://ftp.gnu.org/gnu/$(PACKAGE)
 
 # Used in maint.mk's web-manual rule
 manual_title = Comparing and Merging Files
@@ -24,20 +24,30 @@ manual_title = Comparing and Merging Files
 # Tests not to run as part of "make distcheck".
 local-checks-to-skip =		\
   sc_error_message_period	\
-  sc_error_message_uppercase
+  sc_error_message_uppercase	\
+  sc_indent
 
 # Tools used to bootstrap this package, used for "announcement".
 bootstrap-tools = autoconf,automake,gnulib
 
+# Override the default Cc: used in generating an announcement.
+announcement_Cc_ = $(translation_project_), $(PACKAGE)-devel@gnu.org
+
 # Now that we have better tests, make this the default.
 export VERBOSE = yes
 
-old_NEWS_hash = 0216ec3bf3e3322f33afd4e949a9a29b
+old_NEWS_hash = cfdcec044a361bf0e7b4e860bacc3d53
 
 # Tell maint.mk's syntax-check rules that diff gets config.h directly or
 # via diff.h or system.h.
 config_h_header = (<config\.h>|"(diff|system)\.h")
 
+# Write base64-encoded (not hex) checksums into the announcement.
+announce_gen_args = --cksum-checksums
+
+# Add an exemption for sc_makefile_at_at_check.
+_makefile_at_at_check_exceptions = ' && !/MAKEINFO/'
+
 update-copyright-env = \
   UPDATE_COPYRIGHT_USE_INTERVALS=1 \
   UPDATE_COPYRIGHT_MAX_LINE_LENGTH=79
@@ -66,7 +76,14 @@ config-save:
 	ln -nsf $(_date_time) $(_cf_state_dir)/latest
 	cp lib/config.h config.status $(_cf_state_dir)/latest
 
-exclude_file_name_regexp--sc_space_tab = ^gl/lib/.*\.c\.diff$$
+exclude_file_name_regexp--sc_GPL_version = ^gl/lib/
+exclude_file_name_regexp--sc_bindtextdomain = ^gl/tests/
+exclude_file_name_regexp--sc_doubled_words = ^gl/lib/mcel\.h$$
+exclude_file_name_regexp--sc_prohibit_doubled_word = ^(gl/lib/mcel\.h|tests/y2038-vs-32bit)$$
+exclude_file_name_regexp--sc_prohibit_strcmp = ^gl/lib/
 
-# Tell gnulib's tight_scope rule that we mark externs with XTERN
-export _gl_TS_extern = extern|XTERN
+# Tell gnulib's tight_scope rule that we mark extern inlines with
+# DIFF_INLINE and SYSTEM_INLINE.
+export _gl_TS_extern = extern|DIFF_INLINE|SYSTEM_INLINE
+
+codespell_ignore_words_list = FO,ND,debbugs

configure.ac

@@ -1,6 +1,6 @@
 # Configure template for GNU Diffutils.
 
-# Copyright (C) 1994-1995, 1998, 2001-2002, 2004, 2006, 2009-2013, 2015-2017
+# Copyright (C) 1994-1995, 1998, 2001-2002, 2004, 2006, 2009-2013, 2015-2026
 # Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
@@ -16,145 +16,193 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-AC_PREREQ(2.61)
+AC_PREREQ([2.64])
 
 AC_INIT([GNU diffutils],
         m4_esyscmd([build-aux/git-version-gen .tarball-version]),
-        [bug-diffutils@gnu.org])
+        [bug-diffutils@gnu.org],
+	[diffutils],
+	[https://www.gnu.org/software/diffutils/])
 
 AC_CONFIG_SRCDIR([src/diff.c])
 AC_CONFIG_AUX_DIR([build-aux])
 AC_CONFIG_MACRO_DIR([m4])
 
-AM_INIT_AUTOMAKE([1.11 dist-xz no-dist-gzip subdir-objects
-                       color-tests parallel-tests])
+AM_INIT_AUTOMAKE([1.11 dist-xz subdir-objects color-tests parallel-tests])
 AM_SILENT_RULES([yes]) # make --enable-silent-rules the default.
 
-AC_CONFIG_HEADER([lib/config.h:lib/config.hin])
+AC_CONFIG_HEADERS([lib/config.h:lib/config.hin])
 
 AC_PROG_AWK
 AC_PROG_CC
-AM_PROG_CC_C_O
 AM_MISSING_PROG([HELP2MAN], [help2man])
 AC_PROG_RANLIB
+
+# Do not worry about multibyte C locales.
+# This removes dependencies on hard-locale etc.
+gl_cv_func_mbrtowc_C_locale_sans_EILSEQ="guessing yes"
+gl_cv_func_mbrtoc32_C_locale_sans_EILSEQ="guessing yes"
+
+# Default Gnulib configuration to no multithreading,
+# since diffutils is single-threaded.
+# Although the builder can override this with the --enable-threads option
+# of 'configure', diffutils does not support that on all platforms.
+m4_define([gl_THREADLIB_DEFAULT_NO])
+
 gl_EARLY
 gl_USE_SYSTEM_EXTENSIONS
 gl_INIT
 
+# Ensure VLAs are not used.
+# Note -Wvla is implicitly added by gl_MANYWARN_ALL_GCC
+AC_DEFINE([GNULIB_NO_VLA], [1], [Define to 1 to disable use of VLAs])
+
+# diffutils is single-threaded; optimize for this.
+AC_DEFINE([GNULIB_EXCLUDE_SINGLE_THREAD], [1],
+  ['exclude' code is called only from 1 thread.])
+AC_DEFINE([GNULIB_MBRTOWC_SINGLE_THREAD], [1],
+  ['mbrtowc', 'mbrtoc32', 'regex' code is called only from 1 thread.])
+AC_DEFINE([GNULIB_REGEX_SINGLE_THREAD], [1],
+  ['regex' code is called only from 1 thread.])
+AC_DEFINE([GNULIB_WCHAR_SINGLE_LOCALE], [1],
+  [Locale-sensitive functions are called only after locale is set.])
+
+AC_DEFINE([GNULIB_MBRTOC32_REGULAR], [1],
+  [Do not worry about rare encodings like CP864, EBCDIC, Johab, and Shift JIS
+   that glibc does not support.])
+
+AC_DEFINE([REQUIRE_GNUISH_STRFTIME_AM_PM], [0],
+  [Do not worry about GNU strftime behavior for AM and PM indicators.])
+AC_DEFINE([SUPPORT_NON_GREG_CALENDARS_IN_STRFTIME], [false],
+  [Do not worry about GNU strftime behavior for non-Gregorian calendars.])
+
+# Diffutils translates Gnulib's msgids too.
+AC_DEFINE([GNULIB_TEXT_DOMAIN], [PACKAGE],
+  [Textdomain to use when translating Gnulib's msgids.])
+
+AC_C_INLINE
+
+AC_CHECK_MEMBERS([struct stat.st_rdev])
+AC_HEADER_DIRENT
+AC_HEADER_SYS_WAIT
+AC_TYPE_PID_T
+
+AC_FUNC_FORK
+
+dnl O_PATH exists since Linux 2.6.39, but is supported with fstat() only since
+dnl Linux 3.6. Use a configure test rather than testing `uname -sr`.
+AC_CACHE_CHECK([whether O_PATH supports fstat],
+  [du_cv_O_PATH_fstat],
+  [AC_RUN_IFELSE(
+     [AC_LANG_PROGRAM([[
+        #include <sys/types.h>
+        #include <sys/stat.h>
+        #include <fcntl.h>
+        ]], [[
+        struct stat statbuf;
+        int fd = openat (AT_FDCWD, "conftest.c", O_RDONLY|O_PATH);
+        return !(fd >= 0 && fstat (fd, &statbuf) == 0);
+        ]])
+     ],
+     [du_cv_O_PATH_fstat=yes],
+     [du_cv_O_PATH_fstat=no],
+     [du_cv_O_PATH_fstat="guessing no"])
+  ])
+case "$du_cv_O_PATH_fstat" in
+  yes)
+    AC_DEFINE([O_PATH_SUPPORTS_FSTAT], [1],
+      [Define to 1 if <fcntl.h> defines O_PATH and file descriptors
+       created with this flag are supported by fstat().])
+    ;;
+esac
+
+# gl_GCC_VERSION_IFELSE([major], [minor], [run-if-found], [run-if-not-found])
+# ------------------------------------------------
+# If $CPP is gcc-MAJOR.MINOR or newer, then run RUN-IF-FOUND.
+# Otherwise, run RUN-IF-NOT-FOUND.
+AC_DEFUN([gl_GCC_VERSION_IFELSE],
+  [AC_PREPROC_IFELSE(
+    [AC_LANG_PROGRAM(
+      [[
+#if ($1) < __GNUC__ || (($1) == __GNUC__ && ($2) <= __GNUC_MINOR__)
+/* ok */
+#else
+# error "your version of gcc is older than $1.$2"
+#endif
+      ]]),
+    ], [$3], [$4])
+  ]
+)
+
 AC_ARG_ENABLE([gcc-warnings],
-  [AS_HELP_STRING([--enable-gcc-warnings],
-                  [turn on lots of GCC warnings (for developers)])],
+  [AS_HELP_STRING([[--enable-gcc-warnings[=TYPE]]],
+    [control generation of GCC warnings.  The TYPE 'no' disables
+     warnings (default for non-developer builds); 'yes' generates
+     cheap warnings if available (default for developer builds);
+     'expensive' in addition generates expensive-to-compute warnings
+     if available.])],
   [case $enableval in
-     yes|no) ;;
+     no|yes|expensive) ;;
      *)      AC_MSG_ERROR([bad value $enableval for gcc-warnings option]) ;;
    esac
    gl_gcc_warnings=$enableval],
-  [if test -d "$srcdir"/.git; then
-     gl_gcc_warnings=yes
-   else
-     gl_gcc_warnings=no
-   fi]
+  [
+   # GCC provides fine-grained control over diagnostics which
+   # is used in gnulib for example to suppress warnings from
+   # certain sections of code.  So if this is available and
+   # we're running from a git repo, then auto enable the warnings.
+   gl_gcc_warnings=no
+   gl_GCC_VERSION_IFELSE([4], [6],
+                         [test -d "$srcdir"/.git \
+                          && ! test -f "$srcdir"/.tarball-version \
+                          && gl_gcc_warnings=yes])]
 )
 
-if test "$gl_gcc_warnings" = yes; then
+if test $gl_gcc_warnings != no; then
   gl_WARN_ADD([-Werror], [WERROR_CFLAGS])
   AC_SUBST([WERROR_CFLAGS])
 
-  nw=
+  ew=
+  AS_IF([test $gl_gcc_warnings != expensive],
+    [# -fanalyzer and related options slow GCC considerably.
+     ew="$ew -fanalyzer -Wno-analyzer-malloc-leak"])
+
   # This, $nw, is the list of warnings we disable.
-  nw="$nw -Wdeclaration-after-statement" # too useful to forbid
-  nw="$nw -Waggregate-return"       # anachronistic
-  nw="$nw -Wlong-long"              # C90 is anachronistic (lib/gethrxtime.h)
-  nw="$nw -Wc++-compat"             # We don't care about C++ compilers
-  nw="$nw -Wundef"                  # Warns on '#if GNULIB_FOO' etc in gnulib
-  nw="$nw -Wtraditional"            # Warns on #elif which we use often
-  nw="$nw -Wcast-qual"              # Too many warnings for now
-  nw="$nw -Wconversion"             # Too many warnings for now
-  nw="$nw -Wsystem-headers"         # Don't let system headers trigger warnings
-  nw="$nw -Wsign-conversion"        # Too many warnings for now
-  nw="$nw -Wtraditional-conversion" # Too many warnings for now
-  nw="$nw -Wunreachable-code"       # Too many warnings for now
-  nw="$nw -Wpadded"                 # Our structs are not padded
-  nw="$nw -Wredundant-decls"        # openat.h declares e.g., mkdirat
-  nw="$nw -Wlogical-op"             # any use of fwrite provokes this
-  nw="$nw -Wformat-nonliteral"      # who.c and pinky.c strftime uses
-  nw="$nw -Wvla"                    # warnings in gettext.h
-  nw="$nw -Wnested-externs"         # use of XARGMATCH/verify_function__
-  nw="$nw -Wswitch-enum"            # Too many warnings for now
-  nw="$nw -Wswitch-default"         # Too many warnings for now
-  nw="$nw -Wstack-protector"        # not worth working around
-  # things I might fix soon:
-  nw="$nw -Wfloat-equal"            # sort.c, seq.c
-  nw="$nw -Wmissing-format-attribute" # copy.c
-  nw="$nw -Wunsafe-loop-optimizations" # a few src/*.c
-  nw="$nw -Winline"                 # system.h's readdir_ignoring_dot_and_dotdot
-  nw="$nw -Wstrict-overflow"        # expr.c, pr.c, tr.c, factor.c
-  nw="$nw -Wformat-extra-args"      # sdiff.c
-  # ?? -Wstrict-overflow
+  nw=$ew
+  nw="$nw -Winline"                 # not a correctness warning
+  nw="$nw -Wstack-protector"        # not a correctness warning
 
   gl_MANYWARN_ALL_GCC([ws])
   gl_MANYWARN_COMPLEMENT([ws], [$ws], [$nw])
   for w in $ws; do
     gl_WARN_ADD([$w])
   done
-  gl_WARN_ADD([-Wno-missing-field-initializers]) # We need this one
-  gl_WARN_ADD([-Wno-sign-compare])     # Too many warnings for now
-  gl_WARN_ADD([-Wno-pointer-sign])     # Too many warnings for now
-  gl_WARN_ADD([-Wno-unused-parameter]) # Too many warnings for now
+  gl_WARN_ADD([-Wno-sign-compare])     # Too many false positives
   gl_WARN_ADD([-Wno-format-nonliteral])
 
-  # In spite of excluding -Wlogical-op above, it is enabled, as of
-  # gcc 4.5.0 20090517, and it provokes warnings in cat.c, dd.c, truncate.c
-  gl_WARN_ADD([-Wno-logical-op])
-
-  gl_WARN_ADD([-fdiagnostics-show-option])
-  gl_WARN_ADD([-funit-at-a-time])
-  gl_WARN_ADD([-fno-common])
-
   AC_SUBST([WARN_CFLAGS])
 
-  AC_DEFINE([lint], [1], [Define to 1 if the compiler is checking for lint.])
-  AC_DEFINE([_FORTIFY_SOURCE], [2],
-    [enable compile-time and run-time bounds-checking, and some warnings])
+  AH_VERBATIM([FORTIFY_SOURCE],
+  [/* Enable compile-time and run-time bounds-checking, and some warnings,
+      without upsetting glibc 2.15+. */
+   #if !defined _FORTIFY_SOURCE && defined __OPTIMIZE__ && __OPTIMIZE__ \
+       && !defined __MINGW32__
+   # define _FORTIFY_SOURCE 2
+   #endif
+  ])
   AC_DEFINE([GNULIB_PORTCHECK], [1], [enable some gnulib portability checks])
-
-  # We use a slightly smaller set of warning options for lib/.
-  # Remove the following and save the result in GNULIB_WARN_CFLAGS.
-  nw=
-  nw="$nw -Wunused-macros"
-  nw="$nw -Wmissing-prototypes"
-  nw="$nw -Wold-style-definition"
-  nw="$nw -Wsuggest-attribute=pure"
-  gl_MANYWARN_COMPLEMENT([GNULIB_WARN_CFLAGS], [$WARN_CFLAGS], [$nw])
-  AC_SUBST([GNULIB_WARN_CFLAGS])
 fi
 
-AC_C_INLINE
-AC_C_VARARRAYS
-
 AC_DEFINE([DEFAULT_EDITOR_PROGRAM], ["ed"],
   [Name of editor program, unless overridden.])
 
 AC_PATH_PROG([PR_PROGRAM], [pr], [""])
 AC_DEFINE_UNQUOTED([PR_PROGRAM], ["$PR_PROGRAM"], [Name of "pr" program.])
 
-AC_CHECK_MEMBERS([struct stat.st_blksize])
-AC_CHECK_MEMBERS([struct stat.st_rdev])
-AC_HEADER_DIRENT
-AC_HEADER_SYS_WAIT
-AC_TYPE_PID_T
-
-AC_CHECK_FUNCS_ONCE([sigaction sigprocmask strcasecoll stricoll])
-if test $ac_cv_func_sigprocmask = no; then
-  AC_CHECK_FUNCS([sigblock])
-fi
-AC_FUNC_CLOSEDIR_VOID
-AC_FUNC_FORK
-
 # When .tarball-version exists, we're building from a tarball
 # and must not make man/*.1 files depend on the generated src/version.c,
-# because that would induce a requirement to run the help2man perl script.
-# We are not yet prepared to make perl a build-from-tarball requirement.
+# because that would induce a requirement to run help2man.
+# We are not yet prepared to make help2man a build-from-tarball requirement.
 # Hence, here we detect .tarball-version existence.  When not present,
 # we define a variable to be used in man/Makefile.am to induce the
 # proper dependency (so that man/*.1 will be rebuilt upon any version change),
@@ -164,7 +212,7 @@ test -f $srcdir/.tarball-version \
   && SRC_VERSION_C= \
   || SRC_VERSION_C=../src/version.c
 
-AM_GNU_GETTEXT([external], [need-ngettext])
+AM_GNU_GETTEXT([external], [need-formatstring-macros])
 AM_GNU_GETTEXT_VERSION([0.19.2])
 XGETTEXT="AWK='$AWK' \$(SHELL) \$(top_srcdir)/exgettext $XGETTEXT"
 

doc/Makefile.am

@@ -1,6 +1,6 @@
 # Makefile for GNU diffutils documentation.
 
-# Copyright (C) 2001-2002, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2001-2002, 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -15,6 +15,11 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
+
+# The customization variable CHECK_NORMAL_MENU_STRUCTURE is necessary with
+# makeinfo versions ≥ 6.8.
+MAKEINFO = @MAKEINFO@ -c CHECK_NORMAL_MENU_STRUCTURE=1
+
 AM_MAKEINFOFLAGS = --no-split
 
 info_TEXINFOS = diffutils.texi

doc/diffutils.texi

@@ -3,6 +3,9 @@
 @setfilename diffutils.info
 @include version.texi
 @settitle Comparing and Merging Files
+@documentencoding UTF-8
+@set txicodequoteundirected
+@set txicodequotebacktick
 @syncodeindex vr cp
 @setchapternewpage odd
 @comment %**end of header
@@ -14,8 +17,8 @@ and documents the GNU @command{diff}, @command{diff3},
 differences between files and the GNU @command{patch} command for
 using their output to update files.
 
-Copyright @copyright{} 1992-1994, 1998, 2001-2002, 2004, 2006, 2009-2017 Free
-Software Foundation, Inc.
+Copyright @copyright{} 1992--1994, 1998, 2001--2002, 2004, 2006, 2009--2026
+Free Software Foundation, Inc.
 
 @quotation
 Permission is granted to copy, distribute and/or modify this document
@@ -235,6 +238,7 @@ need not read them to do its job.
 * Case Folding::      Suppressing differences in alphabetic case.
 * Brief::             Summarizing which files are different.
 * Binary::            Comparing binary files or forcing text comparisons.
+* Mutating Files::    Comparing files that are changing while being read.
 @end menu
 
 @node Hunks
@@ -388,12 +392,42 @@ line against each regular expression.
 @section Suppressing Case Differences
 @cindex case difference suppression
 
-GNU @command{diff} can treat lower case letters as
-equivalent to their upper case counterparts, so that, for example, it
+GNU @command{diff} can treat capital letters as
+equivalent to their small counterparts, so that, for example, it
 considers @samp{Funky Stuff}, @samp{funky STUFF}, and @samp{fUNKy
 stuFf} to all be the same.  To request this, use the @option{-i} or
 @option{--ignore-case} option.
 
+When ignoring case @command{diff} downcases each character before comparing it.
+For example, @samp{diff -i} downcases Greek capital ``Δ'' to small
+``δ'' before comparison.  Although this works for many cases, it may
+have problems in some.  For example:
+
+@itemize @bullet
+@item
+The German word ``Straße'' is not treated like ``STRASSE'' as
+many Germans might expect, because the two words downcase to
+``straße'' and ``strasse'' respectively.
+
+@item
+The Greek name ``Κως'' is not treated like ``ΚΩΣ''
+even though they both end in sigma,
+because the latter's trailing ``Σ'' (U+03A3 GREEK CAPITAL LETTER SIGMA)
+downcases to ``σ'' (U+03C3 GREEK SMALL LETTER SIGMA) instead
+of to ``ς'' (U+03C2 GREEK SMALL LETTER FINAL SIGMA).
+
+@c pdftex cannot handle combining characters, alas.
+@ifnottex
+@item
+The ancient Greek phrase ``Πάντα ῥεῖ'' is not treated like ``ΠΆΝΤΑ ῬΕΙ͂'',
+as the former's trailing ``ῖ'' is the single character U+1FD6 (GREEK
+SMALL LETTER IOTA WITH PERISPOMENI) whereas the latter's trailing
+``Ι͂'' has two characters, U+0399 (GREEK CAPITAL LETTER IOTA) and
+U+0342 (COMBINING GREEK PERISPOMENI).
+@end ifnottex
+
+@end itemize
+
 @node Brief
 @section Summarizing Which Files Differ
 @cindex summarizing which files differ
@@ -497,6 +531,28 @@ text files, you can force @command{diff3} to consider all files to be text
 files and compare them line by line by using the @option{-a} or
 @option{--text} option.
 
+@node Mutating Files
+@section Comparing Files While They Are Changing
+@cindex mutating files
+
+Commands like @command{diff} read files and directories by using a
+series of system calls, as operating systems typically do not give a
+program a way to snapshot all its input before processing it.
+If input files change while being read, @command{diff} and related
+commands can generate output that does not correspond to any state of
+the input file system.
+
+For example, if some other process truncates a file to zero size when
+@command{diff} has read just half the file, @command{diff} will use
+the half that it read as its input data.  As another example, if a
+recursive @command{diff} determines that a file is a directory, and
+then some other process replaces the directory with a regular file,
+and then @command{diff} attempts to descend into the directory,
+@command{diff} will issue a diagnostic and fail.
+
+In these situations the output of @command{diff} and related programs
+should be interpreted with some care.
+
 @node Output Formats
 @chapter @command{diff} Output Formats
 @cindex output formats
@@ -766,8 +822,8 @@ or @option{-u} option.
 The argument @var{lines} is the number of lines of context to show.
 When it is not given, it defaults to three.
 
-At present, only GNU @command{diff} can produce this format and
-only GNU @command{patch} can automatically apply diffs in this
+In the early 1990s, only GNU @command{diff} could produce this format and
+only GNU @command{patch} could automatically apply diffs in this
 format.  For proper operation, @command{patch} typically needs at
 least three lines of context.
 
@@ -1737,7 +1793,7 @@ output formats (@pxref{Output Formats}) @command{diff} cannot
 represent an incomplete line, so it pretends there was a newline and
 reports an error.
 
-For example, suppose @file{F} and @file{G} are one-byte files that
+For example, suppose @file{F} and @file{G} are one-character files that
 contain just @samp{f} and @samp{g}, respectively.  Then @samp{diff F G}
 outputs
 
@@ -1796,8 +1852,7 @@ second position.)  To do this, use the @option{--new-file}
 (@option{-N}) option.  This option affects command-line arguments as
 well as files found via directory traversal; for example, @samp{diff
 -N a b} treats @file{a} as empty if @file{a} does not exist but
-@file{b} does, and similarly @samp{diff -N - b} treats standard input
-as empty if it is closed but @file{b} exists.
+@file{b} does.
 
 If the older directory contains large files that are not in
 the newer directory, you can make the patch smaller by using the
@@ -1872,6 +1927,10 @@ These adjustments can be applied to any output format.
 @cindex tab stop alignment
 @cindex aligning tab stops
 
+The tab character moves the cursor to the next tab stop.
+Tab stops are normally every 8 display columns;
+this can be altered by the @option{--tabsize=@var{columns}} option.
+
 The lines of text in some of the @command{diff} output formats are
 preceded by one or two characters that indicate whether the text is
 inserted, deleted, or changed.  The addition of those characters can
@@ -1884,9 +1943,7 @@ number of spaces before outputting them; select this method with the
 @option{--expand-tabs} (@option{-t}) option.  To use this form of output with
 @command{patch}, you must give @command{patch} the @option{-l} or
 @option{--ignore-white-space} option (@pxref{Changed White Space}, for more
-information).  @command{diff} normally assumes that tab stops are set
-every 8 print columns, but this can be altered by the
-@option{--tabsize=@var{columns}} option.
+information).
 
 The other method for making tabs line up correctly is to add a tab
 character instead of a space after the indicator character at the
@@ -1899,6 +1956,15 @@ output format, which does not have a space character after the change
 type indicator character.  Select this method with the @option{-T} or
 @option{--initial-tab} option.
 
+GNU @command{diff} currently assumes that the output device respects tab stops,
+displays each character with column width as given by the operating system,
+and displays each encoding error byte in a single column.
+Unfortunately these assumptions are often incorrect
+for encoding errors and non-ASCII characters,
+so complex input data may not line up properly on output,
+and analysis based on the @option{--ignore-tab-expansion} (@option{-E}) option
+may differ from the display device's behavior.
+
 @node Trailing Blanks
 @section Omitting trailing blanks
 @cindex trailing blanks
@@ -3544,9 +3610,12 @@ line word: @option{-bl} is equivalent to @option{-b -l}.
 @table @option
 @item -b
 @itemx --print-bytes
-Print the differing bytes.  Display control bytes as a
-@samp{^} followed by a letter of the alphabet and precede bytes
-that have the high bit set with @samp{M-} (which stands for ``meta'').
+Print the differing bytes.  Display each control byte as a @samp{^}
+followed by an ASCII letter, @samp{@@}, @samp{[}, @samp{\}, @samp{]},
+@samp{^} or @samp{_}.  Precede each byte with the high bit set with
+@samp{M-}, which stands for ``meta''.  A control byte is any byte
+with the high bit clear that does not represent a printable ASCII
+character including space.
 
 @item --help
 Output a summary of usage and then exit.
@@ -3588,20 +3657,16 @@ In the above table, operands that are byte counts are normally
 decimal, but may be preceded by @samp{0} for octal and @samp{0x} for
 hexadecimal.
 
-A byte count can be followed by a suffix to specify a multiple of that
-count; in this case an omitted integer is understood to be 1.  A bare
-size letter, or one followed by @samp{iB}, specifies a multiple using
-powers of 1024.  A size letter followed by @samp{B} specifies powers
-of 1000 instead.  For example, @option{-n 4M} and @option{-n 4MiB} are
-equivalent to @option{-n 4194304}, whereas @option{-n 4MB} is
-equivalent to @option{-n 4000000}.  This notation is upward compatible
-with the @uref{http://www.bipm.fr/enus/3_SI/si-prefixes.html, SI
-prefixes} for decimal multiples and with the
-@uref{http://physics.nist.gov/cuu/Units/binary.html, IEC 60027-2
-prefixes for binary multiples}.
+A byte count can be followed by a suffix to specify a
+multiple of the count.  A bare letter,
+or one followed by @samp{iB}, specifies
+a multiple using powers of 1024.  A letter followed by @samp{B}
+specifies powers of 1000 instead.  For example, @samp{-n 1M} and
+@samp{-n 1MiB} are equivalent to @samp{-n 1048576}, whereas @samp{-n 1MB} is
+equivalent to @samp{-n 1000000}.
 
-The following suffixes are defined.  Large sizes like @code{1Y} may be
-rejected by your computer due to limitations of its arithmetic.
+The following suffixes are defined.  Large sizes like @code{1Q}
+may be rejected by your computer due to limitations of its arithmetic.
 
 @table @samp
 @item kB
@@ -3611,8 +3676,8 @@ kilobyte: @math{10^3 = 1000}.
 @itemx K
 @itemx KiB
 @cindex kibibyte, definition of
-kibibyte: @math{2^10 = 1024}.  @samp{K} is special: the SI prefix is
-@samp{k} and the IEC 60027-2 prefix is @samp{Ki}, but tradition and
+kibibyte: @math{2^{10} = 1024}.  @samp{K} is special: the SI prefix is
+@samp{k} and the ISO/IEC 80000-13 prefix is @samp{Ki}, but tradition and
 POSIX use @samp{k} to mean @samp{KiB}.
 @item MB
 @cindex megabyte, definition of
@@ -3620,49 +3685,59 @@ megabyte: @math{10^6 = 1,000,000}.
 @item M
 @itemx MiB
 @cindex mebibyte, definition of
-mebibyte: @math{2^20 = 1,048,576}.
+mebibyte: @math{2^{20} = 1,048,576}.
 @item GB
 @cindex gigabyte, definition of
 gigabyte: @math{10^9 = 1,000,000,000}.
 @item G
 @itemx GiB
 @cindex gibibyte, definition of
-gibibyte: @math{2^30 = 1,073,741,824}.
+gibibyte: @math{2^{30} = 1,073,741,824}.
 @item TB
 @cindex terabyte, definition of
-terabyte:  @math{10^12 = 1,000,000,000,000}.
+terabyte:  @math{10^{12} = 1,000,000,000,000}.
 @item T
 @itemx TiB
 @cindex tebibyte, definition of
-tebibyte: @math{2^40 = 1,099,511,627,776}.
+tebibyte: @math{2^{40} = 1,099,511,627,776}.
 @item PB
 @cindex petabyte, definition of
-petabyte: @math{10^15 = 1,000,000,000,000,000}.
+petabyte: @math{10^{15} = 1,000,000,000,000,000}.
 @item P
 @itemx PiB
 @cindex pebibyte, definition of
-pebibyte: @math{2^50 = 1,125,899,906,842,624}.
+pebibyte: @math{2^{50} = 1,125,899,906,842,624}.
 @item EB
 @cindex exabyte, definition of
-exabyte: @math{10^18 = 1,000,000,000,000,000,000}.
+exabyte: @math{10^{18} = 1,000,000,000,000,000,000}.
 @item E
 @itemx EiB
 @cindex exbibyte, definition of
-exbibyte: @math{2^60 = 1,152,921,504,606,846,976}.
+exbibyte: @math{2^{60} = 1,152,921,504,606,846,976}.
 @item ZB
 @cindex zettabyte, definition of
-zettabyte: @math{10^21 = 1,000,000,000,000,000,000,000}
+zettabyte: @math{10^{21} = 1,000,000,000,000,000,000,000}
 @item Z
 @itemx ZiB
-@math{2^70 = 1,180,591,620,717,411,303,424}.
-(@samp{Zi} is a GNU extension to IEC 60027-2.)
+zebibyte: @math{2^{70} = 1,180,591,620,717,411,303,424}.
 @item YB
 @cindex yottabyte, definition of
-yottabyte: @math{10^24 = 1,000,000,000,000,000,000,000,000}.
+yottabyte: @math{10^{24} = 1,000,000,000,000,000,000,000,000}.
 @item Y
 @itemx YiB
-@math{2^80 = 1,208,925,819,614,629,174,706,176}.
-(@samp{Yi} is a GNU extension to IEC 60027-2.)
+yobibyte: @math{2^{80} = 1,208,925,819,614,629,174,706,176}.
+@item RB
+@cindex ronnabyte, definition of
+ronnabyte: @math{10^{27} = 1,000,000,000,000,000,000,000,000,000}.
+@item R
+@itemx RiB
+robibyte: @math{2^{90} = 1,237,940,039,285,380,274,899,124,224}.
+@item QB
+@cindex quettabyte, definition of
+quettabyte: @math{10^{30} = 1,000,000,000,000,000,000,000,000,000,000}.
+@item Q
+@itemx QiB
+quebibyte: @math{2^{100} = 1,267,650,600,228,229,401,496,703,205,376}.
 @end table
 
 @node Invoking diff
@@ -3941,7 +4016,7 @@ found.  @xref{Comparing Directories}.
 
 @item -s
 @itemx --report-identical-files
-Report when two files are the same.  @xref{Comparing Directories}.
+Also report when two files are the same.  @xref{Comparing Directories}.
 
 @item -S @var{file}
 @itemx --starting-file=@var{file}
@@ -4595,39 +4670,37 @@ handle some of the existing formats.  These shortcomings motivate the
 following suggested projects.
 
 @menu
-* Internationalization:: Handling multibyte and varying-width characters.
+* Internationalization:: Handling multi-byte and varying-width characters.
 * Changing Structure::   Handling changes to the directory structure.
 * Special Files::        Handling symbolic links, device special files, etc.
 * Unusual File Names::   Handling file names that contain unusual characters.
-* Timestamp Order::     Outputting diffs in timestamp order.
+* Timestamp Order::      Outputting diffs in timestamp order.
 * Ignoring Changes::     Ignoring certain changes while showing others.
 * Speedups::             Improving performance.
 @end menu
 
 @node Internationalization
-@subsection Handling Multibyte and Varying-Width Characters
-@cindex multibyte characters
+@subsection Handling Multi-byte and Varying-Width Characters
+@cindex multi-byte characters
 @cindex varying-width characters
 
 @command{diff}, @command{diff3} and @command{sdiff} treat each line of
-input as a string of unibyte characters.  This can mishandle multibyte
-characters in some cases.  For example, when asked to ignore spaces,
-@command{diff} does not properly ignore a multibyte space character.
+input as a string of characters and encoding errors, where an
+@dfn{encoding error} is an input byte that is not part of any
+character.  Single-byte and multi-byte characters are supported, along
+with common character encoding systems like UTF-8.  The operating
+system's locale specifies the character encoding, and can be specified
+with the @env{LC_ALL} environment variable.  You can find which
+locales are supported on your system by running the shell command
+@samp{locale -a}.
 
-Also, @command{diff} currently assumes that each byte is one column
-wide, and this assumption is incorrect in some locales, e.g., locales
-that use UTF-8 encoding.  This causes problems with the @option{-y} or
-@option{--side-by-side} option of @command{diff}.
+When counting columns for options like @option{--expand-tabs} (@option{-t}),
+@command{diff} consults the locale for the column width of each character,
+and assumes that each encoding error occupies a single column.
 
-These problems need to be fixed without unduly affecting the
-performance of the utilities in unibyte environments.
-
-The IBM GNU/Linux Technology Center Internationalization Team has
-proposed
-@uref{http://oss.software.ibm.com/developer/opensource/linux/patches/i18n/diffutils-2.7.2-i18n-0.1.patch.gz,patches
-to support internationalized @command{diff}}.
-Unfortunately, these patches are incomplete and are to an older
-version of @command{diff}, so more work needs to be done in this area.
+When ignoring case for @option{--ignore-case} (@option{-i}),
+@command{diff} downcases each character before comparing it,
+regardless of whether it is multi-byte.  @xref{Case Folding}.
 
 @node Changing Structure
 @subsection Handling Changes to the Directory Structure
@@ -4677,13 +4750,14 @@ and @command{patch} should be extended to understand these extensions.
 @subsection File Names that Contain Unusual Characters
 @cindex file names with unusual characters
 
-When a file name contains an unusual character like a newline or
-white space, @samp{diff -r} generates a patch that @command{patch} cannot
-parse.  The problem is with format of @command{diff} output, not just with
-@command{patch}, because with odd enough file names one can cause
-@command{diff} to generate a patch that is syntactically correct but
-patches the wrong files.  The format of @command{diff} output should be
-extended to handle all possible file names.
+Since diffutils-3.3, file names have been encoded to eliminate the ambiguity
+of unusual characters like newline and TAB.
+However, since any file name containing a newline may easily cause trouble,
+in so many contexts (not just diff and patch), a future version of
+diff may well reject attempts to operate on such names.
+@c FIXME-in-2026 consider rejecting NL-afflicted file names
+@c keep an eye on POSIX, e.g.,
+@c https://pubs.opengroup.org/onlinepubs/9799919799/utilities/diff.html
 
 @node Timestamp Order
 @subsection Outputting Diffs in Timestamp Order

doc/fdl.texi

@@ -0,0 +1,506 @@
+@c The GNU Free Documentation License.
+@center Version 1.3, 3 November 2008
+
+@c This file is intended to be included within another document,
+@c hence no sectioning command or @node.
+
+@display
+Copyright @copyright{} 2000--2002, 2007--2008, 2022--2026 Free Software
+Foundation, Inc.
+@uref{https://fsf.org/}
+
+Everyone is permitted to copy and distribute verbatim copies
+of this license document, but changing it is not allowed.
+@end display
+
+@enumerate 0
+@item
+PREAMBLE
+
+The purpose of this License is to make a manual, textbook, or other
+functional and useful document @dfn{free} in the sense of freedom: to
+assure everyone the effective freedom to copy and redistribute it,
+with or without modifying it, either commercially or noncommercially.
+Secondarily, this License preserves for the author and publisher a way
+to get credit for their work, while not being considered responsible
+for modifications made by others.
+
+This License is a kind of ``copyleft'', which means that derivative
+works of the document must themselves be free in the same sense.  It
+complements the GNU General Public License, which is a copyleft
+license designed for free software.
+
+We have designed this License in order to use it for manuals for free
+software, because free software needs free documentation: a free
+program should come with manuals providing the same freedoms that the
+software does.  But this License is not limited to software manuals;
+it can be used for any textual work, regardless of subject matter or
+whether it is published as a printed book.  We recommend this License
+principally for works whose purpose is instruction or reference.
+
+@item
+APPLICABILITY AND DEFINITIONS
+
+This License applies to any manual or other work, in any medium, that
+contains a notice placed by the copyright holder saying it can be
+distributed under the terms of this License.  Such a notice grants a
+world-wide, royalty-free license, unlimited in duration, to use that
+work under the conditions stated herein.  The ``Document'', below,
+refers to any such manual or work.  Any member of the public is a
+licensee, and is addressed as ``you''.  You accept the license if you
+copy, modify or distribute the work in a way requiring permission
+under copyright law.
+
+A ``Modified Version'' of the Document means any work containing the
+Document or a portion of it, either copied verbatim, or with
+modifications and/or translated into another language.
+
+A ``Secondary Section'' is a named appendix or a front-matter section
+of the Document that deals exclusively with the relationship of the
+publishers or authors of the Document to the Document's overall
+subject (or to related matters) and contains nothing that could fall
+directly within that overall subject.  (Thus, if the Document is in
+part a textbook of mathematics, a Secondary Section may not explain
+any mathematics.)  The relationship could be a matter of historical
+connection with the subject or with related matters, or of legal,
+commercial, philosophical, ethical or political position regarding
+them.
+
+The ``Invariant Sections'' are certain Secondary Sections whose titles
+are designated, as being those of Invariant Sections, in the notice
+that says that the Document is released under this License.  If a
+section does not fit the above definition of Secondary then it is not
+allowed to be designated as Invariant.  The Document may contain zero
+Invariant Sections.  If the Document does not identify any Invariant
+Sections then there are none.
+
+The ``Cover Texts'' are certain short passages of text that are listed,
+as Front-Cover Texts or Back-Cover Texts, in the notice that says that
+the Document is released under this License.  A Front-Cover Text may
+be at most 5 words, and a Back-Cover Text may be at most 25 words.
+
+A ``Transparent'' copy of the Document means a machine-readable copy,
+represented in a format whose specification is available to the
+general public, that is suitable for revising the document
+straightforwardly with generic text editors or (for images composed of
+pixels) generic paint programs or (for drawings) some widely available
+drawing editor, and that is suitable for input to text formatters or
+for automatic translation to a variety of formats suitable for input
+to text formatters.  A copy made in an otherwise Transparent file
+format whose markup, or absence of markup, has been arranged to thwart
+or discourage subsequent modification by readers is not Transparent.
+An image format is not Transparent if used for any substantial amount
+of text.  A copy that is not ``Transparent'' is called ``Opaque''.
+
+Examples of suitable formats for Transparent copies include plain
+ASCII without markup, Texinfo input format, La@TeX{} input
+format, SGML or XML using a publicly available
+DTD, and standard-conforming simple HTML,
+PostScript or PDF designed for human modification.  Examples
+of transparent image formats include PNG, XCF and
+JPG@.  Opaque formats include proprietary formats that can be
+read and edited only by proprietary word processors, SGML or
+XML for which the DTD and/or processing tools are
+not generally available, and the machine-generated HTML,
+PostScript or PDF produced by some word processors for
+output purposes only.
+
+The ``Title Page'' means, for a printed book, the title page itself,
+plus such following pages as are needed to hold, legibly, the material
+this License requires to appear in the title page.  For works in
+formats which do not have any title page as such, ``Title Page'' means
+the text near the most prominent appearance of the work's title,
+preceding the beginning of the body of the text.
+
+The ``publisher'' means any person or entity that distributes copies
+of the Document to the public.
+
+A section ``Entitled XYZ'' means a named subunit of the Document whose
+title either is precisely XYZ or contains XYZ in parentheses following
+text that translates XYZ in another language.  (Here XYZ stands for a
+specific section name mentioned below, such as ``Acknowledgements'',
+``Dedications'', ``Endorsements'', or ``History''.)  To ``Preserve the Title''
+of such a section when you modify the Document means that it remains a
+section ``Entitled XYZ'' according to this definition.
+
+The Document may include Warranty Disclaimers next to the notice which
+states that this License applies to the Document.  These Warranty
+Disclaimers are considered to be included by reference in this
+License, but only as regards disclaiming warranties: any other
+implication that these Warranty Disclaimers may have is void and has
+no effect on the meaning of this License.
+
+@item
+VERBATIM COPYING
+
+You may copy and distribute the Document in any medium, either
+commercially or noncommercially, provided that this License, the
+copyright notices, and the license notice saying this License applies
+to the Document are reproduced in all copies, and that you add no other
+conditions whatsoever to those of this License.  You may not use
+technical measures to obstruct or control the reading or further
+copying of the copies you make or distribute.  However, you may accept
+compensation in exchange for copies.  If you distribute a large enough
+number of copies you must also follow the conditions in section 3.
+
+You may also lend copies, under the same conditions stated above, and
+you may publicly display copies.
+
+@item
+COPYING IN QUANTITY
+
+If you publish printed copies (or copies in media that commonly have
+printed covers) of the Document, numbering more than 100, and the
+Document's license notice requires Cover Texts, you must enclose the
+copies in covers that carry, clearly and legibly, all these Cover
+Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
+the back cover.  Both covers must also clearly and legibly identify
+you as the publisher of these copies.  The front cover must present
+the full title with all words of the title equally prominent and
+visible.  You may add other material on the covers in addition.
+Copying with changes limited to the covers, as long as they preserve
+the title of the Document and satisfy these conditions, can be treated
+as verbatim copying in other respects.
+
+If the required texts for either cover are too voluminous to fit
+legibly, you should put the first ones listed (as many as fit
+reasonably) on the actual cover, and continue the rest onto adjacent
+pages.
+
+If you publish or distribute Opaque copies of the Document numbering
+more than 100, you must either include a machine-readable Transparent
+copy along with each Opaque copy, or state in or with each Opaque copy
+a computer-network location from which the general network-using
+public has access to download using public-standard network protocols
+a complete Transparent copy of the Document, free of added material.
+If you use the latter option, you must take reasonably prudent steps,
+when you begin distribution of Opaque copies in quantity, to ensure
+that this Transparent copy will remain thus accessible at the stated
+location until at least one year after the last time you distribute an
+Opaque copy (directly or through your agents or retailers) of that
+edition to the public.
+
+It is requested, but not required, that you contact the authors of the
+Document well before redistributing any large number of copies, to give
+them a chance to provide you with an updated version of the Document.
+
+@item
+MODIFICATIONS
+
+You may copy and distribute a Modified Version of the Document under
+the conditions of sections 2 and 3 above, provided that you release
+the Modified Version under precisely this License, with the Modified
+Version filling the role of the Document, thus licensing distribution
+and modification of the Modified Version to whoever possesses a copy
+of it.  In addition, you must do these things in the Modified Version:
+
+@enumerate A
+@item
+Use in the Title Page (and on the covers, if any) a title distinct
+from that of the Document, and from those of previous versions
+(which should, if there were any, be listed in the History section
+of the Document).  You may use the same title as a previous version
+if the original publisher of that version gives permission.
+
+@item
+List on the Title Page, as authors, one or more persons or entities
+responsible for authorship of the modifications in the Modified
+Version, together with at least five of the principal authors of the
+Document (all of its principal authors, if it has fewer than five),
+unless they release you from this requirement.
+
+@item
+State on the Title page the name of the publisher of the
+Modified Version, as the publisher.
+
+@item
+Preserve all the copyright notices of the Document.
+
+@item
+Add an appropriate copyright notice for your modifications
+adjacent to the other copyright notices.
+
+@item
+Include, immediately after the copyright notices, a license notice
+giving the public permission to use the Modified Version under the
+terms of this License, in the form shown in the Addendum below.
+
+@item
+Preserve in that license notice the full lists of Invariant Sections
+and required Cover Texts given in the Document's license notice.
+
+@item
+Include an unaltered copy of this License.
+
+@item
+Preserve the section Entitled ``History'', Preserve its Title, and add
+to it an item stating at least the title, year, new authors, and
+publisher of the Modified Version as given on the Title Page.  If
+there is no section Entitled ``History'' in the Document, create one
+stating the title, year, authors, and publisher of the Document as
+given on its Title Page, then add an item describing the Modified
+Version as stated in the previous sentence.
+
+@item
+Preserve the network location, if any, given in the Document for
+public access to a Transparent copy of the Document, and likewise
+the network locations given in the Document for previous versions
+it was based on.  These may be placed in the ``History'' section.
+You may omit a network location for a work that was published at
+least four years before the Document itself, or if the original
+publisher of the version it refers to gives permission.
+
+@item
+For any section Entitled ``Acknowledgements'' or ``Dedications'', Preserve
+the Title of the section, and preserve in the section all the
+substance and tone of each of the contributor acknowledgements and/or
+dedications given therein.
+
+@item
+Preserve all the Invariant Sections of the Document,
+unaltered in their text and in their titles.  Section numbers
+or the equivalent are not considered part of the section titles.
+
+@item
+Delete any section Entitled ``Endorsements''.  Such a section
+may not be included in the Modified Version.
+
+@item
+Do not retitle any existing section to be Entitled ``Endorsements'' or
+to conflict in title with any Invariant Section.
+
+@item
+Preserve any Warranty Disclaimers.
+@end enumerate
+
+If the Modified Version includes new front-matter sections or
+appendices that qualify as Secondary Sections and contain no material
+copied from the Document, you may at your option designate some or all
+of these sections as invariant.  To do this, add their titles to the
+list of Invariant Sections in the Modified Version's license notice.
+These titles must be distinct from any other section titles.
+
+You may add a section Entitled ``Endorsements'', provided it contains
+nothing but endorsements of your Modified Version by various
+parties---for example, statements of peer review or that the text has
+been approved by an organization as the authoritative definition of a
+standard.
+
+You may add a passage of up to five words as a Front-Cover Text, and a
+passage of up to 25 words as a Back-Cover Text, to the end of the list
+of Cover Texts in the Modified Version.  Only one passage of
+Front-Cover Text and one of Back-Cover Text may be added by (or
+through arrangements made by) any one entity.  If the Document already
+includes a cover text for the same cover, previously added by you or
+by arrangement made by the same entity you are acting on behalf of,
+you may not add another; but you may replace the old one, on explicit
+permission from the previous publisher that added the old one.
+
+The author(s) and publisher(s) of the Document do not by this License
+give permission to use their names for publicity for or to assert or
+imply endorsement of any Modified Version.
+
+@item
+COMBINING DOCUMENTS
+
+You may combine the Document with other documents released under this
+License, under the terms defined in section 4 above for modified
+versions, provided that you include in the combination all of the
+Invariant Sections of all of the original documents, unmodified, and
+list them all as Invariant Sections of your combined work in its
+license notice, and that you preserve all their Warranty Disclaimers.
+
+The combined work need only contain one copy of this License, and
+multiple identical Invariant Sections may be replaced with a single
+copy.  If there are multiple Invariant Sections with the same name but
+different contents, make the title of each such section unique by
+adding at the end of it, in parentheses, the name of the original
+author or publisher of that section if known, or else a unique number.
+Make the same adjustment to the section titles in the list of
+Invariant Sections in the license notice of the combined work.
+
+In the combination, you must combine any sections Entitled ``History''
+in the various original documents, forming one section Entitled
+``History''; likewise combine any sections Entitled ``Acknowledgements'',
+and any sections Entitled ``Dedications''.  You must delete all
+sections Entitled ``Endorsements.''
+
+@item
+COLLECTIONS OF DOCUMENTS
+
+You may make a collection consisting of the Document and other documents
+released under this License, and replace the individual copies of this
+License in the various documents with a single copy that is included in
+the collection, provided that you follow the rules of this License for
+verbatim copying of each of the documents in all other respects.
+
+You may extract a single document from such a collection, and distribute
+it individually under this License, provided you insert a copy of this
+License into the extracted document, and follow this License in all
+other respects regarding verbatim copying of that document.
+
+@item
+AGGREGATION WITH INDEPENDENT WORKS
+
+A compilation of the Document or its derivatives with other separate
+and independent documents or works, in or on a volume of a storage or
+distribution medium, is called an ``aggregate'' if the copyright
+resulting from the compilation is not used to limit the legal rights
+of the compilation's users beyond what the individual works permit.
+When the Document is included in an aggregate, this License does not
+apply to the other works in the aggregate which are not themselves
+derivative works of the Document.
+
+If the Cover Text requirement of section 3 is applicable to these
+copies of the Document, then if the Document is less than one half of
+the entire aggregate, the Document's Cover Texts may be placed on
+covers that bracket the Document within the aggregate, or the
+electronic equivalent of covers if the Document is in electronic form.
+Otherwise they must appear on printed covers that bracket the whole
+aggregate.
+
+@item
+TRANSLATION
+
+Translation is considered a kind of modification, so you may
+distribute translations of the Document under the terms of section 4.
+Replacing Invariant Sections with translations requires special
+permission from their copyright holders, but you may include
+translations of some or all Invariant Sections in addition to the
+original versions of these Invariant Sections.  You may include a
+translation of this License, and all the license notices in the
+Document, and any Warranty Disclaimers, provided that you also include
+the original English version of this License and the original versions
+of those notices and disclaimers.  In case of a disagreement between
+the translation and the original version of this License or a notice
+or disclaimer, the original version will prevail.
+
+If a section in the Document is Entitled ``Acknowledgements'',
+``Dedications'', or ``History'', the requirement (section 4) to Preserve
+its Title (section 1) will typically require changing the actual
+title.
+
+@item
+TERMINATION
+
+You may not copy, modify, sublicense, or distribute the Document
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense, or distribute it is void, and
+will automatically terminate your rights under this License.
+
+However, if you cease all violation of this License, then your license
+from a particular copyright holder is reinstated (a) provisionally,
+unless and until the copyright holder explicitly and finally
+terminates your license, and (b) permanently, if the copyright holder
+fails to notify you of the violation by some reasonable means prior to
+60 days after the cessation.
+
+Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, receipt of a copy of some or all of the same material does
+not give you any rights to use it.
+
+@item
+FUTURE REVISIONS OF THIS LICENSE
+
+The Free Software Foundation may publish new, revised versions
+of the GNU Free Documentation License from time to time.  Such new
+versions will be similar in spirit to the present version, but may
+differ in detail to address new problems or concerns.  See
+@uref{https://www.gnu.org/licenses/}.
+
+Each version of the License is given a distinguishing version number.
+If the Document specifies that a particular numbered version of this
+License ``or any later version'' applies to it, you have the option of
+following the terms and conditions either of that specified version or
+of any later version that has been published (not as a draft) by the
+Free Software Foundation.  If the Document does not specify a version
+number of this License, you may choose any version ever published (not
+as a draft) by the Free Software Foundation.  If the Document
+specifies that a proxy can decide which future versions of this
+License can be used, that proxy's public statement of acceptance of a
+version permanently authorizes you to choose that version for the
+Document.
+
+@item
+RELICENSING
+
+``Massive Multiauthor Collaboration Site'' (or ``MMC Site'') means any
+World Wide Web server that publishes copyrightable works and also
+provides prominent facilities for anybody to edit those works.  A
+public wiki that anybody can edit is an example of such a server.  A
+``Massive Multiauthor Collaboration'' (or ``MMC'') contained in the
+site means any set of copyrightable works thus published on the MMC
+site.
+
+``CC-BY-SA'' means the Creative Commons Attribution-Share Alike 3.0
+license published by Creative Commons Corporation, a not-for-profit
+corporation with a principal place of business in San Francisco,
+California, as well as future copyleft versions of that license
+published by that same organization.
+
+``Incorporate'' means to publish or republish a Document, in whole or
+in part, as part of another Document.
+
+An MMC is ``eligible for relicensing'' if it is licensed under this
+License, and if all works that were first published under this License
+somewhere other than this MMC, and subsequently incorporated in whole
+or in part into the MMC, (1) had no cover texts or invariant sections,
+and (2) were thus incorporated prior to November 1, 2008.
+
+The operator of an MMC Site may republish an MMC contained in the site
+under CC-BY-SA on the same site at any time before August 1, 2009,
+provided the MMC is eligible for relicensing.
+
+@end enumerate
+
+@page
+@heading ADDENDUM: How to use this License for your documents
+
+To use this License in a document you have written, include a copy of
+the License in the document and put the following copyright and
+license notices just after the title page:
+
+@smallexample
+@group
+  Copyright (C)  @var{year}  @var{your name}.
+  Permission is granted to copy, distribute and/or modify this document
+  under the terms of the GNU Free Documentation License, Version 1.3
+  or any later version published by the Free Software Foundation;
+  with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
+  Texts.  A copy of the license is included in the section entitled ``GNU
+  Free Documentation License''.
+@end group
+@end smallexample
+
+If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
+replace the ``with@dots{}Texts.''@: line with this:
+
+@smallexample
+@group
+    with the Invariant Sections being @var{list their titles}, with
+    the Front-Cover Texts being @var{list}, and with the Back-Cover Texts
+    being @var{list}.
+@end group
+@end smallexample
+
+If you have Invariant Sections without Cover Texts, or some other
+combination of the three, merge those two alternatives to suit the
+situation.
+
+If your document contains nontrivial examples of program code, we
+recommend releasing these examples in parallel under your choice of
+free software license, such as the GNU General Public License,
+to permit their use in free software.
+
+@c Local Variables:
+@c ispell-local-pdict: "ispell-dict"
+@c End:

exgettext

@@ -1,7 +1,7 @@
 #! /bin/sh
 # Wrapper around gettext for programs using the msgid convention.
 
-# Copyright (C) 1998, 2001, 2004, 2009-2013, 2015-2017 Free Software
+# Copyright (C) 1998, 2001, 2004, 2009-2013, 2015-2026 Free Software
 # Foundation, Inc.
 
 # Written by Paul Eggert <eggert@twinsun.com>.
@@ -97,7 +97,7 @@ generate_emsgids='
       line = substr(line, percent_index + 2)
       bracket_index = index(line, "}")
       if (bracket_index == 0) {
-	continue
+        continue
       }
       msgid = substr(line, 1, bracket_index - 1)
       if (index(msgid, "%") != 0) {

gl/lib/sh-quote.c.diff

@@ -1,15 +0,0 @@
-diff --git a/lib/sh-quote.c b/lib/sh-quote.c
-index e349f40..788989a 100644
---- a/lib/sh-quote.c
-+++ b/lib/sh-quote.c
-@@ -30,7 +30,7 @@ static struct quoting_options *sh_quoting_options;
-
- /* Initializes the sh_quoting_options variable.  */
- static void
--init_sh_quoting_options ()
-+init_sh_quoting_options (void)
- {
-   sh_quoting_options = clone_quoting_options (NULL);
-   set_quoting_style (sh_quoting_options, shell_quoting_style);
-
---

gnulib

@@ -1 +1 @@
-Subproject commit 809f19dba7cf85e4ea16246786cf19269590abf8
+Subproject commit 4f6ac2c3c689cd7312b5f9da97791b14bbc2ee53

gnulib-tests/Makefile.am

@@ -1 +1,3 @@
+AM_CFLAGS =
+
 include gnulib.mk

lib/Makefile.am

@@ -1,6 +1,6 @@
 # Automakefile for GNU Diffutils library.
 
-# Copyright (C) 2001-2002, 2004, 2006, 2009-2013, 2015-2017 Free Software
+# Copyright (C) 2001-2002, 2004, 2006, 2009-2013, 2015-2026 Free Software
 # Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
@@ -29,7 +29,7 @@ noinst_HEADERS =
 
 include gnulib.mk
 
-noinst_HEADERS += cmpbuf.h prepargs.h
-libdiffutils_a_SOURCES += cmpbuf.c prepargs.c
+noinst_HEADERS += cmpbuf.h diagnose.h
+libdiffutils_a_SOURCES += cmpbuf.c diagnose.c
 
 AM_CFLAGS += $(GNULIB_WARN_CFLAGS) $(WERROR_CFLAGS)

lib/cmpbuf.c

@@ -1,6 +1,6 @@
 /* Buffer primitives for comparison operations.
 
-   Copyright (C) 1993, 1995, 1998, 2001-2002, 2006, 2009-2013, 2015-2017 Free
+   Copyright (C) 1993, 1995, 1998, 2001-2002, 2006, 2009-2013, 2015-2026 Free
    Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
@@ -18,64 +18,77 @@
 
 #include <config.h>
 
+#include "cmpbuf.h"
+
 #include <errno.h>
 #include <limits.h>
 #include <signal.h>
+#include <stdckdint.h>
 #include <unistd.h>
 #include <stdint.h>
 #include <inttypes.h>
 #include <sys/types.h>
-#include "cmpbuf.h"
+
 #include "intprops.h"
-
-#ifndef SSIZE_MAX
-# define SSIZE_MAX TYPE_MAXIMUM (ssize_t)
-#endif
-
-#undef MIN
-#define MIN(a, b) ((a) <= (b) ? (a) : (b))
+#include "minmax.h"
 
 /* Read NBYTES bytes from descriptor FD into BUF.
-   NBYTES must not be SIZE_MAX.
    Return the number of characters successfully read.
-   On error, return SIZE_MAX, setting errno.
+   On error, return -1, setting errno.
    The number returned is always NBYTES unless end-of-file or error.  */
 
-size_t
-block_read (int fd, char *buf, size_t nbytes)
+ptrdiff_t
+block_read (int fd, char *buf, idx_t nbytes)
 {
   char *bp = buf;
   char const *buflim = buf + nbytes;
-  size_t readlim = MIN (SSIZE_MAX, SIZE_MAX);
+  idx_t readlim = MIN (IDX_MAX, MIN (SSIZE_MAX, SIZE_MAX));
 
   do
     {
-      size_t bytes_remaining = buflim - bp;
-      size_t bytes_to_read = MIN (bytes_remaining, readlim);
+      idx_t bytes_remaining = buflim - bp;
+      idx_t bytes_to_read = MIN (bytes_remaining, readlim);
       ssize_t nread = read (fd, bp, bytes_to_read);
       if (nread <= 0)
-	{
-	  if (nread == 0)
-	    break;
+        {
+          if (nread == 0)
+            break;
 
-	  /* Accommodate Tru64 5.1, which can't read more than INT_MAX
-	     bytes at a time.  They call that a 64-bit OS?  */
-	  if (errno == EINVAL && INT_MAX < bytes_to_read)
+          /* Accommodate FreeBSD 13, which can't read more than INT_MAX bytes
+	     when debug.iosize_max_clamp is nonzero.  Prefer a power
+	     of two for the read limit in this case.
+
+	     Also, work around a bug in Linux kernel 6.3.8 tmpfs,
+	     which fails if the current offset + bytes_to_read exceeds
+	     TYPE_MAXIMUM (off_t), even if EOF occurs before then.  */
+	  if (errno == EINVAL)
 	    {
-	      readlim = INT_MAX;
-	      continue;
+	      if (bytes_to_read <= 1)
+		{
+		  if (lseek (fd, 0, SEEK_CUR) == TYPE_MAXIMUM (off_t))
+		    {
+		      nread = 0;
+		      break;
+		    }
+		  errno = EINVAL;
+		}
+	      else
+		{
+		  readlim = MIN (bytes_to_read >> 1, INT_MAX / 2 + 1);
+		  continue;
+		}
 	    }
 
-	  /* This is needed for programs that have signal handlers on
-	     older hosts without SA_RESTART.  It also accommodates
-	     ancient AIX hosts that set errno to EINTR after uncaught
-	     SIGCONT.  See <news:1r77ojINN85n@ftp.UU.NET>
-	     (1993-04-22).  */
-	  if (! SA_RESTART && errno == EINTR)
-	    continue;
+          /* This is needed for programs that have signal handlers on
+             older hosts without SA_RESTART.  It also accommodates
+             ancient AIX hosts that set errno to EINTR after uncaught
+             SIGCONT.  See <news:1r77ojINN85n@ftp.UU.NET>
+             (1993-04-22).  */
+          if (! SA_RESTART && errno == EINTR)
+            continue;
 
-	  return SIZE_MAX;
-	}
+          return -1;
+        }
       bp += nread;
     }
   while (bp < buflim);
@@ -87,23 +100,23 @@ block_read (int fd, char *buf, size_t nbytes)
    either A or B is zero, or if the multiple is greater than LCM_MAX,
    return a reasonable buffer size.  */
 
-size_t
-buffer_lcm (size_t a, size_t b, size_t lcm_max)
+idx_t
+buffer_lcm (idx_t a, idx_t b, idx_t lcm_max)
 {
-  size_t lcm, m, n, q, r;
-
-  /* Yield reasonable values if buffer sizes are zero.  */
-  if (!a)
-    return b ? b : 8 * 1024;
-  if (!b)
+  /* Yield reasonable values if buffer sizes are zero.  Although A and
+     B must be nonnegative, GCC generates better code from (A <= 0)
+     than from (A != 0).  */
+  if (a <= 0)
+    return b <= 0 ? 8 * 1024 : b;
+  if (b <= 0)
     return a;
 
   /* n = gcd (a, b) */
+  idx_t m, n, r;
   for (m = a, n = b;  (r = m % n) != 0;  m = n, n = r)
     continue;
 
   /* Yield a if there is an overflow.  */
-  q = a / n;
-  lcm = q * b;
-  return lcm <= lcm_max && lcm / b == q ? lcm : a;
+  idx_t q = a / n, lcm;
+  return !ckd_mul (&lcm, b, q) && lcm <= lcm_max ? lcm : a;
 }

lib/cmpbuf.h

@@ -1,6 +1,6 @@
 /* Buffer primitives for comparison operations.
 
-   Copyright (C) 2002, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+   Copyright (C) 2002, 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -15,5 +15,7 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
-size_t block_read (int, char *, size_t);
-size_t buffer_lcm (size_t, size_t, size_t) _GL_ATTRIBUTE_CONST;
+#include "idx.h"
+#include <stddef.h>
+ptrdiff_t block_read (int, char *, idx_t);
+idx_t buffer_lcm (idx_t, idx_t, idx_t) _GL_ATTRIBUTE_CONST;

lib/diagnose.c

@@ -0,0 +1,52 @@
+/* Diagnostics for GNU diffutils
+
+   Copyright 2023-2026 Free Software Foundation, Inc.
+
+   This file is part of GNU DIFF.
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#include <config.h>
+
+#include "diagnose.h"
+
+#include "gettext.h"
+#include "progname.h"
+#include "quotearg.h"
+
+#include <error.h>
+
+/* In slot N return NAME, quoted for the shell if NAME has unusual characters.
+   This is for messages that historically did not quote names,
+   where we want to keep the message unchanged for compatibility reasons
+   unless the unusual characters might cause confusion.
+   Also, POSIX requires that names be unquoted in some cases,
+   and we want to respect that for POSIX portable filenames.  */
+char *
+squote (int n, char const *name)
+{
+  return quotearg_n_style (n, shell_escape_quoting_style, name);
+}
+
+/* Issue help for the program.  If REASON_MSGID, first issue a
+   diagnostic with that reason and with optional operand OPERAND.
+   Suggest --help regardless.  */
+void
+try_help (char const *reason_msgid, char const *operand)
+{
+  if (reason_msgid)
+    error (0, 0, gettext (reason_msgid), operand);
+  error (EXIT_TROUBLE, 0,
+	 gettext ("Try '%s --help' for more information."), program_name);
+}

lib/diagnose.h

@@ -0,0 +1,3 @@
+enum { EXIT_TROUBLE = 2 };
+char *squote (int, char const *);
+_Noreturn void try_help (char const *, char const *);

lib/prepargs.c

@@ -1,91 +0,0 @@
-/* Parse arguments from a string and prepend them to an argv.
-
-   Copyright (C) 1999-2002, 2006, 2009-2013, 2015-2017 Free Software
-   Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation, either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* Written by Paul Eggert <eggert@twinsun.com>.  */
-
-#include <config.h>
-
-#include "prepargs.h"
-#include <string.h>
-#include <sys/types.h>
-#include <xalloc.h>
-
-#include <ctype.h>
-
-/* IN_CTYPE_DOMAIN (C) is nonzero if the unsigned char C can safely be given
-   as an argument to <ctype.h> macros like "isspace".  */
-#ifdef STDC_HEADERS
-# define IN_CTYPE_DOMAIN(c) 1
-#else
-# define IN_CTYPE_DOMAIN(c) ((c) <= 0177)
-#endif
-
-#define ISSPACE(c) (IN_CTYPE_DOMAIN (c) && isspace (c))
-
-/* Find the white-space-separated options specified by OPTIONS, and
-   using BUF to store copies of these options, set ARGV[0], ARGV[1],
-   etc. to the option copies.  Return the number N of options found.
-   Do not set ARGV[N].  If ARGV is zero, do not store ARGV[0] etc.
-   Backslash can be used to escape whitespace (and backslashes).  */
-static int
-prepend_args (char const *options, char *buf, char **argv)
-{
-  char const *o = options;
-  char *b = buf;
-  int n = 0;
-
-  for (;;)
-    {
-      while (ISSPACE ((unsigned char) *o))
-	o++;
-      if (!*o)
-	return n;
-      if (argv)
-	argv[n] = b;
-      n++;
-
-      do
-	if ((*b++ = *o++) == '\\' && *o)
-	  b[-1] = *o++;
-      while (*o && ! ISSPACE ((unsigned char) *o));
-
-      *b++ = '\0';
-    }
-}
-
-/* Prepend the whitespace-separated options in OPTIONS to the argument
-   vector of a main program with argument count *PARGC and argument
-   vector *PARGV.  */
-void
-prepend_default_options (char const *options, int *pargc, char ***pargv)
-{
-  if (options)
-    {
-      char *buf = xmalloc (strlen (options) + 1);
-      int prepended = prepend_args (options, buf, (char **) 0);
-      int argc = *pargc;
-      char * const *argv = *pargv;
-      char **pp = xmalloc ((prepended + argc + 1) * sizeof *pp);
-      *pargc = prepended + argc;
-      *pargv = pp;
-      *pp++ = *argv++;
-      pp += prepend_args (options, buf, pp);
-      while ((*pp++ = *argv++))
-	continue;
-    }
-}

lib/prepargs.h

@@ -1,3 +0,0 @@
-/* Parse arguments from a string and prepend them to an argv.  */
-
-void prepend_default_options (char const *, int *, char ***);

man/Makefile.am

@@ -1,6 +1,6 @@
 # Automakefile for GNU diffutils man pages
 
-# Copyright (C) 2002, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2002, 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -16,7 +16,7 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 dist_man1_MANS = cmp.1 diff.1 diff3.1 sdiff.1
-EXTRA_DIST = $(dist_man1_MANS:%.1=%.x) help2man
+EXTRA_DIST = $(dist_man1_MANS:%.1=%.x)
 MAINTAINERCLEANFILES = $(dist_man1_MANS)
 
 S = $(top_srcdir)/src
@@ -29,11 +29,13 @@ sdiff.1: $S/sdiff.c sdiff.x
 # to ensure help2man invokes them via the use of PATH below.
 bin_dir = ../src
 
-# Depend on the former to get version number changes.
-$(dist_man1_MANS): $(SRC_VERSION_C) help2man
+# Depend on $(SRC_VERSION_C) to get version number changes.
+$(dist_man1_MANS): $(SRC_VERSION_C)
 	$(AM_V_GEN)base=`expr $@ : '\(.*\).1'`				\
 	  && test -x $(bin_dir)/$$base					\
-	  && (echo '[NAME]' && sed 's@/\* *@@; s/-/\\-/; q' $S/$$base.c) \
+	  && (echo '[NAME]'						\
+	      && sed 's@/\* *@@; s/-/\\-/;s/^GNU //; q' $S/$$base.c)	\
 	     | PATH="$(bin_dir)$(PATH_SEPARATOR)$$PATH"			\
-	       $(srcdir)/help2man -i - -i $(srcdir)/$$base.x		\
-		 -S '$(PACKAGE) $(VERSION)' $$base > $@-t && mv $@-t $@
+	       $(HELP2MAN) -i - -i $(srcdir)/$$base.x			\
+		 -S '$(PACKAGE) $(VERSION)' $$base > $$base.1-t		\
+	  && mv $$base.1-t $(srcdir)/$$base.1

man/cmp.x

@@ -1,2 +1,4 @@
 [SEE ALSO]
-diff(1), diff3(1), sdiff(1)
+.BR diff (1),
+.BR diff3 (1),
+.BR sdiff (1)

man/diff.x

@@ -1,2 +1,6 @@
 [SEE ALSO]
-wdiff(1), cmp(1), diff3(1), sdiff(1), patch(1)
+.BR wdiff (1),
+.BR cmp (1),
+.BR diff3 (1),
+.BR sdiff (1),
+.BR patch (1)

man/diff3.x

@@ -1,2 +1,4 @@
 [SEE ALSO]
-cmp(1), diff(1), sdiff(1)
+.BR cmp (1),
+.BR diff (1),
+.BR sdiff (1)

man/help2man

@@ -1,671 +0,0 @@
-#!/usr/bin/perl -w
-
-# Generate a short man page from --help and --version output.
-# Copyright (C) 1997-2005, 2009-2011, 2013, 2015-2017 Free Software Foundation,
-# Inc.
-
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 3, or (at your option)
-# any later version.
-
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-# GNU General Public License for more details.
-
-# You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software Foundation,
-# Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
-
-# Written by Brendan O'Dea <bod@debian.org>
-# Available from ftp://ftp.gnu.org/gnu/help2man/
-
-use 5.008;
-use strict;
-use Getopt::Long;
-use Text::Tabs qw(expand);
-use POSIX qw(strftime setlocale LC_ALL);
-
-my $this_program = 'help2man';
-my $this_version = '1.40.4';
-
-sub _ { $_[0] }
-sub configure_locale
-{
-    my $locale = shift;
-    die "$this_program: no locale support (Locale::gettext required)\n"
-	unless $locale eq 'C';
-}
-
-sub dec { $_[0] }
-sub enc { $_[0] }
-sub enc_user { $_[0] }
-sub kark { die +(sprintf shift, @_), "\n" }
-sub N_ { $_[0] }
-
-my $version_info = enc_user sprintf _(<<'EOT'), $this_program, $this_version;
-GNU %s %s
-
-Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2009, 2010,
-2011 Free Software Foundation, Inc.
-This is free software; see the source for copying conditions.  There is NO
-warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
-
-Written by Brendan O'Dea <bod@debian.org>
-EOT
-
-my $help_info = enc_user sprintf _(<<'EOT'), $this_program, $this_program;
-'%s' generates a man page out of '--help' and '--version' output.
-
-Usage: %s [OPTION]... EXECUTABLE
-
- -n, --name=STRING       description for the NAME paragraph
- -s, --section=SECTION   section number for manual page (1, 6, 8)
- -m, --manual=TEXT       name of manual (User Commands, ...)
- -S, --source=TEXT       source of program (FSF, Debian, ...)
- -L, --locale=STRING     select locale (default "C")
- -i, --include=FILE      include material from 'FILE'
- -I, --opt-include=FILE  include material from 'FILE' if it exists
- -o, --output=FILE       send output to 'FILE'
- -p, --info-page=TEXT    name of Texinfo manual
- -N, --no-info           suppress pointer to Texinfo manual
- -l, --libtool           exclude the 'lt-' from the program name
-     --help              print this help, then exit
-     --version           print version number, then exit
-
-EXECUTABLE should accept '--help' and '--version' options and produce output on
-stdout although alternatives may be specified using:
-
- -h, --help-option=STRING     help option string
- -v, --version-option=STRING  version option string
- --version-string=STRING      version string
- --no-discard-stderr          include stderr when parsing option output
-
-Report bugs to <bug-help2man@gnu.org>.
-EOT
-
-my $section = 1;
-my $manual = '';
-my $source = '';
-my $help_option = '--help';
-my $version_option = '--version';
-my $discard_stderr = 1;
-my ($opt_name, @opt_include, $opt_output, $opt_info, $opt_no_info, $opt_libtool,
-    $version_text);
-
-my %opt_def = (
-    'n|name=s'		 => \$opt_name,
-    's|section=s'	 => \$section,
-    'm|manual=s'	 => \$manual,
-    'S|source=s'	 => \$source,
-    'L|locale=s'	 => sub { configure_locale pop },
-    'i|include=s'	 => sub { push @opt_include, [ pop, 1 ] },
-    'I|opt-include=s'	 => sub { push @opt_include, [ pop, 0 ] },
-    'o|output=s'	 => \$opt_output,
-    'p|info-page=s'	 => \$opt_info,
-    'N|no-info'		 => \$opt_no_info,
-    'l|libtool'		 => \$opt_libtool,
-    'help'		 => sub { print $help_info; exit },
-    'version'		 => sub { print $version_info; exit },
-    'h|help-option=s'	 => \$help_option,
-    'v|version-option=s' => \$version_option,
-    'version-string=s'	 => \$version_text,
-    'discard-stderr!'	 => \$discard_stderr,
-);
-
-# Parse options.
-Getopt::Long::config('bundling');
-die $help_info unless GetOptions %opt_def and @ARGV == 1;
-
-my %include = ();
-my %append = ();
-my @include = (); # retain order given in include file
-
-# Process include file (if given).  Format is:
-#
-#   [section name]
-#   verbatim text
-#
-# or
-#
-#   /pattern/
-#   verbatim text
-#
-
-while (@opt_include)
-{
-    my ($inc, $required) = @{shift @opt_include};
-
-    next unless -f $inc or $required;
-    kark N_("%s: can't open '%s' (%s)"), $this_program, $inc, $!
-	unless open INC, $inc;
-
-    my $key;
-    my $hash = \%include;
-
-    while (<INC>)
-    {
-	# Convert input to internal Perl format, so that multibyte
-	# sequences are treated as single characters.
-	$_ = dec $_;
-
-	# [section]
-	if (/^\[([^]]+)\]\s*$/)
-	{
-	    $key = uc $1;
-	    $key =~ s/^\s+//;
-	    $key =~ s/\s+$//;
-	    $hash = \%include;
-	    push @include, $key unless $include{$key};
-	    next;
-	}
-
-	# /pattern/
-	if (m!^/(.*)/([ims]*)\s*$!)
-	{
-	    my $pat = $2 ? "(?$2)$1" : $1;
-
-	    # Check pattern.
-	    eval { $key = qr($pat) };
-	    if ($@)
-	    {
-		$@ =~ s/ at .*? line \d.*//;
-		die "$inc:$.:$@";
-	    }
-
-	    $hash = \%append;
-	    next;
-	}
-
-	# Check for options before the first section--anything else is
-	# silently ignored, allowing the first for comments and
-	# revision info.
-	unless ($key)
-	{
-	    # handle options
-	    if (/^-/)
-	    {
-		local @ARGV = split;
-		GetOptions %opt_def;
-	    }
-
-	    next;
-	}
-
-	$hash->{$key} ||= '';
-	$hash->{$key} .= $_;
-    }
-
-    close INC;
-
-    kark N_("%s: no valid information found in '%s'"), $this_program, $inc
-	unless $key;
-}
-
-# Compress trailing blank lines.
-for my $hash (\(%include, %append))
-{
-    for (keys %$hash) { $hash->{$_} =~ s/\n+$/\n/ }
-}
-
-sub get_option_value;
-
-# Grab help and version info from executable.
-my $help_text   = get_option_value $ARGV[0], $help_option;
-$version_text ||= get_option_value $ARGV[0], $version_option;
-
-# Translators: the following message is a strftime(3) format string, which in
-# the English version expands to the month as a word and the full year.  It
-# is used on the footer of the generated manual pages.  If in doubt, you may
-# just use %x as the value (which should be the full locale-specific date).
-my $date = enc strftime _("%B %Y"), localtime;
-(my $program = $ARGV[0]) =~ s!.*/!!;
-my $package = $program;
-my $version;
-
-if ($opt_output)
-{
-    unlink $opt_output or kark N_("%s: can't unlink %s (%s)"),
-	$this_program, $opt_output, $! if -e $opt_output;
-
-    open STDOUT, ">$opt_output"
-	or kark N_("%s: can't create %s (%s)"), $this_program, $opt_output, $!;
-}
-
-# The first line of the --version information is assumed to be in one
-# of the following formats:
-#
-#   <version>
-#   <program> <version>
-#   {GNU,Free} <program> <version>
-#   <program> ({GNU,Free} <package>) <version>
-#   <program> - {GNU,Free} <package> <version>
-#
-# and separated from any copyright/author details by a blank line.
-
-($_, $version_text) = ((split /\n+/, $version_text, 2), '');
-
-if (/^(\S+) +\(((?:GNU|Free) +[^)]+)\) +(.*)/ or
-    /^(\S+) +- *((?:GNU|Free) +\S+) +(.*)/)
-{
-    $program = $1;
-    $package = $2;
-    $version = $3;
-}
-elsif (/^((?:GNU|Free) +)?(\S+) +(.*)/)
-{
-    $program = $2;
-    $package = $1 ? "$1$2" : $2;
-    $version = $3;
-}
-else
-{
-    $version = $_;
-}
-
-$program =~ s!.*/!!;
-
-# No info for 'info' itself.
-$opt_no_info = 1 if $program eq 'info';
-
-# Translators: "NAME", "SYNOPSIS" and other one or two word strings in all
-# upper case are manual page section headings.  The man(1) manual page in your
-# language, if available should provide the conventional translations.
-for ($include{_('NAME')})
-{
-    if ($opt_name) # --name overrides --include contents.
-    {
-	$_ = "$program \\- $opt_name\n";
-    }
-    elsif ($_) # Use first name given as $program
-    {
-	$program = $1 if /^([^\s,]+)(?:,?\s*[^\s,\\-]+)*\s+\\?-/;
-    }
-    else # Set a default (useless) NAME paragraph.
-    {
-	$_ = sprintf _("%s \\- manual page for %s %s") . "\n", $program,
-	    $program, $version;
-    }
-}
-
-# Man pages traditionally have the page title in caps.
-my $PROGRAM = uc $program;
-
-# Set default page head/footers
-$source ||= "$program $version";
-unless ($manual)
-{
-    for ($section)
-    {
-	if (/^(1[Mm]|8)/) { $manual = enc _('System Administration Utilities') }
-	elsif (/^6/)	  { $manual = enc _('Games') }
-	else		  { $manual = enc _('User Commands') }
-    }
-}
-
-# Extract usage clause(s) [if any] for SYNOPSIS.
-# Translators: "Usage" and "or" here are patterns (regular expressions) which
-# are used to match the usage synopsis in program output.  An example from cp
-# (GNU coreutils) which contains both strings:
-#  Usage: cp [OPTION]... [-T] SOURCE DEST
-#    or:  cp [OPTION]... SOURCE... DIRECTORY
-#    or:  cp [OPTION]... -t DIRECTORY SOURCE...
-my $PAT_USAGE = _('Usage');
-my $PAT_USAGE_CONT = _('or');
-if ($help_text =~ s/^($PAT_USAGE):( +(\S+))(.*)((?:\n(?: {6}\1| *($PAT_USAGE_CONT): +\S).*)*)//om)
-{
-    my @syn = $3 . $4;
-
-    if ($_ = $5)
-    {
-	s/^\n//;
-	for (split /\n/) { s/^ *(($PAT_USAGE_CONT): +)?//o; push @syn, $_ }
-    }
-
-    my $synopsis = '';
-    for (@syn)
-    {
-	$synopsis .= ".br\n" if $synopsis;
-	s!^\S*/!!;
-	s/^lt-// if $opt_libtool;
-	s/^(\S+) *//;
-	$synopsis .= ".B $1\n";
-	s/\s+$//;
-	s/(([][]|\.\.+)+)/\\fR$1\\fI/g;
-	s/^/\\fI/ unless s/^\\fR//;
-	$_ .= '\fR';
-	s/(\\fI)( *)/$2$1/g;
-	s/\\fI\\fR//g;
-	s/^\\fR//;
-	s/\\fI$//;
-	s/^\./\\&./;
-
-	$synopsis .= "$_\n";
-    }
-
-    $include{_('SYNOPSIS')} ||= $synopsis;
-}
-
-# Process text, initial section is DESCRIPTION.
-my $sect = _('DESCRIPTION');
-$_ = "$help_text\n\n$version_text";
-
-# Normalise paragraph breaks.
-s/^\n+//;
-s/\n*$/\n/;
-s/\n\n+/\n\n/g;
-
-# Join hyphenated lines.
-s/([A-Za-z])-\n *([A-Za-z])/$1$2/g;
-
-# Temporarily exchange leading dots, apostrophes and backslashes for
-# tokens.
-s/^\./\x80/mg;
-s/^'/\x81/mg;
-s/\\/\x82/g;
-
-# Translators: patterns are used to match common program output. In the source
-# these strings are all of the form of "my $PAT_something = _('...');" and are
-# regular expressions.  If there is more than one commonly used string, you
-# may separate alternatives with "|".  Spaces in these expressions are written
-# as " +" to indicate that more than one space may be matched.  The string
-# "(?:[\\w-]+ +)?" in the bug reporting pattern is used to indicate an
-# optional word, so that either "Report bugs" or "Report _program_ bugs" will
-# be matched.
-my $PAT_BUGS		= _('Report +(?:[\w-]+ +)?bugs|Email +bug +reports +to');
-my $PAT_AUTHOR		= _('Written +by');
-my $PAT_OPTIONS		= _('Options');
-my $PAT_ENVIRONMENT	= _('Environment');
-my $PAT_FILES		= _('Files');
-my $PAT_EXAMPLES	= _('Examples');
-my $PAT_FREE_SOFTWARE	= _('This +is +free +software');
-
-# Start a new paragraph (if required) for these.
-s/([^\n])\n($PAT_BUGS|$PAT_AUTHOR) /$1\n\n$2 /og;
-
-# Convert iso-8859-1 copyright symbol or (c) to nroff
-# character.
-s/^Copyright +(?:\xa9|\([Cc]\))/Copyright \\(co/mg;
-
-sub convert_option;
-
-while (length)
-{
-    # Convert some standard paragraph names.
-    if (s/^($PAT_OPTIONS): *\n//o)
-    {
-	$sect = _('OPTIONS');
-	next;
-    }
-    if (s/^($PAT_ENVIRONMENT): *\n//o)
-    {
-	$sect = _('ENVIRONMENT');
-	next;
-    }
-    if (s/^($PAT_FILES): *\n//o)
-    {
-	$sect = _('FILES');
-	next;
-    }
-    elsif (s/^($PAT_EXAMPLES): *\n//o)
-    {
-	$sect = _('EXAMPLES');
-	next;
-    }
-
-    # Copyright section
-    if (/^Copyright /)
-    {
-	$sect = _('COPYRIGHT');
-    }
-
-    # Bug reporting section.
-    elsif (/^($PAT_BUGS) /o)
-    {
-	$sect = _('REPORTING BUGS');
-    }
-
-    # Author section.
-    elsif (/^($PAT_AUTHOR)/o)
-    {
-	$sect = _('AUTHOR');
-    }
-
-    # Examples, indicated by an indented leading $, % or > are
-    # rendered in a constant width font.
-    if (/^( +)([\$\%>] )\S/)
-    {
-	my $indent = $1;
-	my $prefix = $2;
-	my $break = '.IP';
-	$include{$sect} ||= '';
-	while (s/^$indent\Q$prefix\E(\S.*)\n*//)
-	{
-	    $include{$sect} .= "$break\n\\f(CW$prefix$1\\fR\n";
-	    $break = '.br';
-	}
-
-	next;
-    }
-
-    my $matched = '';
-    $include{$sect} ||= '';
-
-    # Sub-sections have a trailing colon and the second line indented.
-    if (s/^(\S.*:) *\n / /)
-    {
-	$matched .= $& if %append;
-	$include{$sect} .= qq(.SS "$1"\n);
-    }
-
-    my $indent = 0;
-    my $content = '';
-
-    # Option with description.
-    if (s/^( {1,10}([+-]\S.*?))(?:(  +(?!-))|\n( {20,}))(\S.*)\n//)
-    {
-	$matched .= $& if %append;
-	$indent = length ($4 || "$1$3");
-	$content = ".TP\n\x84$2\n\x84$5\n";
-	unless ($4)
-	{
-	    # Indent may be different on second line.
-	    $indent = length $& if /^ {20,}/;
-	}
-    }
-
-    # Option without description.
-    elsif (s/^ {1,10}([+-]\S.*)\n//)
-    {
-	$matched .= $& if %append;
-	$content = ".HP\n\x84$1\n";
-	$indent = 80; # not continued
-    }
-
-    # Indented paragraph with tag.
-    elsif (s/^( +(\S.*?)  +)(\S.*)\n//)
-    {
-	$matched .= $& if %append;
-	$indent = length $1;
-	$content = ".TP\n\x84$2\n\x84$3\n";
-    }
-
-    # Indented paragraph.
-    elsif (s/^( +)(\S.*)\n//)
-    {
-	$matched .= $& if %append;
-	$indent = length $1;
-	$content = ".IP\n\x84$2\n";
-    }
-
-    # Left justified paragraph.
-    else
-    {
-	s/(.*)\n//;
-	$matched .= $& if %append;
-	$content = ".PP\n" if $include{$sect};
-	$content .= "$1\n";
-    }
-
-    # Append continuations.
-    while ($indent ? s/^ {$indent}(\S.*)\n// : s/^(\S.*)\n//)
-    {
-	$matched .= $& if %append;
-	$content .= "\x84$1\n";
-    }
-
-    # Move to next paragraph.
-    s/^\n+//;
-
-    for ($content)
-    {
-	# Leading dot and apostrophe protection.
-	s/\x84\./\x80/g;
-	s/\x84'/\x81/g;
-	s/\x84//g;
-
-	# Convert options.
-	s/(^| |\()(-[][\w=-]+)/$1 . convert_option $2/mge;
-
-	# Escape remaining hyphens
-	s/-/\x83/g;
-
-	if ($sect eq 'COPYRIGHT')
-	{
-	    # Insert line breaks before additional copyright messages
-	    # and the disclaimer.
-	    s/\n(Copyright |$PAT_FREE_SOFTWARE)/\n.br\n$1/og;
-	}
-	elsif ($sect eq 'REPORTING BUGS')
-	{
-	    # Handle multi-line bug reporting sections of the form:
-	    #
-	    #   Report <program> bugs to <addr>
-	    #   GNU <package> home page: <url>
-	    #   ...
-	    s/\n([[:upper:]])/\n.br\n$1/g;
-	}
-    }
-
-    # Check if matched paragraph contains /pat/.
-    if (%append)
-    {
-	for my $pat (keys %append)
-	{
-	    if ($matched =~ $pat)
-	    {
-		$content .= ".PP\n" unless $append{$pat} =~ /^\./;
-		$content .= $append{$pat};
-	    }
-	}
-    }
-
-    $include{$sect} .= $content;
-}
-
-# Refer to the real documentation.
-unless ($opt_no_info)
-{
-    my $info_page = $opt_info || $program;
-
-    $sect = _('SEE ALSO');
-    $include{$sect} ||= '';
-    $include{$sect} .= ".PP\n" if $include{$sect};
-    $include{$sect} .= sprintf _(<<'EOT'), $program, $program, $info_page;
-The full documentation for
-.B %s
-is maintained as a Texinfo manual.  If the
-.B info
-and
-.B %s
-programs are properly installed at your site, the command
-.IP
-.B info %s
-.PP
-should give you access to the complete manual.
-EOT
-}
-
-# Output header.
-print <<EOT;
-.\\" DO NOT MODIFY THIS FILE!  It was generated by $this_program $this_version.
-.TH $PROGRAM "$section" "$date" "$source" "$manual"
-EOT
-
-# Section ordering.
-my @pre = (_('NAME'), _('SYNOPSIS'), _('DESCRIPTION'), _('OPTIONS'),
-    _('ENVIRONMENT'), _('FILES'), _('EXAMPLES'));
-
-my @post = (_('AUTHOR'), _('REPORTING BUGS'), _('COPYRIGHT'), _('SEE ALSO'));
-my $filter = join '|', @pre, @post;
-
-# Output content.
-for my $sect (@pre, (grep ! /^($filter)$/o, @include), @post)
-{
-    if ($include{$sect})
-    {
-	my $quote = $sect =~ /\W/ ? '"' : '';
-	print enc ".SH $quote$sect$quote\n";
-
-	for ($include{$sect})
-	{
-	    # Replace leading dot, apostrophe, backslash and hyphen
-	    # tokens.
-	    s/\x80/\\&./g;
-	    s/\x81/\\&'/g;
-	    s/\x82/\\e/g;
-	    s/\x83/\\-/g;
-
-	    # Convert some latin1 chars to troff equivalents
-	    s/\xa0/\\ /g; # non-breaking space
-
-	    print enc $_;
-	}
-    }
-}
-
-close STDOUT or kark N_("%s: error writing to %s (%s)"), $this_program,
-    $opt_output || 'stdout', $!;
-
-exit;
-
-# Call program with given option and return results.
-sub get_option_value
-{
-    my ($prog, $opt) = @_;
-    my $stderr = $discard_stderr ? '/dev/null' : '&1';
-    my $value = join '',
-	map { s/ +$//; expand $_ }
-	map { dec $_ }
-	`$prog $opt 2>$stderr`;
-
-    unless ($value)
-    {
-	my $err = N_("%s: can't get '%s' info from %s%s");
-	my $extra = $discard_stderr
-	    ? "\n" . N_("Try '--no-discard-stderr' if option outputs to stderr")
-	    : '';
-
-	kark $err, $this_program, $opt, $prog, $extra;
-    }
-
-    return $value;
-}
-
-# Convert option dashes to \- to stop nroff from hyphenating 'em, and
-# embolden.  Option arguments get italicised.
-sub convert_option
-{
-    local $_ = '\fB' . shift;
-
-    s/-/\x83/g;
-    unless (s/\[=(.*)\]$/\\fR[=\\fI$1\\fR]/)
-    {
-	s/=(.)/\\fR=\\fI$1/;
-	s/ (.)/ \\fI$1/;
-	$_ .= '\fR';
-    }
-
-    $_;
-}

man/sdiff.x

@@ -1,2 +1,4 @@
 [SEE ALSO]
-cmp(1), diff(1), diff3(1)
+.BR cmp (1),
+.BR diff (1),
+.BR diff3 (1)

po/POTFILES.in

@@ -1,6 +1,6 @@
 # List of files that contain translatable strings.
 
-# Copyright (C) 2001-2002, 2009-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2001-2002, 2009-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -16,16 +16,19 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
 lib/argmatch.c
+lib/argmatch.h
+lib/c-file-type.c
 lib/c-stack.c
+lib/diagnose.c
 lib/error.c
-lib/file-type.c
 lib/getopt.c
+lib/openat-die.c
 lib/quotearg.c
 lib/regcomp.c
+lib/version-etc.c
 lib/xalloc-die.c
 lib/xfreopen.c
-lib/xstrtol-error.c
-lib/version-etc.c
+lib/xstdopen.c
 
 src/analyze.c
 src/cmp.c

po/en.po

@@ -1,30 +0,0 @@
-# English messages for GNU diffutils
-# Copyright 1998, 2001-2003, 2009-2013, 2015-2017 Free Software Foundation,
-# Inc.
-# Paul Eggert <eggert@twinsun.com>, 1998
-#
-msgid ""
-msgstr ""
-"Project-Id-Version: GNU diffutils 3.2\n"
-"POT-Creation-Date: 2002-06-16 23:44-0700\n"
-"PO-Revision-Date: 2012-01-25 23:11-0700\n"
-"Last-Translator: Paul Eggert <eggert@cs.ucla.edu>\n"
-"Language-Team: English <en@translate.freefriends.org>\n"
-"MIME-Version: 1.0\n"
-"Content-Type: text/plain; charset=ISO-8859-1\n"
-"Content-Transfer-Encoding: 8bit\n"
-
-#. TRANSLATORS: Please translate "(C)" to the C-in-a-circle symbol
-#. (U+00A9, COPYRIGHT SIGN) if possible, as this has some minor
-#. technical advantages in international copyright law.  If the
-#. copyright symbol is not available, please leave it as "(C)".
-#: lib/version-etc.c:50
-msgid "(C)"
-msgstr "©"
-
-#. TRANSLATORS: Please translate the second "o" in "Torbjorn Granlund"
-#. to an o-with-umlaut (U+00F6, LATIN SMALL LETTER O WITH DIAERESIS)
-#. if possible.
-#: src/cmp.c:47
-msgid "Written by Torbjorn Granlund and David MacKenzie."
-msgstr "Written by Torbjörn Granlund and David MacKenzie."

scripts/git-hooks/applypatch-msg

@@ -10,6 +10,6 @@
 # To enable this hook, rename this file to "applypatch-msg".
 
 . git-sh-setup
-commitmsg="$(git rev-parse --git-path hooks/commit-msg)"
-test -x "$commitmsg" && exec "$commitmsg" ${1+"$@"}
+test -x "$GIT_DIR/hooks/commit-msg" &&
+	exec "$GIT_DIR/hooks/commit-msg" ${1+"$@"}
 :

scripts/git-hooks/commit-msg

@@ -16,7 +16,7 @@ $editor = "vi" if $? != 0 or $editor =~ /^\s*\z/;
 my @valid = qw(
     diff diff3 cmp sdiff
 
-    gnulib tests maint doc build scripts
+    all gnulib tests maint doc build scripts
     );
 my $v_or = join '|', @valid;
 my $valid_regex = qr/^(?:$v_or)$/;
@@ -79,7 +79,7 @@ sub check_msg($$)
   my ($log_file, $line_ref) = @_;
 
   local *LOG;
-  open LOG, '<', $log_file
+  open LOG, '<:utf8', $log_file
     or return "failed to open for reading: $!";
   @$line_ref = <LOG>;
   close LOG;
@@ -109,21 +109,24 @@ sub check_msg($$)
     and return 'second line must be empty';
 
   # Limit line length to allow for the ChangeLog's leading TAB.
+  my $max_len = 72;
   foreach my $line (@line)
     {
-      72 < length $line && $line =~ /^[^#]/
-        and return 'line longer than 72';
+      last if $line =~ '.*-{24} >8 -{24}$';
+      my $len = length $line;
+      $max_len < $len && $line =~ /^[^#]/
+        and return "line length ($len) greater than than max: $max_len";
     }
 
   my $buf = join ("\n", @line) . "\n";
   $buf =~ m!https?://bugzilla\.redhat\.com/show_bug\.cgi\?id=(\d+)!s
-    and return "use shorter http://bugzilla.redhat.com/$1";
+    and return "use shorter https://bugzilla.redhat.com/$1";
 
   $buf =~ m!https?://debbugs\.gnu\.org/(?:cgi/bugreport\.cgi\?bug=)?(\d+)!s
-    and return "use shorter http://bugs.gnu.org/$1";
+    and return "use shorter https://bugs.gnu.org/$1";
 
-  $buf =~ /^ *Signed-off-by:/mi
-    and return q(do not use "Signed-off-by:");
+  $buf =~ m!https://lists\.gnu\.org/archive/html/!s
+    and return "use '/r/' in place of '/archive/html/' in lists.gnu.org URLs";
 
   return '';
 }
@@ -141,6 +144,7 @@ sub check_msg($$)
       $err eq ''
         and last;
       $err = "$ME: $err\n";
+      -t STDOUT or die $err;
       warn $err;
       # Insert the diagnostic as a comment on the first line of $log_file.
       rewrite $log_file, $err, \@line;

scripts/git-hooks/pre-applypatch

@@ -9,6 +9,6 @@
 # To enable this hook, rename this file to "pre-applypatch".
 
 . git-sh-setup
-precommit="$(git rev-parse --git-path hooks/pre-commit)"
-test -x "$precommit" && exec "$precommit" ${1+"$@"}
+test -x "$GIT_DIR/hooks/pre-commit" &&
+	exec "$GIT_DIR/hooks/pre-commit" ${1+"$@"}
 :

scripts/git-hooks/pre-commit

@@ -15,13 +15,13 @@ else
 	against=4b825dc642cb6eb9a060e54bf8d69288fbee4904
 fi
 
-# If you want to allow non-ASCII filenames set this variable to true.
-allownonascii=$(git config --bool hooks.allownonascii)
+# If you want to allow non-ascii filenames set this variable to true.
+allownonascii=$(git config hooks.allownonascii)
 
 # Redirect output to stderr.
 exec 1>&2
 
-# Cross platform projects tend to avoid non-ASCII filenames; prevent
+# Cross platform projects tend to avoid non-ascii filenames; prevent
 # them from being added to the repository. We exploit the fact that the
 # printable range starts at the space character and ends with tilde.
 if [ "$allownonascii" != "true" ] &&
@@ -31,17 +31,18 @@ if [ "$allownonascii" != "true" ] &&
 	test $(git diff --cached --name-only --diff-filter=A -z $against |
 	  LC_ALL=C tr -d '[ -~]\0' | wc -c) != 0
 then
-	cat <<\EOF
-Error: Attempt to add a non-ASCII file name.
-
-This can cause problems if you want to work with people on other platforms.
-
-To be portable it is advisable to rename the file.
-
-If you know what you are doing you can disable this check using:
-
-  git config hooks.allownonascii true
-EOF
+	echo "Error: Attempt to add a non-ascii file name."
+	echo
+	echo "This can cause problems if you want to work"
+	echo "with people on other platforms."
+	echo
+	echo "To be portable it is advisable to rename the file ..."
+	echo
+	echo "If you know what you are doing you can disable this"
+	echo "check using:"
+	echo
+	echo "  git config hooks.allownonascii true"
+	echo
 	exit 1
 fi
 

src/Makefile.am

@@ -1,6 +1,6 @@
 # Automakefile for GNU diffutils programs.
 
-# Copyright (C) 2001-2002, 2006, 2009-2013, 2015-2017 Free Software Foundation,
+# Copyright (C) 2001-2002, 2006, 2009-2013, 2015-2026 Free Software Foundation,
 # Inc.
 
 # This program is free software: you can redistribute it and/or modify
@@ -26,31 +26,33 @@ AM_CFLAGS = $(WARN_CFLAGS) $(WERROR_CFLAGS)
 LDADD = \
   libver.a \
   ../lib/libdiffutils.a \
+  $(CLOCK_TIME_LIB) \
+  $(HARD_LOCALE_LIB) \
+  $(LIBTHREAD) \
   $(LIBCSTACK) \
   $(LIBINTL) \
-  $(LIBICONV) \
   $(LIBSIGSEGV) \
-  $(LIB_CLOCK_GETTIME)
+  $(LIBUNISTRING) \
+  $(MBRTOWC_LIB) \
+  $(LIBC32CONV) \
+  $(SETLOCALE_NULL_LIB)
 
 diff_LDADD = $(LDADD)
 cmp_LDADD = $(LDADD)
-sdiff_LDADD = $(LDADD)
+sdiff_LDADD = $(LDADD) $(GETRANDOM_LIB)
 diff3_LDADD = $(LDADD)
 
-cmp_SOURCES = cmp.c
-diff3_SOURCES = diff3.c
-sdiff_SOURCES = sdiff.c
+cmp_SOURCES = cmp.c system.c
+diff3_SOURCES = diff3.c system.c
+sdiff_SOURCES = sdiff.c system.c
 diff_SOURCES = \
   analyze.c context.c diff.c dir.c ed.c ifdef.c io.c \
-  normal.c side.c util.c
-noinst_HEADERS =	\
-  die.h			\
-  diff.h		\
-  system.h
+  normal.c side.c syncsig.c system.c util.c
+noinst_HEADERS = diff.h syncsig.h system.h
 
 MOSTLYCLEANFILES = paths.h paths.ht
 
-cmp.$(OBJEXT) diff3.$(OBJEXT) diff.$(OBJEXT) sdiff.$(OBJEXT): paths.h
+cmp.$(OBJEXT) diff3.$(OBJEXT) diff.$(OBJEXT) sdiff.$(OBJEXT): paths.h version.h
 
 gdiff = `echo diff|sed '$(transform)'`
 BUILT_SOURCES = paths.h
@@ -65,6 +67,7 @@ BUILT_SOURCES += version.c
 version.c: Makefile
 	$(AM_V_GEN)rm -f $@
 	$(AM_V_at)printf '#include <config.h>\n' > $@t
+	$(AM_V_at)printf '#include <version.h>\n' >> $@t
 	$(AM_V_at)printf 'char const *Version = "$(PACKAGE_VERSION)";\n' >> $@t
 	$(AM_V_at)chmod a-w $@t
 	$(AM_V_at)mv $@t $@

src/context.c

@@ -1,7 +1,7 @@
 /* Context-format output routines for GNU DIFF.
 
    Copyright (C) 1988-1989, 1991-1995, 1998, 2001-2002, 2004, 2006, 2009-2013,
-   2015-2017 Free Software Foundation, Inc.
+   2015-2026 Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
 
@@ -19,11 +19,11 @@
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
 #include "diff.h"
-#include "c-ctype.h"
+#include <c-ctype.h>
 #include <stat-time.h>
 #include <strftime.h>
 
-static char const *find_function (char const * const *, lin);
+static char const *find_function (char const *const *, lin);
 static struct change *find_hunk (struct change *);
 static void mark_ignorable (struct change *);
 static void pr_context_hunk (struct change *);
@@ -39,40 +39,55 @@ static lin find_function_last_match;
 
 static void
 print_context_label (char const *mark,
-		     struct file_data *inf,
-		     char const *name,
-		     char const *label)
+                     struct file_data *inf,
+                     char const *name,
+                     char const *label)
 {
+  set_color_context (HEADER_CONTEXT);
   if (label)
-    fprintf (outfile, "%s %s\n", mark, label);
+    fprintf (outfile, "%s %s", mark, label);
   else
     {
-      char buf[MAX (INT_STRLEN_BOUND (int) + 32,
-		    INT_STRLEN_BOUND (time_t) + 11)];
-      struct tm const *tm = localtime (&inf->stat.st_mtime);
-      int nsec = get_stat_mtime_ns (&inf->stat);
-      if (! (tm && nstrftime (buf, sizeof buf, time_format, tm, 0, nsec)))
+      /* POSIX requires current time for stdin.  */
+      struct timespec ts;
+      if (inf->desc == STDIN_FILENO)
 	{
-	  verify (TYPE_IS_INTEGER (time_t));
-	  if (LONG_MIN <= TYPE_MINIMUM (time_t)
-	      && TYPE_MAXIMUM (time_t) <= LONG_MAX)
-	    {
-	      long int sec = inf->stat.st_mtime;
-	      sprintf (buf, "%ld.%.9d", sec, nsec);
-	    }
-	  else if (TYPE_MAXIMUM (time_t) <= INTMAX_MAX)
-	    {
-	      intmax_t sec = inf->stat.st_mtime;
-	      sprintf (buf, "%"PRIdMAX".%.9d", sec, nsec);
-	    }
-	  else
-	    {
-	      uintmax_t sec = inf->stat.st_mtime;
-	      sprintf (buf, "%"PRIuMAX".%.9d", sec, nsec);
-	    }
+	  static struct timespec now;
+	  if (!now.tv_sec)
+	    timespec_get (&now, TIME_UTC);
+	  ts = now;
+	}
+      else
+	ts = get_stat_mtime (&inf->stat);
+
+      /* Buffer for nstftime output, big enough to handle any
+	 timestamp formatted according to time_format.
+	 Its size is an upper bound for the format "%Y-%m-%d %H:%M:%S.%N %z",
+	 with an int for year and a time_t for time zone hour.
+	 The format "%Y-%m-%d %H:%M:%S %z" generates fewer bytes,
+	 and although the format "%a %b %e %T %Y" could in theory
+	 generate more bytes in practice it never does.  */
+      char buf[INT_STRLEN_BOUND (int) + INT_STRLEN_BOUND (time_t)
+	       + sizeof "-%m-%d %H:%M:%S.000000000 +00"];
+
+      struct tm const *tm = localtime (&ts.tv_sec);
+      int nsec = ts.tv_nsec;
+      if (tm
+	  && 0 <= nstrftime (buf, sizeof buf, time_format, tm, localtz, nsec))
+	fprintf (outfile, "%s %s\t%s", mark, name, buf);
+      else if (TYPE_SIGNED (time_t))
+        {
+	  intmax_t sec = inf->stat.st_mtime;
+	  fprintf (outfile, "%s %s\t%"PRIdMAX".%.9d", mark, name, sec, nsec);
+	}
+      else
+	{
+	  uintmax_t sec = inf->stat.st_mtime;
+	  fprintf (outfile, "%s %s\t%"PRIuMAX".%.9d", mark, name, sec, nsec);
 	}
-      fprintf (outfile, "%s %s\t%s\n", mark, name, buf);
     }
+  set_color_context (RESET_CONTEXT);
+  putc ('\n', outfile);
 }
 
 /* Print a header for a context diff, with the file names and dates.  */
@@ -80,7 +95,6 @@ print_context_label (char const *mark,
 void
 print_context_header (struct file_data inf[], char const *const *names, bool unidiff)
 {
-  set_color_context (HEADER_CONTEXT);
   if (unidiff)
     {
       print_context_label ("---", &inf[0], names[0], file_label[0]);
@@ -91,7 +105,6 @@ print_context_header (struct file_data inf[], char const *const *names, bool uni
       print_context_label ("***", &inf[0], names[0], file_label[0]);
       print_context_label ("---", &inf[1], names[1], file_label[1]);
     }
-  set_color_context (RESET_CONTEXT);
 }
 
 /* Print an edit script in context format.  */
@@ -102,13 +115,10 @@ print_context_script (struct change *script, bool unidiff)
   if (ignore_blank_lines || ignore_regexp.fastmap)
     mark_ignorable (script);
   else
-    {
-      struct change *e;
-      for (e = script; e; e = e->link)
-	e->ignore = false;
-    }
+    for (struct change *e = script; e; e = e->link)
+      e->ignore = false;
 
-  find_function_last_search = - files[0].prefix_lines;
+  find_function_last_search = - curr.file[0].prefix_lines;
   find_function_last_match = LIN_MAX;
 
   if (unidiff)
@@ -126,7 +136,7 @@ print_context_script (struct change *script, bool unidiff)
 static void
 print_context_number_range (struct file_data const *file, lin a, lin b)
 {
-  printint trans_a, trans_b;
+  lin trans_a, trans_b;
   translate_range (file, a, b, &trans_a, &trans_b);
 
   /* We can have B <= A in the case of a range of no lines.
@@ -169,38 +179,33 @@ print_context_function (FILE *out, char const *function)
 static void
 pr_context_hunk (struct change *hunk)
 {
-  lin first0, last0, first1, last1, i;
-  char const *prefix;
-  char const *function;
-  FILE *out;
-
   /* Determine range of line numbers involved in each file.  */
-
+  lin first0, last0, first1, last1;
   enum changes changes = analyze_hunk (hunk, &first0, &last0, &first1, &last1);
   if (! changes)
     return;
 
   /* Include a context's width before and after.  */
 
-  i = - files[0].prefix_lines;
-  first0 = MAX (first0 - context, i);
-  first1 = MAX (first1 - context, i);
-  if (last0 < files[0].valid_lines - context)
+  lin minus_prefix_lines = - curr.file[0].prefix_lines;
+  first0 = MAX (first0 - context, minus_prefix_lines);
+  first1 = MAX (first1 - context, minus_prefix_lines);
+  if (last0 < curr.file[0].valid_lines - context)
     last0 += context;
   else
-    last0 = files[0].valid_lines - 1;
-  if (last1 < files[1].valid_lines - context)
+    last0 = curr.file[0].valid_lines - 1;
+  if (last1 < curr.file[1].valid_lines - context)
     last1 += context;
   else
-    last1 = files[1].valid_lines - 1;
+    last1 = curr.file[1].valid_lines - 1;
 
   /* If desired, find the preceding function definition line in file 0.  */
-  function = NULL;
+  char const *function = nullptr;
   if (function_regexp.fastmap)
-    function = find_function (files[0].linbuf, first0);
+    function = find_function (curr.file[0].linbuf, first0);
 
   begin_output ();
-  out = outfile;
+  FILE *out = outfile;
 
   fputs ("***************", out);
 
@@ -210,7 +215,7 @@ pr_context_hunk (struct change *hunk)
   putc ('\n', out);
   set_color_context (LINE_NUMBER_CONTEXT);
   fputs ("*** ", out);
-  print_context_number_range (&files[0], first0, last0);
+  print_context_number_range (&curr.file[0], first0, last0);
   fputs (" ****", out);
   set_color_context (RESET_CONTEXT);
   putc ('\n', out);
@@ -219,38 +224,36 @@ pr_context_hunk (struct change *hunk)
     {
       struct change *next = hunk;
 
-      if (first0 <= last0)
-        set_color_context (DELETE_CONTEXT);
+      for (lin i = first0; i <= last0; i++)
+        {
+          set_color_context (DELETE_CONTEXT);
 
-      for (i = first0; i <= last0; i++)
-	{
-	  /* Skip past changes that apply (in file 0)
-	     only to lines before line I.  */
+          /* Skip past changes that apply (in file 0)
+             only to lines before line I.  */
 
-	  while (next && next->line0 + next->deleted <= i)
-	    next = next->link;
+          while (next && next->line0 + next->deleted <= i)
+            next = next->link;
 
-	  /* Compute the marking for line I.  */
+          /* Compute the marking for line I.  */
 
-	  prefix = " ";
-	  if (next && next->line0 <= i)
+          char const *prefix = " ";
+          if (next && next->line0 <= i)
             {
               /* The change NEXT covers this line.
                  If lines were inserted here in file 1, this is "changed".
                  Otherwise it is "deleted".  */
               prefix = (next->inserted > 0 ? "!" : "-");
             }
-	  print_1_line_nl (prefix, &files[0].linbuf[i], true);
-          if (i == last0)
-            set_color_context (RESET_CONTEXT);
-          if (files[0].linbuf[i + 1][-1] == '\n')
+	  print_1_line_nl (prefix, &curr.file[0].linbuf[i], true);
+          set_color_context (RESET_CONTEXT);
+	  if (curr.file[0].linbuf[i + 1][-1] == '\n')
             putc ('\n', out);
-	}
+        }
     }
 
   set_color_context (LINE_NUMBER_CONTEXT);
   fputs ("--- ", out);
-  print_context_number_range (&files[1], first1, last1);
+  print_context_number_range (&curr.file[1], first1, last1);
   fputs (" ----", out);
   set_color_context (RESET_CONTEXT);
   putc ('\n', out);
@@ -259,33 +262,31 @@ pr_context_hunk (struct change *hunk)
     {
       struct change *next = hunk;
 
-      if (first1 <= last1)
-        set_color_context (ADD_CONTEXT);
+      for (lin i = first1; i <= last1; i++)
+        {
+          set_color_context (ADD_CONTEXT);
 
-      for (i = first1; i <= last1; i++)
-	{
-	  /* Skip past changes that apply (in file 1)
-	     only to lines before line I.  */
+          /* Skip past changes that apply (in file 1)
+             only to lines before line I.  */
 
-	  while (next && next->line1 + next->inserted <= i)
-	    next = next->link;
+          while (next && next->line1 + next->inserted <= i)
+            next = next->link;
 
-	  /* Compute the marking for line I.  */
+          /* Compute the marking for line I.  */
 
-	  prefix = " ";
-	  if (next && next->line1 <= i)
+          char const *prefix = " ";
+          if (next && next->line1 <= i)
             {
               /* The change NEXT covers this line.
                  If lines were deleted here in file 0, this is "changed".
                  Otherwise it is "inserted".  */
               prefix = (next->deleted > 0 ? "!" : "+");
             }
-	  print_1_line_nl (prefix, &files[1].linbuf[i], true);
-          if (i == last1)
-            set_color_context (RESET_CONTEXT);
-          if (files[1].linbuf[i + 1][-1] == '\n')
+	  print_1_line_nl (prefix, &curr.file[1].linbuf[i], true);
+          set_color_context (RESET_CONTEXT);
+	  if (curr.file[1].linbuf[i + 1][-1] == '\n')
             putc ('\n', out);
-	}
+        }
     }
 }
 
@@ -299,7 +300,7 @@ pr_context_hunk (struct change *hunk)
 static void
 print_unidiff_number_range (struct file_data const *file, lin a, lin b)
 {
-  printint trans_a, trans_b;
+  lin trans_a, trans_b;
   translate_range (file, a, b, &trans_a, &trans_b);
 
   /* We can have B < A in the case of a range of no lines.
@@ -322,44 +323,38 @@ print_unidiff_number_range (struct file_data const *file, lin a, lin b)
 static void
 pr_unidiff_hunk (struct change *hunk)
 {
-  lin first0, last0, first1, last1;
-  lin i, j, k;
-  struct change *next;
-  char const *function;
-  FILE *out;
-
   /* Determine range of line numbers involved in each file.  */
-
+  lin first0, last0, first1, last1;
   if (! analyze_hunk (hunk, &first0, &last0, &first1, &last1))
     return;
 
   /* Include a context's width before and after.  */
 
-  i = - files[0].prefix_lines;
-  first0 = MAX (first0 - context, i);
-  first1 = MAX (first1 - context, i);
-  if (last0 < files[0].valid_lines - context)
+  lin minus_prefix_lines = - curr.file[0].prefix_lines;
+  first0 = MAX (first0 - context, minus_prefix_lines);
+  first1 = MAX (first1 - context, minus_prefix_lines);
+  if (last0 < curr.file[0].valid_lines - context)
     last0 += context;
   else
-    last0 = files[0].valid_lines - 1;
-  if (last1 < files[1].valid_lines - context)
+    last0 = curr.file[0].valid_lines - 1;
+  if (last1 < curr.file[1].valid_lines - context)
     last1 += context;
   else
-    last1 = files[1].valid_lines - 1;
+    last1 = curr.file[1].valid_lines - 1;
 
   /* If desired, find the preceding function definition line in file 0.  */
-  function = NULL;
+  char const *function = nullptr;
   if (function_regexp.fastmap)
-    function = find_function (files[0].linbuf, first0);
+    function = find_function (curr.file[0].linbuf, first0);
 
   begin_output ();
-  out = outfile;
+  FILE *out = outfile;
 
   set_color_context (LINE_NUMBER_CONTEXT);
   fputs ("@@ -", out);
-  print_unidiff_number_range (&files[0], first0, last0);
+  print_unidiff_number_range (&curr.file[0], first0, last0);
   fputs (" +", out);
-  print_unidiff_number_range (&files[1], first1, last1);
+  print_unidiff_number_range (&curr.file[1], first1, last1);
   fputs (" @@", out);
   set_color_context (RESET_CONTEXT);
 
@@ -368,9 +363,9 @@ pr_unidiff_hunk (struct change *hunk)
 
   putc ('\n', out);
 
-  next = hunk;
-  i = first0;
-  j = first1;
+  struct change *next = hunk;
+  lin i = first0;
+  lin j = first1;
 
   while (i <= last0 || j <= last1)
     {
@@ -378,61 +373,57 @@ pr_unidiff_hunk (struct change *hunk)
       /* If the line isn't a difference, output the context from file 0. */
 
       if (!next || i < next->line0)
-	{
-	  char const *const *line = &files[0].linbuf[i++];
-	  if (! (suppress_blank_empty && **line == '\n'))
-	    putc (initial_tab ? '\t' : ' ', out);
-	  print_1_line (NULL, line);
-	  j++;
-	}
+        {
+	  char const *const *line = &curr.file[0].linbuf[i++];
+          if (! (suppress_blank_empty && **line == '\n'))
+            putc (initial_tab ? '\t' : ' ', out);
+          print_1_line (nullptr, line);
+          j++;
+        }
       else
-	{
-	  /* For each difference, first output the deleted part. */
+        {
+          /* For each difference, first output the deleted part. */
 
-	  k = next->deleted;
-          if (k)
-            set_color_context (DELETE_CONTEXT);
-
-	  while (k--)
-	    {
-	      char const * const *line = &files[0].linbuf[i++];
-	      putc ('-', out);
-	      if (initial_tab && ! (suppress_blank_empty && **line == '\n'))
-		putc ('\t', out);
-	      print_1_line_nl (NULL, line, true);
-
-              if (!k)
-                set_color_context (RESET_CONTEXT);
-
-              if (line[1][-1] == '\n')
-                putc ('\n', out);
-	    }
-
-	  /* Then output the inserted part. */
-
-	  k = next->inserted;
-          if (k)
-            set_color_context (ADD_CONTEXT);
+          lin k = next->deleted;
 
           while (k--)
-	    {
-	      char const * const *line = &files[1].linbuf[j++];
-	      putc ('+', out);
-	      if (initial_tab && ! (suppress_blank_empty && **line == '\n'))
-		putc ('\t', out);
-	      print_1_line_nl (NULL, line, true);
+            {
+	      char const *const *line = &curr.file[0].linbuf[i++];
+              set_color_context (DELETE_CONTEXT);
+              putc ('-', out);
+              if (initial_tab && ! (suppress_blank_empty && **line == '\n'))
+                putc ('\t', out);
+              print_1_line_nl (nullptr, line, true);
 
-              if (!k)
-                set_color_context (RESET_CONTEXT);
+              set_color_context (RESET_CONTEXT);
 
               if (line[1][-1] == '\n')
                 putc ('\n', out);
-	    }
+            }
 
-	  /* We're done with this hunk, so on to the next! */
+          /* Then output the inserted part. */
 
-	  next = next->link;
-	}
+          k = next->inserted;
+
+          while (k--)
+            {
+	      char const *const *line = &curr.file[1].linbuf[j++];
+              set_color_context (ADD_CONTEXT);
+              putc ('+', out);
+              if (initial_tab && ! (suppress_blank_empty && **line == '\n'))
+                putc ('\t', out);
+              print_1_line_nl (nullptr, line, true);
+
+              set_color_context (RESET_CONTEXT);
+
+              if (line[1][-1] == '\n')
+                putc ('\n', out);
+            }
+
+          /* We're done with this hunk, so on to the next! */
+
+          next = next->link;
+        }
     }
 }
 
@@ -440,39 +431,33 @@ pr_unidiff_hunk (struct change *hunk)
    2*CONTEXT unchanged lines appear, and return a pointer
    to the 'struct change' for the last change before those lines.  */
 
-static struct change * _GL_ATTRIBUTE_PURE
-find_hunk (struct change *start)
+static struct change * ATTRIBUTE_PURE
+find_hunk (struct change *script)
 {
-  struct change *prev;
-  lin top0, top1;
-  lin thresh;
-
   /* Threshold distance is CONTEXT if the second change is ignorable,
-     2 * CONTEXT + 1 otherwise.  Integer overflow can't happen, due
-     to CONTEXT_LIM.  */
+     min (2 * CONTEXT + 1, LIN_MAX) otherwise.  */
   lin ignorable_threshold = context;
-  lin non_ignorable_threshold = 2 * context + 1;
+  lin non_ignorable_threshold = (ckd_mul (&non_ignorable_threshold, context, 2)
+				 ? LIN_MAX
+				 : non_ignorable_threshold + 1);
 
-  do
+  for (struct change *next; ; script = next)
     {
+      next = script->link;
       /* Compute number of first line in each file beyond this changed.  */
-      top0 = start->line0 + start->deleted;
-      top1 = start->line1 + start->inserted;
-      prev = start;
-      start = start->link;
-      thresh = (start && start->ignore
-		? ignorable_threshold
-		: non_ignorable_threshold);
-      /* It is not supposed to matter which file we check in the end-test.
-	 If it would matter, crash.  */
-      if (start && start->line0 - top0 != start->line1 - top1)
-	abort ();
-    } while (start
-	     /* Keep going if less than THRESH lines
-		elapse before the affected line.  */
-	     && start->line0 - top0 < thresh);
+      lin top0 = script->line0 + script->deleted;
+      lin top1 = script->line1 + script->inserted;
+      lin thresh = (next && next->ignore
+		    ? ignorable_threshold
+		    : non_ignorable_threshold);
+      /* It is not supposed to matter which file we check in the end-test.  */
+      dassert (!next || next->line0 - top0 == next->line1 - top1);
 
-  return prev;
+      /* Keep going if less than THRESH lines elapse
+	 before the affected line.  */
+      if (!next || thresh <= next->line0 - top0)
+	return script;
+    }
 }
 
 /* Set the 'ignore' flag properly in each change in SCRIPT.
@@ -485,14 +470,14 @@ mark_ignorable (struct change *script)
   while (script)
     {
       struct change *next = script->link;
-      lin first0, last0, first1, last1;
 
       /* Turn this change into a hunk: detach it from the others.  */
-      script->link = NULL;
+      script->link = nullptr;
 
       /* Determine whether this change is ignorable.  */
+      lin first0, last0, first1, last1;
       script->ignore = ! analyze_hunk (script,
-				       &first0, &last0, &first1, &last1);
+                                       &first0, &last0, &first1, &last1);
 
       /* Reconnect the chain as before.  */
       script->link = next;
@@ -504,10 +489,10 @@ mark_ignorable (struct change *script)
 
 /* Find the last function-header line in LINBUF prior to line number LINENUM.
    This is a line containing a match for the regexp in 'function_regexp'.
-   Return the address of the text, or NULL if no function-header is found.  */
+   Return the address of the text, or null if no function-header is found.  */
 
 static char const *
-find_function (char const * const *linbuf, lin linenum)
+find_function (char const *const *linbuf, lin linenum)
 {
   lin i = linenum;
   lin last = find_function_last_search;
@@ -517,21 +502,22 @@ find_function (char const * const *linbuf, lin linenum)
     {
       /* See if this line is what we want.  */
       char const *line = linbuf[i];
-      size_t linelen = linbuf[i + 1] - line - 1;
+      idx_t linelen = linbuf[i + 1] - line - 1;
 
-      /* FIXME: re_search's size args should be size_t, not int.  */
-      int len = MIN (linelen, INT_MAX);
+      /* This line is for documentation; in practice it's equivalent
+	 to LEN = LINELEN and no machine code is generated.  */
+      regoff_t len = MIN (linelen, TYPE_MAXIMUM (regoff_t));
 
-      if (0 <= re_search (&function_regexp, line, len, 0, len, NULL))
-	{
-	  find_function_last_match = i;
-	  return line;
-	}
+      if (0 <= re_search (&function_regexp, line, len, 0, len, nullptr))
+        {
+          find_function_last_match = i;
+          return line;
+        }
     }
   /* If we search back to where we started searching the previous time,
      find the line we found last time.  */
   if (find_function_last_match != LIN_MAX)
     return linbuf[find_function_last_match];
 
-  return NULL;
+  return nullptr;
 }

src/die.h

@@ -1,31 +0,0 @@
-/* Report an error and exit.
-   Copyright 2016-2017 Free Software Foundation, Inc.
-
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3, or (at your option)
-   any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program; if not, write to the Free Software
-   Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
-   02110-1301, USA.  */
-
-#ifndef DIE_H
-# define DIE_H
-
-# include <error.h>
-# include <stdbool.h>
-# include <verify.h>
-
-/* Like 'error (STATUS, ...)', except STATUS must be a nonzero constant.
-   This may pacify the compiler or help it generate better code.  */
-# define die(status, ...) \
-  verify_expr (status, (error (status, __VA_ARGS__), assume (false)))
-
-#endif /* DIE_H */

src/diff.h

@@ -1,7 +1,7 @@
 /* Shared definitions for GNU DIFF
 
    Copyright (C) 1988-1989, 1991-1995, 1998, 2001-2002, 2004, 2009-2013,
-   2015-2017 Free Software Foundation, Inc.
+   2015-2026 Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
 
@@ -23,6 +23,12 @@
 #include <stdio.h>
 #include <unlocked-io.h>
 
+_GL_INLINE_HEADER_BEGIN
+
+#ifndef DIFF_INLINE
+# define DIFF_INLINE _GL_INLINE
+#endif
+
 /* What kind of changes a hunk contains.  */
 enum changes
 {
@@ -54,12 +60,6 @@ enum colors_style
 
 /* Variables for command line options */
 
-#ifndef GDIFF_MAIN
-# define XTERN extern
-#else
-# define XTERN
-#endif
-
 enum output_style
 {
   /* No output style specified.  */
@@ -91,27 +91,31 @@ enum output_style
 };
 
 /* True for output styles that are robust,
-   i.e. can handle a file that ends in a non-newline.  */
-#define ROBUST_OUTPUT_STYLE(S) ((S) != OUTPUT_ED && (S) != OUTPUT_FORWARD_ED)
+   i.e. can handle a file that ends in a non-newline.
+   This is indented unusually to pacify 'make syntax-check'.  */
+DIFF_INLINE bool robust_output_style (enum output_style s)
+{
+  return s != OUTPUT_ED && s != OUTPUT_FORWARD_ED;
+}
 
-XTERN enum output_style output_style;
+extern enum output_style output_style;
 
 /* Define the current color context used to print a line.  */
-XTERN enum colors_style colors_style;
+extern enum colors_style colors_style;
 
 /* Nonzero if output cannot be generated for identical files.  */
-XTERN bool no_diff_means_no_output;
+extern bool no_diff_means_no_output;
 
 /* Number of lines of context to show in each set of diffs.
    This is zero when context is not to be shown.  */
-XTERN lin context;
+extern lin context;
 
 /* Consider all files as text files (-a).
    Don't interpret codes over 0177 as implying a "binary file".  */
-XTERN bool text;
+extern bool text;
 
 /* Number of lines to keep in identical prefix and suffix.  */
-XTERN lin horizon_lines;
+extern lin horizon_lines;
 
 /* The significance of white space during comparisons.  */
 enum DIFF_white_space
@@ -136,100 +140,107 @@ enum DIFF_white_space
   /* Ignore all horizontal white space (-w).  */
   IGNORE_ALL_SPACE
 };
-XTERN enum DIFF_white_space ignore_white_space;
+extern enum DIFF_white_space ignore_white_space;
 
 /* Ignore changes that affect only blank lines (-B).  */
-XTERN bool ignore_blank_lines;
+extern bool ignore_blank_lines;
 
 /* Files can be compared byte-by-byte, as if they were binary.
    This depends on various options.  */
-XTERN bool files_can_be_treated_as_binary;
+extern bool files_can_be_treated_as_binary;
 
 /* Ignore differences in case of letters (-i).  */
-XTERN bool ignore_case;
+extern bool ignore_case;
 
 /* Ignore differences in case of letters in file names.  */
-XTERN bool ignore_file_name_case;
+extern bool ignore_file_name_case;
 
 /* Act on symbolic links themselves rather than on their target
    (--no-dereference).  */
-XTERN bool no_dereference_symlinks;
+extern bool no_dereference_symlinks;
+
+/* Local timezone for 'c' output headers, if needed.  */
+#if HAVE_TM_GMTOFF
+# define localtz 0 /* Placeholder since localtz is never needed.  */
+#else
+extern timezone_t localtz;
+#endif
 
 /* File labels for '-c' output headers (--label).  */
-XTERN char *file_label[2];
+extern char *file_label[2];
 
 /* Regexp to identify function-header lines (-F).  */
-XTERN struct re_pattern_buffer function_regexp;
+extern struct re_pattern_buffer function_regexp;
 
 /* Ignore changes that affect only lines matching this regexp (-I).  */
-XTERN struct re_pattern_buffer ignore_regexp;
+extern struct re_pattern_buffer ignore_regexp;
 
 /* Say only whether files differ, not how (-q).  */
-XTERN bool brief;
+extern bool brief;
 
 /* Expand tabs in the output so the text lines up properly
    despite the characters added to the front of each line (-t).  */
-XTERN bool expand_tabs;
+extern bool expand_tabs;
 
 /* Number of columns between tab stops.  */
-XTERN size_t tabsize;
+extern intmax_t tabsize;
 
 /* Use a tab in the output, rather than a space, before the text of an
    input line, so as to keep the proper alignment in the input line
    without changing the characters in it (-T).  */
-XTERN bool initial_tab;
+extern bool initial_tab;
 
 /* Do not output an initial space or tab before the text of an empty line.  */
-XTERN bool suppress_blank_empty;
+extern bool suppress_blank_empty;
 
 /* Remove trailing carriage returns from input.  */
-XTERN bool strip_trailing_cr;
+extern bool strip_trailing_cr;
 
 /* In directory comparison, specify file to start with (-S).
    This is used for resuming an aborted comparison.
    All file names less than this name are ignored.  */
-XTERN char const *starting_file;
+extern char const *starting_file;
 
 /* Pipe each file's output through pr (-l).  */
-XTERN bool paginate;
+extern bool paginate;
 
 /* Line group formats for unchanged, old, new, and changed groups.  */
-XTERN char const *group_format[CHANGED + 1];
+extern char const *group_format[CHANGED + 1];
 
 /* Line formats for unchanged, old, and new lines.  */
-XTERN char const *line_format[NEW + 1];
+extern char const *line_format[NEW + 1];
 
 /* If using OUTPUT_SDIFF print extra information to help the sdiff filter.  */
-XTERN bool sdiff_merge_assist;
+extern bool sdiff_merge_assist;
 
 /* Tell OUTPUT_SDIFF to show only the left version of common lines.  */
-XTERN bool left_column;
+extern bool left_column;
 
 /* Tell OUTPUT_SDIFF to not show common lines.  */
-XTERN bool suppress_common_lines;
+extern bool suppress_common_lines;
 
 /* The half line width and column 2 offset for OUTPUT_SDIFF.  */
-XTERN size_t sdiff_half_width;
-XTERN size_t sdiff_column2_offset;
+extern intmax_t sdiff_half_width;
+extern intmax_t sdiff_column2_offset;
 
 /* String containing all the command options diff received,
    with spaces between and at the beginning but none at the end.
    If there were no options given, this string is empty.  */
-XTERN char *switch_string;
+extern char *switch_string;
 
 /* Use heuristics for better speed with large files with a small
    density of changes.  */
-XTERN bool speed_large_files;
+extern bool speed_large_files;
 
 /* Patterns that match file names to be excluded.  */
-XTERN struct exclude *excluded;
+extern struct exclude *excluded;
 
 /* Don't discard lines.  This makes things slower (sometimes much
    slower) but will find a guaranteed minimal set of changes.  */
-XTERN bool minimal;
+extern bool minimal;
 
 /* The strftime format to use for time strings.  */
-XTERN char const *time_format;
+extern char const *time_format;
 
 /* The result of comparison is an "edit script": a chain of 'struct change'.
    Each 'struct change' represents one place where some lines are deleted
@@ -254,30 +265,57 @@ struct change
 
 /* Structures that describe the input files.  */
 
+/* Directory entry types.  Like dirent DT_* macros, but portable and
+   safe to use as 'char'.  Use the same values as GNU/Linux, as this
+   may help compilers on that platform.  */
+enum detype
+  {
+    DE_UNKNOWN,
+    DE_FIFO,
+    DE_CHR,
+    DE_DIR = 4,
+    DE_BLK = 6,
+    DE_REG = 8,
+    DE_LNK = 10,
+    DE_SOCK = 12,
+    DE_WHT = 14,
+
+    /* This one is not in GNU/Linux; it means the directory entry
+       type has been determined but is none of the above.  */
+    DE_OTHER
+  };
+
 /* Data on one input file being compared.  */
 
 struct file_data {
     int             desc;	/* File descriptor  */
+    int             openerr;	/* openat errno, or 0  */
+    int             err;	/* openat or fstatat or fstat errno, or 0  */
     char const      *name;	/* File name  */
+    char const      *filetype;	/* file type as untranslated string  */
     struct stat     stat;	/* File status */
 
+    /* Directory stream corresponding to DESC, if it has been fdopendir'ed.
+       Null otherwise.  */
+    DIR *dirstream;
+
     /* Buffer in which text of file is read.  */
     word *buffer;
 
     /* Allocated size of buffer, in bytes.  Always a multiple of
        sizeof *buffer.  */
-    size_t bufsize;
+    idx_t bufsize;
 
     /* Number of valid bytes now in the buffer.  */
-    size_t buffered;
+    idx_t buffered;
 
     /* Array of pointers to lines in the file.  */
     char const **linbuf;
 
-    /* linbuf_base <= buffered_lines <= valid_lines <= alloc_lines.
-       linebuf[linbuf_base ... buffered_lines - 1] are possibly differing.
-       linebuf[linbuf_base ... valid_lines - 1] contain valid data.
-       linebuf[linbuf_base ... alloc_lines - 1] are allocated.  */
+    /* linbuf_base <= 0 <= buffered_lines <= valid_lines <= alloc_lines.
+       linbuf[0 ... buffered_lines - 1] are possibly differing.
+       linbuf[linbuf_base ... valid_lines - 1] contain valid data.
+       linbuf[linbuf_base ... alloc_lines - 1] are allocated.  */
     lin linbuf_base, buffered_lines, valid_lines, alloc_lines;
 
     /* Pointer to end of prefix of this file to ignore when hashing.  */
@@ -307,9 +345,9 @@ struct file_data {
     lin nondiscarded_lines;
 
     /* Vector, indexed by real origin-0 line number,
-       containing 1 for a line that is an insertion or a deletion.
+       containing true for a line that is an insertion or a deletion.
        The results of comparison are stored here.  */
-    char *changed;
+    bool *changed;
 
     /* 1 if file ends in a line with no final newline.  */
     bool missing_newline;
@@ -322,25 +360,38 @@ struct file_data {
     lin equiv_max;
 };
 
-/* The file buffer, considered as an array of bytes rather than
-   as an array of words.  */
-#define FILE_BUFFER(f) ((char *) (f)->buffer)
+/* struct file_data.desc markers.
+   A top level parent directory desc can be AT_FDCWD;
+   it is OK if AT_FDCWD is one of these other values.  */
+enum { OPEN_FAILED = -1 }; /* open was attempted but failed */
+enum { NONEXISTENT = -2 }; /* nonexistent file */
+enum { UNOPENED = -3 }; /* unopened file (e.g., file type mismatch) */
 
 /* Data on two input files being compared.  */
 
 struct comparison
   {
+    /* The two files.  */
     struct file_data file[2];
-    struct comparison const *parent;  /* parent, if a recursive comparison */
+
+    /* The parent comparison, or &noparent if at the top level.  */
+    struct comparison const *parent;
   };
 
 /* Describe the two files currently being compared.  */
 
-XTERN struct file_data files[2];
+extern struct comparison curr;
+
+/* A placeholder for the parent of the top level comparison.
+   Only the desc slots are used; although they are typically AT_FDCWD,
+   one might be nonnegative for a directory at the top level
+   for 'diff DIR FILE' or 'diff FILE DIR'.  */
+
+extern struct comparison noparent;
 
 /* Stdio stream to output diffs to.  */
 
-XTERN FILE *outfile;
+extern FILE *outfile;
 
 /* Declare various functions.  */
 
@@ -348,14 +399,20 @@ XTERN FILE *outfile;
 extern int diff_2_files (struct comparison *);
 
 /* context.c */
-extern void print_context_header (struct file_data[], char const * const *, bool);
+extern void print_context_header (struct file_data[],
+				  char const *const *, bool);
 extern void print_context_script (struct change *, bool);
 
+/* diff.c */
+extern int compare_files (struct comparison const *, enum detype const[2],
+			  char const *, char const *);
+
 /* dir.c */
-extern int diff_dirs (struct comparison const *,
-                      int (*) (struct comparison const *,
-                               char const *, char const *));
-extern char *find_dir_file_pathname (char const *, char const *);
+extern int diff_dirs (struct comparison *);
+extern char *find_dir_file_pathname (struct file_data *, char const *,
+				     enum detype *)
+  ATTRIBUTE_MALLOC ATTRIBUTE_DEALLOC_FREE
+  ATTRIBUTE_RETURNS_NONNULL;
 
 /* ed.c */
 extern void print_ed_script (struct change *);
@@ -365,7 +422,7 @@ extern void pr_forward_ed_script (struct change *);
 extern void print_ifdef_script (struct change *);
 
 /* io.c */
-extern void file_block_read (struct file_data *, size_t);
+extern void file_block_read (struct file_data *, idx_t);
 extern bool read_files (struct file_data[], bool);
 
 /* normal.c */
@@ -380,33 +437,28 @@ extern void print_sdiff_script (struct change *);
 /* util.c */
 extern char const change_letter[4];
 extern char const pr_program[];
-extern char *concat (char const *, char const *, char const *);
-extern bool lines_differ (char const *, char const *) _GL_ATTRIBUTE_PURE;
-extern lin translate_line_number (struct file_data const *, lin);
-extern struct change *find_change (struct change *);
-extern struct change *find_reverse_change (struct change *);
-extern void *zalloc (size_t);
+extern lin translate_line_number (struct file_data const *, lin)
+  ATTRIBUTE_PURE;
+extern struct change *find_change (struct change *) ATTRIBUTE_CONST;
 extern enum changes analyze_hunk (struct change *, lin *, lin *, lin *, lin *);
 extern void begin_output (void);
+extern void cleanup_signal_handlers (void);
 extern void debug_script (struct change *);
-extern void fatal (char const *) __attribute__((noreturn));
+extern _Noreturn void fatal (char const *);
 extern void finish_output (void);
-extern void message (char const *, char const *, char const *);
-extern void message5 (char const *, char const *, char const *,
-                      char const *, char const *);
+extern void message (char const *, ...) ATTRIBUTE_FORMAT ((printf, 1, 2));
 extern void output_1_line (char const *, char const *, char const *,
                            char const *);
 extern void perror_with_name (char const *);
-extern void pfatal_with_name (char const *) __attribute__((noreturn));
-extern void print_1_line (char const *, char const * const *);
-extern void print_1_line_nl (char const *, char const * const *, bool);
+extern _Noreturn void pfatal_with_name (char const *);
+extern void print_1_line (char const *, char const *const *);
+extern void print_1_line_nl (char const *, char const *const *, bool);
 extern void print_message_queue (void);
 extern void print_number_range (char, struct file_data *, lin, lin);
 extern void print_script (struct change *, struct change * (*) (struct change *),
                           void (*) (struct change *));
 extern void setup_output (char const *, char const *, bool);
-extern void translate_range (struct file_data const *, lin, lin,
-                             printint *, printint *);
+extern void translate_range (struct file_data const *, lin, lin, lin *, lin *);
 
 enum color_context
 {
@@ -417,7 +469,9 @@ enum color_context
   LINE_NUMBER_CONTEXT,
 };
 
-XTERN bool presume_output_tty;
+extern bool presume_output_tty;
 
 extern void set_color_context (enum color_context color_context);
-extern void set_color_palette (char const *palette);
+extern void set_color_palette (char *palette);
+
+_GL_INLINE_HEADER_END

src/dir.c

@@ -1,7 +1,7 @@
 /* Read, sort and compare two directories.  Used for GNU DIFF.
 
    Copyright (C) 1988-1989, 1992-1995, 1998, 2001-2002, 2004, 2006-2007,
-   2009-2013, 2015-2017 Free Software Foundation, Inc.
+   2009-2013, 2015-2026 Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
 
@@ -19,20 +19,28 @@
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
 #include "diff.h"
+
+#include <diagnose.h>
+#include <dirname.h>
 #include <error.h>
 #include <exclude.h>
 #include <filenamecat.h>
+#include <mcel.h>
+#include <quote.h>
 #include <setjmp.h>
 #include <xalloc.h>
 
-/* Read the directory named by DIR and store into DIRDATA a sorted vector
-   of filenames for its contents.  DIR->desc == -1 means this directory is
-   known to be nonexistent, so set DIRDATA to an empty vector.
-   Return -1 (setting errno) if error, 0 otherwise.  */
+#ifndef HAVE_STRUCT_DIRENT_D_TYPE
+# define HAVE_STRUCT_DIRENT_D_TYPE false
+#endif
+
+/* A sorted vector of file names obtained by reading a directory.
+   If HAVE_STRUCT_DIRENT_D_TYPE, each name is preceded by a byte
+   giving the file type as an enum detype value.  */
 
 struct dirdata
 {
-  size_t nnames;	/* Number of names.  */
+  idx_t nnames;	/* Number of names.  */
   char const **names;	/* Sorted names of files in dir, followed by 0.  */
   char *data;	/* Allocated storage for file names.  */
 };
@@ -44,98 +52,132 @@ static bool locale_specific_sorting;
 /* Where to go if locale-specific sorting fails.  */
 static jmp_buf failed_locale_specific_sorting;
 
+static int compare_names (char const *, char const *);
 static bool dir_loop (struct comparison const *, int);
 
 
-/* Read a directory and get its vector of names.  */
+/* Given the parent directory PARENTDIRFD (negative for current dir),
+   read the directory named by DIR and store into DIRDATA a sorted
+   vector of filenames for its contents.
+   Use DIR's basename if PARENTDIRFD is nonnegative, for efficiency.
+   If DIR->desc == NONEXISTENT, this directory is known to be
+   nonexistent so set DIRDATA to an empty vector;
+   otherwise, update DIR->desc and DIR->dirstream as needed.
+   If STARTFILE, ignore directory entries less than STARTFILE, and if
+   STARTFILE_ONLY, also ignore directory entries greater than STARTFILE.
+   Return true if successful, false (setting errno) otherwise.  */
 
 static bool
-dir_read (struct file_data const *dir, struct dirdata *dirdata)
+dir_read (int parentdirfd, struct file_data *dir, struct dirdata *dirdata,
+	  char const *startfile, bool startfile_only)
 {
-  register struct dirent *next;
-  register size_t i;
-
-  /* Address of block containing the files that are described.  */
-  char const **names;
-
   /* Number of files in directory.  */
-  size_t nnames;
+  idx_t nnames = 0;
 
   /* Allocated and used storage for file name data.  */
   char *data;
-  size_t data_alloc, data_used;
 
-  dirdata->names = 0;
-  dirdata->data = 0;
-  nnames = 0;
-  data = 0;
+  dirdata->names = nullptr;
+  dirdata->data = nullptr;
 
-  if (dir->desc != -1)
+  if (dir->desc != NONEXISTENT)
     {
       /* Open the directory and check for errors.  */
-      register DIR *reading = opendir (dir->name);
+      int dirfd = dir->desc;
+      if (dirfd < 0)
+	{
+	  dirfd = openat (parentdirfd,
+			  (parentdirfd < 0 ? dir->name
+			   : last_component (dir->name)),
+			  (O_RDONLY | O_CLOEXEC | O_DIRECTORY
+			   | (no_dereference_symlinks ? O_NOFOLLOW : 0)));
+	  if (dirfd < 0)
+	    return false;
+	  dir->desc = dirfd;
+	}
+      DIR *reading = fdopendir (dirfd);
       if (!reading)
-	return false;
+        return false;
+      dir->dirstream = reading;
 
       /* Initialize the table of filenames.  */
 
-      data_alloc = 512;
-      data_used = 0;
-      dirdata->data = data = xmalloc (data_alloc);
+      idx_t data_alloc = 512;
+      idx_t data_used = 0;
+      dirdata->data = data = ximalloc (data_alloc);
 
       /* Read the directory entries, and insert the subfiles
-	 into the 'data' table.  */
+         into the 'data' table.  */
 
-      while ((errno = 0, (next = readdir (reading)) != 0))
-	{
-	  char *d_name = next->d_name;
-	  size_t d_size = _D_EXACT_NAMLEN (next) + 1;
+      while (true)
+        {
+	  errno = 0;
+	  struct dirent *next = readdir (reading);
+	  if (!next)
+	    break;
 
-	  /* Ignore "." and "..".  */
-	  if (d_name[0] == '.'
-	      && (d_name[1] == 0 || (d_name[1] == '.' && d_name[2] == 0)))
-	    continue;
+          char *d_name = next->d_name;
 
-	  if (excluded_file_name (excluded, d_name))
-	    continue;
+          /* Ignore "." and "..".  */
+          if (d_name[0] == '.'
+              && (d_name[1] == 0 || (d_name[1] == '.' && d_name[2] == 0)))
+            continue;
 
-	  while (data_alloc < data_used + d_size)
+	  if (startfile)
 	    {
-	      if (PTRDIFF_MAX / 2 <= data_alloc)
-		xalloc_die ();
-	      dirdata->data = data = xrealloc (data, data_alloc *= 2);
+	      int cmp = compare_names (d_name, startfile);
+	      if (cmp < 0 || (startfile_only && !!cmp))
+		continue;
 	    }
 
-	  memcpy (data + data_used, d_name, d_size);
-	  data_used += d_size;
-	  nnames++;
-	}
-      if (errno)
-	{
-	  int e = errno;
-	  closedir (reading);
-	  errno = e;
-	  return false;
-	}
-#if CLOSEDIR_VOID
-      closedir (reading);
-#else
-      if (closedir (reading) != 0)
-	return false;
+          if (excluded_file_name (excluded, d_name))
+            continue;
+
+	  idx_t d_size = HAVE_STRUCT_DIRENT_D_TYPE + _D_EXACT_NAMLEN (next) + 1;
+          if (data_alloc - data_used < d_size)
+	    dirdata->data = data
+	      = xpalloc (data, &data_alloc,
+			 d_size - (data_alloc - data_used), -1, 1);
+#if HAVE_STRUCT_DIRENT_D_TYPE
+	  char detype;
+	  switch (next->d_type)
+	    {
+	    case DT_BLK:  detype = DE_BLK;  break;
+	    case DT_CHR:  detype = DE_CHR;  break;
+	    case DT_DIR:  detype = DE_DIR;  break;
+	    case DT_FIFO: detype = DE_FIFO; break;
+	    case DT_LNK:  detype = DE_LNK;  break;
+	    case DT_REG:  detype = DE_REG;  break;
+	    case DT_SOCK: detype = DE_SOCK; break;
+# ifdef DT_WHT
+	    case DT_WHT:  detype = DE_WHT;  break;
+# endif
+	    case DT_UNKNOWN: detype = DE_UNKNOWN; break;
+	    default:         detype = DE_OTHER;   break;
+	    }
+	  data[data_used++] = detype;
+	  d_size--;
 #endif
+          memcpy (data + data_used, d_name, d_size);
+          data_used += d_size;
+          nnames++;
+        }
+
+      if (errno)
+	return false;
     }
 
   /* Create the 'names' table from the 'data' table.  */
-  if (PTRDIFF_MAX / sizeof *names - 1 <= nnames)
-    xalloc_die ();
-  dirdata->names = names = xmalloc ((nnames + 1) * sizeof *names);
+  char const **names = xinmalloc (nnames + 1, sizeof *names);
+  dirdata->names = names;
   dirdata->nnames = nnames;
-  for (i = 0;  i < nnames;  i++)
+  for (idx_t i = 0; i < nnames; i++)
     {
+      data += HAVE_STRUCT_DIRENT_D_TYPE;
       names[i] = data;
       data += strlen (data) + 1;
     }
-  names[nnames] = 0;
+  names[nnames] = nullptr;
   return true;
 }
 
@@ -145,16 +187,12 @@ dir_read (struct file_data const *dir, struct dirdata *dirdata)
 static int
 compare_collated (char const *name1, char const *name2)
 {
-  int r;
   errno = 0;
-  if (ignore_file_name_case)
-    r = strcasecoll (name1, name2);
-  else
-    r = strcoll (name1, name2);
+  int r = strcoll (name1, name2);
   if (errno)
     {
-      error (0, errno, _("cannot compare file names '%s' and '%s'"),
-	     name1, name2);
+      error (0, errno, _("cannot compare file names %s and %s"),
+	     quote_n (0, name1), quote_n (1, name2));
       longjmp (failed_locale_specific_sorting, 1);
     }
   return r;
@@ -165,11 +203,14 @@ compare_collated (char const *name1, char const *name2)
 static int
 compare_names (char const *name1, char const *name2)
 {
+  if (ignore_file_name_case)
+    return mbscasecmp (name1, name2);  /* Best we can do.  */
+
   if (locale_specific_sorting)
     {
       int diff = compare_collated (name1, name2);
-      if (diff || ignore_file_name_case)
-	return diff;
+      if (diff)
+        return diff;
     }
   return file_name_cmp (name1, name2);
 }
@@ -182,138 +223,108 @@ compare_names_for_qsort (void const *file1, void const *file2)
 {
   char const *const *f1 = file1;
   char const *const *f2 = file2;
-  char const *name1 = *f1;
-  char const *name2 = *f2;
-  if (locale_specific_sorting)
-    {
-      int diff = compare_collated (name1, name2);
-      if (diff)
-	return diff;
-    }
-  return file_name_cmp (name1, name2);
+  return compare_names (*f1, *f2);
 }
 
 /* Compare the contents of two directories named in CMP.
    This is a top-level routine; it does everything necessary for diff
    on two directories.
 
-   CMP->file[0].desc == -1 says directory CMP->file[0] doesn't exist,
-   but pretend it is empty.  Likewise for CMP->file[1].
+   If CMP->file[0].desc == NONEXISTENT, directory CMP->file[0] doesn't exist
+   and pretend it is empty.  Otherwise, update CMP->file[0].desc and
+   CMP->file[0].dirstream as needed.  Likewise for CMP->file[1].
 
-   HANDLE_FILE is a caller-provided subroutine called to handle each file.
-   It gets three operands: CMP, name of file in dir 0, name of file in dir 1.
-   These names are relative to the original working directory.
-
-   For a file that appears in only one of the dirs, one of the name-args
-   to HANDLE_FILE is zero.
-
-   Returns the maximum of all the values returned by HANDLE_FILE,
+   Returns the maximum of all the values returned by compare_files,
    or EXIT_TROUBLE if trouble is encountered in opening files.  */
 
 int
-diff_dirs (struct comparison const *cmp,
-	   int (*handle_file) (struct comparison const *,
-			       char const *, char const *))
+diff_dirs (struct comparison *cmp)
 {
-  struct dirdata dirdata[2];
-  int volatile val = EXIT_SUCCESS;
-  int i;
-
-  if ((cmp->file[0].desc == -1 || dir_loop (cmp, 0))
-      && (cmp->file[1].desc == -1 || dir_loop (cmp, 1)))
+  if ((cmp->file[0].desc == NONEXISTENT || dir_loop (cmp, 0))
+      && (cmp->file[1].desc == NONEXISTENT || dir_loop (cmp, 1)))
     {
       error (0, 0, _("%s: recursive directory loop"),
-	     cmp->file[cmp->file[0].desc == -1].name);
+	     squote (0, cmp->file[cmp->file[0].desc == NONEXISTENT].name));
       return EXIT_TROUBLE;
     }
 
   /* Get contents of both dirs.  */
-  for (i = 0; i < 2; i++)
-    if (! dir_read (&cmp->file[i], &dirdata[i]))
+  struct dirdata dirdata[2];
+  int val = EXIT_SUCCESS;
+  for (int i = 0; i < 2; i++)
+    if (! dir_read (cmp->parent->file[i].desc, &cmp->file[i], &dirdata[i],
+		    cmp->parent == &noparent ? starting_file : nullptr, false))
       {
-	perror_with_name (cmp->file[i].name);
-	val = EXIT_TROUBLE;
+        perror_with_name (cmp->file[i].name);
+        val = EXIT_TROUBLE;
       }
 
   if (val == EXIT_SUCCESS)
     {
-      char const **volatile names[2];
-      names[0] = dirdata[0].names;
-      names[1] = dirdata[1].names;
-
       /* Use locale-specific sorting if possible, else native byte order.  */
       locale_specific_sorting = true;
-      if (setjmp (failed_locale_specific_sorting))
-	locale_specific_sorting = false;
+      if (! ignore_file_name_case)
+	if (setjmp (failed_locale_specific_sorting))
+	  locale_specific_sorting = false;
 
       /* Sort the directories.  */
-      for (i = 0; i < 2; i++)
-	qsort (names[i], dirdata[i].nnames, sizeof *dirdata[i].names,
-	       compare_names_for_qsort);
-
-      /* If '-S name' was given, and this is the topmost level of comparison,
-	 ignore all file names less than the specified starting name.  */
-
-      if (starting_file && ! cmp->parent)
-	{
-	  while (*names[0] && compare_names (*names[0], starting_file) < 0)
-	    names[0]++;
-	  while (*names[1] && compare_names (*names[1], starting_file) < 0)
-	    names[1]++;
-	}
+      for (int i = 0; i < 2; i++)
+        qsort (dirdata[i].names, dirdata[i].nnames, sizeof *dirdata[i].names,
+               compare_names_for_qsort);
 
       /* Loop while files remain in one or both dirs.  */
-      while (*names[0] || *names[1])
-	{
-	  /* Compare next name in dir 0 with next name in dir 1.
-	     At the end of a dir,
-	     pretend the "next name" in that dir is very large.  */
-	  int nameorder = (!*names[0] ? 1 : !*names[1] ? -1
-			   : compare_names (*names[0], *names[1]));
+      char const **n0 = dirdata[0].names;
+      char const **n1 = dirdata[1].names;
+      while (*n0 || *n1)
+        {
+          /* Compare next name in dir 0 with next name in dir 1.
+             At the end of a dir,
+             pretend the "next name" in that dir is very large.  */
+          int nameorder = !*n0 ? 1 : !*n1 ? -1 : compare_names (*n0, *n1);
 
-	  /* Prefer a file_name_cmp match if available.  This algorithm is
-	     O(N**2), where N is the number of names in a directory
-	     that compare_names says are all equal, but in practice N
-	     is so small it's not worth tuning.  */
-	  if (nameorder == 0 && ignore_file_name_case)
-	    {
-	      int raw_order = file_name_cmp (*names[0], *names[1]);
-	      if (raw_order != 0)
-		{
-		  int greater_side = raw_order < 0;
-		  int lesser_side = 1 - greater_side;
-		  char const **lesser = names[lesser_side];
-		  char const *greater_name = *names[greater_side];
-		  char const **p;
+          /* Prefer a file_name_cmp match if available.  This algorithm is
+             O(N**2), where N is the number of names in a directory
+             that compare_names says are all equal, but in practice N
+             is so small it's not worth tuning.  */
+          if (nameorder == 0 && ignore_file_name_case)
+            {
+              int raw_order = file_name_cmp (*n0, *n1);
+              if (raw_order != 0)
+                {
+                  char const **lesser = raw_order < 0 ? n0 : n1;
+                  char const *greater_name = *(raw_order < 0 ? n1 : n0);
 
-		  for (p = lesser + 1;
-		       *p && compare_names (*p, greater_name) == 0;
-		       p++)
-		    {
-		      int c = file_name_cmp (*p, greater_name);
-		      if (0 <= c)
-			{
-			  if (c == 0)
-			    {
-			      memmove (lesser + 1, lesser,
-				       (char *) p - (char *) lesser);
-			      *lesser = greater_name;
-			    }
-			  break;
-			}
-		    }
-		}
-	    }
+                  for (char const **p = lesser + 1;
+                       *p && compare_names (*p, greater_name) == 0;
+                       p++)
+                    {
+                      int c = file_name_cmp (*p, greater_name);
+                      if (0 <= c)
+                        {
+                          if (c == 0)
+                            {
+                              memmove (lesser + 1, lesser,
+                                       (char *) p - (char *) lesser);
+                              *lesser = greater_name;
+                            }
+                          break;
+                        }
+                    }
+                }
+            }
 
-	  int v1 = (*handle_file) (cmp,
-				   0 < nameorder ? 0 : *names[0]++,
-				   nameorder < 0 ? 0 : *names[1]++);
-	  if (val < v1)
-	    val = v1;
-	}
+	  enum detype detypes[]
+	    = { HAVE_STRUCT_DIRENT_D_TYPE && *n0 ? (*n0)[-1] : DE_UNKNOWN,
+		HAVE_STRUCT_DIRENT_D_TYPE && *n1 ? (*n1)[-1] : DE_UNKNOWN };
+	  int v1 = compare_files (cmp, detypes,
+				  0 < nameorder ? nullptr : *n0++,
+				  nameorder < 0 ? nullptr : *n1++);
+          if (val < v1)
+            val = v1;
+        }
     }
 
-  for (i = 0; i < 2; i++)
+  for (int i = 0; i < 2; i++)
     {
       free (dirdata[i].names);
       free (dirdata[i].data);
@@ -324,12 +335,11 @@ diff_dirs (struct comparison const *cmp,
 
 /* Return nonzero if CMP is looping recursively in argument I.  */
 
-static bool _GL_ATTRIBUTE_PURE
+static bool ATTRIBUTE_PURE
 dir_loop (struct comparison const *cmp, int i)
 {
-  struct comparison const *p = cmp;
-  while ((p = p->parent))
-    if (0 < same_file (&p->file[i].stat, &cmp->file[i].stat))
+  for (struct comparison const *p = cmp; (p = p->parent) != &noparent; )
+    if (same_file (&p->file[i].stat, &cmp->file[i].stat))
       return true;
   return false;
 }
@@ -337,48 +347,29 @@ dir_loop (struct comparison const *cmp, int i)
 /* Find a matching filename in a directory.  */
 
 char *
-find_dir_file_pathname (char const *dir, char const *file)
+find_dir_file_pathname (struct file_data *dir, char const *file,
+			enum detype *detype)
 {
-  /* The 'IF_LINT (volatile)' works around what appears to be a bug in
-     gcc 4.8.0 20120825; see
-     <http://lists.gnu.org/archive/html/bug-diffutils/2012-08/msg00007.html>.
-     */
-  char const * IF_LINT (volatile) match = file;
+  char const *match = file;
 
-  char *val;
   struct dirdata dirdata;
-  dirdata.names = NULL;
-  dirdata.data = NULL;
+  dirdata.names = nullptr;
+  dirdata.data = nullptr;
 
-  if (ignore_file_name_case)
-    {
-      struct file_data filedata;
-      filedata.name = dir;
-      filedata.desc = 0;
+  if (ignore_file_name_case && dir_read (AT_FDCWD, dir, &dirdata, file, true))
+    for (char const **p = dirdata.names; *p; p++)
+      {
+	if (file_name_cmp (*p, file) == 0)
+	  {
+	    match = *p;
+	    break;
+	  }
+	if (match == file)
+	  match = *p;
+      }
 
-      if (dir_read (&filedata, &dirdata))
-	{
-	  locale_specific_sorting = true;
-	  if (setjmp (failed_locale_specific_sorting))
-	    match = file; /* longjmp may mess up MATCH.  */
-	  else
-	    {
-	      for (char const **p = dirdata.names; *p; p++)
-		if (compare_names (*p, file) == 0)
-		  {
-		    if (file_name_cmp (*p, file) == 0)
-		      {
-			match = *p;
-			break;
-		      }
-		    if (match == file)
-		      match = *p;
-		  }
-	    }
-	}
-    }
-
-  val = file_name_concat (dir, match, NULL);
+  *detype = HAVE_STRUCT_DIRENT_D_TYPE && match != file ? match[-1] : DE_UNKNOWN;
+  char *val = file_name_concat (dir->name, match, nullptr);
   free (dirdata.names);
   free (dirdata.data);
   return val;

src/ed.c

@@ -1,7 +1,7 @@
 /* Output routines for ed-script format.
 
    Copyright (C) 1988-1989, 1991-1993, 1995, 1998, 2001, 2004, 2006, 2009-2013,
-   2015-2017 Free Software Foundation, Inc.
+   2015-2026 Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
 
@@ -29,7 +29,8 @@ static void pr_forward_ed_hunk (struct change *);
 void
 print_ed_script (struct change *script)
 {
-  print_script (script, find_reverse_change, print_ed_hunk);
+  /* The script is reversed, so plain find_change suffices.  */
+  print_script (script, find_change, print_ed_hunk);
 }
 
 /* Print a hunk of an ed diff */
@@ -37,52 +38,50 @@ print_ed_script (struct change *script)
 static void
 print_ed_hunk (struct change *hunk)
 {
-  lin f0, l0, f1, l1;
-  enum changes changes;
-
 #ifdef DEBUG
   debug_script (hunk);
 #endif
 
   /* Determine range of line numbers involved in each file.  */
-  changes = analyze_hunk (hunk, &f0, &l0, &f1, &l1);
+  lin f0, l0, f1, l1;
+  enum changes changes = analyze_hunk (hunk, &f0, &l0, &f1, &l1);
   if (!changes)
     return;
 
   begin_output ();
 
   /* Print out the line number header for this hunk */
-  print_number_range (',', &files[0], f0, l0);
+  print_number_range (',', &curr.file[0], f0, l0);
   fputc (change_letter[changes], outfile);
   fputc ('\n', outfile);
 
   /* Print new/changed lines from second file, if needed */
   if (changes != OLD)
     {
-      lin i;
       bool insert_mode = true;
 
-      for (i = f1; i <= l1; i++)
-	{
-	  if (!insert_mode)
-	    {
-	      fputs ("a\n", outfile);
-	      insert_mode = true;
-	    }
-	  if (files[1].linbuf[i][0] == '.' && files[1].linbuf[i][1] == '\n')
-	    {
-	      /* The file's line is just a dot, and it would exit
-		 insert mode.  Precede the dot with another dot, exit
-		 insert mode and remove the extra dot.  */
-	      fputs ("..\n.\ns/.//\n", outfile);
-	      insert_mode = false;
-	    }
-	  else
-	    print_1_line ("", &files[1].linbuf[i]);
-	}
+      for (lin i = f1; i <= l1; i++)
+        {
+          if (!insert_mode)
+            {
+              fputs ("a\n", outfile);
+              insert_mode = true;
+            }
+	  if (curr.file[1].linbuf[i][0] == '.'
+	      && curr.file[1].linbuf[i][1] == '\n')
+            {
+              /* The file's line is just a dot, and it would exit
+                 insert mode.  Precede the dot with another dot, exit
+                 insert mode and remove the extra dot.  */
+              fputs ("..\n.\ns/.//\n", outfile);
+              insert_mode = false;
+            }
+          else
+	    print_1_line ("", &curr.file[1].linbuf[i]);
+        }
 
       if (insert_mode)
-	fputs (".\n", outfile);
+        fputs (".\n", outfile);
     }
 }
 
@@ -101,9 +100,8 @@ pr_forward_ed_script (struct change *script)
 static void
 pr_forward_ed_hunk (struct change *hunk)
 {
-  lin i, f0, l0, f1, l1;
-
   /* Determine range of line numbers involved in each file.  */
+  lin f0, l0, f1, l1;
   enum changes changes = analyze_hunk (hunk, &f0, &l0, &f1, &l1);
   if (!changes)
     return;
@@ -111,7 +109,7 @@ pr_forward_ed_hunk (struct change *hunk)
   begin_output ();
 
   fputc (change_letter[changes], outfile);
-  print_number_range (' ', files, f0, l0);
+  print_number_range (' ', &curr.file[0], f0, l0);
   fputc ('\n', outfile);
 
   /* If deletion only, print just the number range.  */
@@ -122,8 +120,8 @@ pr_forward_ed_hunk (struct change *hunk)
   /* For insertion (with or without deletion), print the number range
      and the lines from file 2.  */
 
-  for (i = f1; i <= l1; i++)
-    print_1_line ("", &files[1].linbuf[i]);
+  for (lin i = f1; i <= l1; i++)
+    print_1_line ("", &curr.file[1].linbuf[i]);
 
   fputs (".\n", outfile);
 }
@@ -143,35 +141,34 @@ print_rcs_script (struct change *script)
 static void
 print_rcs_hunk (struct change *hunk)
 {
-  lin i, f0, l0, f1, l1;
-  printint tf0, tl0, tf1, tl1;
-
   /* Determine range of line numbers involved in each file.  */
+  lin f0, l0, f1, l1;
   enum changes changes = analyze_hunk (hunk, &f0, &l0, &f1, &l1);
   if (!changes)
     return;
 
   begin_output ();
 
-  translate_range (&files[0], f0, l0, &tf0, &tl0);
+  lin tf0, tl0, tf1, tl1;
+  translate_range (&curr.file[0], f0, l0, &tf0, &tl0);
 
   if (changes & OLD)
     {
       /* For deletion, print just the starting line number from file 0
-	 and the number of lines deleted.  */
+         and the number of lines deleted.  */
       fprintf (outfile, "d%"pI"d %"pI"d\n", tf0,
-	       tf0 <= tl0 ? tl0 - tf0 + 1 : 1);
+               tf0 <= tl0 ? tl0 - tf0 + 1 : 1);
     }
 
   if (changes & NEW)
     {
       /* Take last-line-number from file 0 and # lines from file 1.  */
-      translate_range (&files[1], f1, l1, &tf1, &tl1);
+      translate_range (&curr.file[1], f1, l1, &tf1, &tl1);
       fprintf (outfile, "a%"pI"d %"pI"d\n", tl0,
-	       tf1 <= tl1 ? tl1 - tf1 + 1 : 1);
+               tf1 <= tl1 ? tl1 - tf1 + 1 : 1);
 
       /* Print the inserted lines.  */
-      for (i = f1; i <= l1; i++)
-	print_1_line ("", &files[1].linbuf[i]);
+      for (lin i = f1; i <= l1; i++)
+	print_1_line ("", &curr.file[1].linbuf[i]);
     }
 }

src/ifdef.c

@@ -1,6 +1,6 @@
 /* #ifdef-format output routines for GNU DIFF.
 
-   Copyright (C) 1989, 1991-1994, 2001-2002, 2004, 2006, 2009-2013, 2015-2017
+   Copyright (C) 1989, 1991-1994, 2001-2002, 2004, 2006, 2009-2013, 2015-2026
    Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
@@ -22,7 +22,8 @@
 
 #include "diff.h"
 
-#include <xalloc.h>
+#include <c-ctype.h>
+#include <xmalloca.h>
 
 struct group
 {
@@ -31,10 +32,10 @@ struct group
 };
 
 static char const *format_group (FILE *, char const *, char,
-				 struct group const *);
+                                 struct group const *);
 static char const *do_printf_spec (FILE *, char const *,
-				   struct file_data const *, lin,
-				   struct group const *);
+                                   struct file_data const *, lin,
+                                   struct group const *);
 static char const *scan_char_literal (char const *, char *);
 static lin groups_letter_value (struct group const *, char);
 static void format_ifdef (char const *, lin, lin, lin, lin);
@@ -49,15 +50,15 @@ static lin next_line1;
 void
 print_ifdef_script (struct change *script)
 {
-  next_line0 = next_line1 = - files[0].prefix_lines;
+  next_line0 = next_line1 = - curr.file[0].prefix_lines;
   print_script (script, find_change, print_ifdef_hunk);
-  if (next_line0 < files[0].valid_lines
-      || next_line1 < files[1].valid_lines)
+  if (next_line0 < curr.file[0].valid_lines
+      || next_line1 < curr.file[1].valid_lines)
     {
       begin_output ();
       format_ifdef (group_format[UNCHANGED],
-		    next_line0, files[0].valid_lines,
-		    next_line1, files[1].valid_lines);
+		    next_line0, curr.file[0].valid_lines,
+		    next_line1, curr.file[1].valid_lines);
     }
 }
 
@@ -68,9 +69,8 @@ print_ifdef_script (struct change *script)
 static void
 print_ifdef_hunk (struct change *hunk)
 {
-  lin first0, last0, first1, last1;
-
   /* Determine range of line numbers involved in each file.  */
+  lin first0, last0, first1, last1;
   enum changes changes = analyze_hunk (hunk, &first0, &last0, &first1, &last1);
   if (!changes)
     return;
@@ -80,15 +80,15 @@ print_ifdef_hunk (struct change *hunk)
   /* Print lines up to this change.  */
   if (next_line0 < first0 || next_line1 < first1)
     format_ifdef (group_format[UNCHANGED],
-		  next_line0, first0,
-		  next_line1, first1);
+                  next_line0, first0,
+                  next_line1, first1);
 
   /* Print this change.  */
   next_line0 = last0 + 1;
   next_line1 = last1 + 1;
   format_ifdef (group_format[changes],
-		first0, next_line0,
-		first1, next_line1);
+                first0, next_line0,
+                first1, next_line1);
 }
 
 /* Print a set of lines according to FORMAT.
@@ -98,15 +98,10 @@ print_ifdef_hunk (struct change *hunk)
 static void
 format_ifdef (char const *format, lin beg0, lin end0, lin beg1, lin end1)
 {
-  struct group groups[2];
-
-  groups[0].file = &files[0];
-  groups[0].from = beg0;
-  groups[0].upto = end0;
-  groups[1].file = &files[1];
-  groups[1].from = beg1;
-  groups[1].upto = end1;
-  format_group (outfile, format, 0, groups);
+  format_group (outfile, format, '\0',
+		((struct group const[])
+		 {{.file = &curr.file[0], .from = beg0, .upto = end0},
+		  {.file = &curr.file[1], .from = beg1, .upto = end1}}));
 }
 
 /* Print to file OUT a set of lines according to FORMAT.
@@ -116,91 +111,91 @@ format_ifdef (char const *format, lin beg0, lin end0, lin beg1, lin end1)
    If OUT is zero, do not actually print anything; just scan the format.  */
 
 static char const *
-format_group (register FILE *out, char const *format, char endchar,
-	      struct group const *groups)
+format_group (FILE *out, char const *format, char endchar,
+              struct group const *groups)
 {
-  register char c;
-  register char const *f = format;
+  char const *f = format;
 
-  while ((c = *f) != endchar && c != 0)
+  for (char c; (c = *f) != endchar && c; )
     {
       char const *f1 = ++f;
       if (c == '%')
-	switch ((c = *f++))
-	  {
-	  case '%':
-	    break;
+       {
+	c = *f++;
+	switch (c)
+          {
+          case '%':
+            break;
 
-	  case '(':
-	    /* Print if-then-else format e.g. '%(n=1?thenpart:elsepart)'.  */
-	    {
-	      int i;
-	      uintmax_t value[2];
-	      FILE *thenout, *elseout;
+          case '(':
+            /* Print if-then-else format e.g. '%(n=1?thenpart:elsepart)'.  */
+            {
+              intmax_t value[2];
 
-	      for (i = 0; i < 2; i++)
-		{
-		  if (ISDIGIT (*f))
-		    {
-		      char *fend;
-		      errno = 0;
-		      value[i] = strtoumax (f, &fend, 10);
-		      if (errno)
-			goto bad_format;
-		      f = fend;
-		    }
-		  else
-		    {
-		      value[i] = groups_letter_value (groups, *f);
-		      if (value[i] == -1)
-			goto bad_format;
-		      f++;
-		    }
-		  if (*f++ != "=?"[i])
-		    goto bad_format;
-		}
-	      if (value[0] == value[1])
-		thenout = out, elseout = 0;
-	      else
-		thenout = 0, elseout = out;
-	      f = format_group (thenout, f, ':', groups);
-	      if (*f)
-		{
-		  f = format_group (elseout, f + 1, ')', groups);
-		  if (*f)
-		    f++;
-		}
-	    }
-	    continue;
+              for (int i = 0; i < 2; i++)
+                {
+                  if (c_isdigit (*f))
+                    {
+                      char *fend;
+                      errno = 0;
+                      value[i] = strtoimax (f, &fend, 10);
+                      if (errno)
+                        goto bad_format;
+                      f = fend;
+                    }
+                  else
+                    {
+                      value[i] = groups_letter_value (groups, *f);
+                      if (value[i] < 0)
+                        goto bad_format;
+                      f++;
+                    }
+                  if (*f++ != "=?"[i])
+                    goto bad_format;
+                }
 
-	  case '<':
-	    /* Print lines deleted from first file.  */
-	    print_ifdef_lines (out, line_format[OLD], &groups[0]);
-	    continue;
+	      bool equal_values = value[0] == value[1];
+              FILE *thenout = equal_values ? out : nullptr;
+	      FILE *elseout = equal_values ? nullptr : out;
+              f = format_group (thenout, f, ':', groups);
+              if (*f)
+                {
+                  f = format_group (elseout, f + 1, ')', groups);
+                  if (*f)
+                    f++;
+                }
+            }
+            continue;
 
-	  case '=':
-	    /* Print common lines.  */
-	    print_ifdef_lines (out, line_format[UNCHANGED], &groups[0]);
-	    continue;
+          case '<':
+            /* Print lines deleted from first file.  */
+            print_ifdef_lines (out, line_format[OLD], &groups[0]);
+            continue;
 
-	  case '>':
-	    /* Print lines inserted from second file.  */
-	    print_ifdef_lines (out, line_format[NEW], &groups[1]);
-	    continue;
+          case '=':
+            /* Print common lines.  */
+            print_ifdef_lines (out, line_format[UNCHANGED], &groups[0]);
+            continue;
 
-	  default:
-	    f = do_printf_spec (out, f - 2, 0, 0, groups);
-	    if (f)
-	      continue;
-	    /* Fall through. */
-	  bad_format:
-	    c = '%';
-	    f = f1;
-	    break;
-	  }
+          case '>':
+            /* Print lines inserted from second file.  */
+            print_ifdef_lines (out, line_format[NEW], &groups[1]);
+            continue;
+
+          default:
+            f = do_printf_spec (out, f - 2, nullptr, 0, groups);
+            if (f)
+              continue;
+            /* Fall through. */
+          bad_format:
+            c = '%';
+            f = f1;
+            break;
+          }
+       }
 
       if (out)
-	putc (c, out);
+        putc (c, out);
     }
 
   return f;
@@ -234,154 +229,149 @@ groups_letter_value (struct group const *g, char letter)
 /* Print to file OUT, using FORMAT to print the line group GROUP.
    But do nothing if OUT is zero.  */
 static void
-print_ifdef_lines (register FILE *out, char const *format,
-		   struct group const *group)
+print_ifdef_lines (FILE *out, char const *format,
+                   struct group const *group)
 {
-  struct file_data const *file = group->file;
-  char const * const *linbuf = file->linbuf;
-  lin from = group->from, upto = group->upto;
-
   if (!out)
     return;
 
+  struct file_data const *file = group->file;
+  char const *const *linbuf = file->linbuf;
+  lin from = group->from, upto = group->upto;
+
   /* If possible, use a single fwrite; it's faster.  */
   if (!expand_tabs && format[0] == '%')
     {
       if (format[1] == 'l' && format[2] == '\n' && !format[3] && from < upto)
-	{
-	  fwrite (linbuf[from], sizeof (char),
-		  linbuf[upto] + (linbuf[upto][-1] != '\n') -  linbuf[from],
-		  out);
-	  return;
-	}
+        {
+          fwrite (linbuf[from], sizeof (char),
+                  linbuf[upto] + (linbuf[upto][-1] != '\n') -  linbuf[from],
+                  out);
+          return;
+        }
       if (format[1] == 'L' && !format[2])
-	{
-	  fwrite (linbuf[from], sizeof (char),
-		  linbuf[upto] -  linbuf[from], out);
-	  return;
-	}
+        {
+          fwrite (linbuf[from], sizeof (char),
+                  linbuf[upto] -  linbuf[from], out);
+          return;
+        }
     }
 
   for (;  from < upto;  from++)
     {
-      register char c;
-      register char const *f = format;
+      char c;
+      char const *f = format;
 
-      while ((c = *f++) != 0)
-	{
-	  char const *f1 = f;
-	  if (c == '%')
-	    switch ((c = *f++))
-	      {
-	      case '%':
-		break;
+      while ((c = *f++))
+        {
+          char const *f1 = f;
+          if (c == '%')
+	   {
+	    c = *f++;
+	    switch (c)
+              {
+              case '%':
+                break;
 
-	      case 'l':
-		output_1_line (linbuf[from],
-			       (linbuf[from + 1]
-				- (linbuf[from + 1][-1] == '\n')),
-			       0, 0);
-		continue;
+              case 'l':
+                output_1_line (linbuf[from],
+                               (linbuf[from + 1]
+                                - (linbuf[from + 1][-1] == '\n')),
+                               nullptr, nullptr);
+                continue;
 
-	      case 'L':
-		output_1_line (linbuf[from], linbuf[from + 1], 0, 0);
-		continue;
+              case 'L':
+                output_1_line (linbuf[from], linbuf[from + 1],
+			       nullptr, nullptr);
+                continue;
 
-	      default:
-		f = do_printf_spec (out, f - 2, file, from, 0);
-		if (f)
-		  continue;
-		c = '%';
-		f = f1;
-		break;
-	      }
+              default:
+                f = do_printf_spec (out, f - 2, file, from, nullptr);
+                if (f)
+                  continue;
+                c = '%';
+                f = f1;
+                break;
+              }
+	   }
 
-	  putc (c, out);
-	}
+          putc (c, out);
+        }
     }
 }
 
 static char const *
 do_printf_spec (FILE *out, char const *spec,
-		struct file_data const *file, lin n,
-		struct group const *groups)
+                struct file_data const *file, lin n,
+                struct group const *groups)
 {
   char const *f = spec;
   char c;
-  char c1;
 
   /* Scan printf-style SPEC of the form %[-'0]*[0-9]*(.[0-9]*)?[cdoxX].  */
-  /* assert (*f == '%'); */
+  dassert (*f == '%');
   f++;
   while ((c = *f++) == '-' || c == '\'' || c == '0')
     continue;
-  while (ISDIGIT (c))
+  while (c_isdigit (c))
     c = *f++;
   if (c == '.')
-    while (ISDIGIT (c = *f++))
+    while (c_isdigit (c = *f++))
       continue;
-  c1 = *f++;
+  char c1 = *f++;
 
   switch (c)
     {
     case 'c':
       if (c1 != '\'')
-	return 0;
+        return nullptr;
       else
-	{
-	  char value IF_LINT (= 0);
-	  f = scan_char_literal (f, &value);
-	  if (!f)
-	    return 0;
-	  if (out)
-	    putc (value, out);
-	}
+        {
+          char value;
+          f = scan_char_literal (f, &value);
+          if (!f)
+            return nullptr;
+          if (out)
+            putc (value, out);
+        }
       break;
 
     case 'd': case 'o': case 'x': case 'X':
       {
-	lin value;
+        lin value;
 
-	if (file)
-	  {
-	    if (c1 != 'n')
-	      return 0;
-	    value = translate_line_number (file, n);
-	  }
-	else
-	  {
-	    value = groups_letter_value (groups, c1);
-	    if (value < 0)
-	      return 0;
-	  }
+        if (file)
+          {
+            if (c1 != 'n')
+              return nullptr;
+            value = translate_line_number (file, n);
+          }
+        else
+          {
+            value = groups_letter_value (groups, c1);
+            if (value < 0)
+              return nullptr;
+          }
 
-	if (out)
-	  {
-	    /* For example, if the spec is "%3xn" and pI is "l", use the printf
-	       format spec "%3lx".  Here the spec prefix is "%3".  */
-	    printint print_value = value;
-	    size_t spec_prefix_len = f - spec - 2;
-	    size_t pI_len = sizeof pI - 1;
-#if 0
-	    char format[spec_prefix_len + pI_len + 2];
-#else
-	    char *format = xmalloc (spec_prefix_len + pI_len + 2);
-#endif
-	    char *p = format + spec_prefix_len + pI_len;
-	    memcpy (format, spec, spec_prefix_len);
-	    memcpy (format + spec_prefix_len, pI, pI_len);
-	    *p++ = c;
-	    *p = '\0';
-	    fprintf (out, format, print_value);
-#if ! HAVE_C_VARARRAYS
-	    free (format);
-#endif
-	  }
+        if (out)
+          {
+            /* For example, if the spec is "%3xn" and pI is "l", use the printf
+               format spec "%3lx".  Here the spec prefix is "%3".  */
+            idx_t spec_prefix_len = f - spec - 2;
+            idx_t pI_len = sizeof pI - 1;
+            char *format = xmalloca (spec_prefix_len + pI_len + 2);
+            char *p = mempcpy (format, spec, spec_prefix_len);
+            p = stpcpy (p, pI);
+            *p++ = c;
+            *p = '\0';
+            fprintf (out, format, value);
+            freea (format);
+          }
       }
       break;
 
     default:
-      return 0;
+      return nullptr;
     }
 
   return f;
@@ -394,36 +384,35 @@ do_printf_spec (FILE *out, char const *spec,
 static char const *
 scan_char_literal (char const *lit, char *valptr)
 {
-  register char const *p = lit;
+  char const *p = lit;
   char value;
-  ptrdiff_t digits;
   char c = *p++;
 
   switch (c)
     {
       case 0:
       case '\'':
-	return NULL;
+        return nullptr;
 
       case '\\':
-	value = 0;
-	while ((c = *p++) != '\'')
-	  {
-	    unsigned int digit = c - '0';
-	    if (8 <= digit)
-	      return NULL;
-	    value = 8 * value + digit;
-	  }
-	digits = p - lit - 2;
-	if (! (1 <= digits && digits <= 3))
-	  return NULL;
-	break;
+        value = '\0';
+        while ((c = *p++) != '\'')
+          {
+            unsigned int digit = c - '0';
+            if (8 <= digit)
+              return nullptr;
+            value = 8 * value + digit;
+          }
+        ptrdiff_t digits = p - lit - 2;
+        if (! (1 <= digits && digits <= 3))
+          return nullptr;
+        break;
 
       default:
-	value = c;
-	if (*p++ != '\'')
-	  return NULL;
-	break;
+        value = c;
+        if (*p++ != '\'')
+          return nullptr;
+        break;
     }
 
   *valptr = value;

src/normal.c

@@ -1,6 +1,6 @@
 /* Normal-format output routines for GNU DIFF.
 
-   Copyright (C) 1988-1989, 1993, 1995, 1998, 2001, 2006, 2009-2013, 2015-2017
+   Copyright (C) 1988-1989, 1993, 1995, 1998, 2001, 2006, 2009-2013, 2015-2026
    Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
@@ -38,10 +38,8 @@ print_normal_script (struct change *script)
 static void
 print_normal_hunk (struct change *hunk)
 {
-  lin first0, last0, first1, last1;
-  register lin i;
-
   /* Determine range of line numbers involved in each file.  */
+  lin first0, last0, first1, last1;
   enum changes changes = analyze_hunk (hunk, &first0, &last0, &first1, &last1);
   if (!changes)
     return;
@@ -50,23 +48,21 @@ print_normal_hunk (struct change *hunk)
 
   /* Print out the line number header for this hunk */
   set_color_context (LINE_NUMBER_CONTEXT);
-  print_number_range (',', &files[0], first0, last0);
+  print_number_range (',', &curr.file[0], first0, last0);
   fputc (change_letter[changes], outfile);
-  print_number_range (',', &files[1], first1, last1);
+  print_number_range (',', &curr.file[1], first1, last1);
   set_color_context (RESET_CONTEXT);
   fputc ('\n', outfile);
 
   /* Print the lines that the first file has.  */
   if (changes & OLD)
     {
-      if (first0 <= last0)
-        set_color_context (DELETE_CONTEXT);
-      for (i = first0; i <= last0; i++)
+      for (lin i = first0; i <= last0; i++)
         {
-          print_1_line_nl ("<", &files[0].linbuf[i], true);
-          if (i == last0)
-            set_color_context (RESET_CONTEXT);
-          if (files[0].linbuf[i + 1][-1] == '\n')
+          set_color_context (DELETE_CONTEXT);
+	  print_1_line_nl ("<", &curr.file[0].linbuf[i], true);
+          set_color_context (RESET_CONTEXT);
+	  if (curr.file[0].linbuf[i + 1][-1] == '\n')
             putc ('\n', outfile);
         }
     }
@@ -77,14 +73,12 @@ print_normal_hunk (struct change *hunk)
   /* Print the lines that the second file has.  */
   if (changes & NEW)
     {
-      if (first1 <= last1)
-        set_color_context (ADD_CONTEXT);
-      for (i = first1; i <= last1; i++)
+      for (lin i = first1; i <= last1; i++)
         {
-          print_1_line_nl (">", &files[1].linbuf[i], true);
-          if (i == last1)
-            set_color_context (RESET_CONTEXT);
-          if (files[1].linbuf[i + 1][-1] == '\n')
+          set_color_context (ADD_CONTEXT);
+	  print_1_line_nl (">", &curr.file[1].linbuf[i], true);
+          set_color_context (RESET_CONTEXT);
+	  if (curr.file[1].linbuf[i + 1][-1] == '\n')
             putc ('\n', outfile);
         }
     }

src/side.c

@@ -1,6 +1,6 @@
 /* sdiff-format output routines for GNU DIFF.
 
-   Copyright (C) 1991-1993, 1998, 2001-2002, 2004, 2009-2013, 2015-2017 Free
+   Copyright (C) 1991-1993, 1998, 2001-2002, 2004, 2009-2013, 2015-2026 Free
    Software Foundation, Inc.
 
    This file is part of GNU DIFF.
@@ -22,7 +22,7 @@
 
 #include "diff.h"
 
-#include <wchar.h>
+#include <mcel.h>
 
 static void print_sdiff_common_lines (lin, lin);
 static void print_sdiff_hunk (struct change *);
@@ -37,157 +37,169 @@ print_sdiff_script (struct change *script)
 {
   begin_output ();
 
-  next0 = next1 = - files[0].prefix_lines;
+  next0 = next1 = - curr.file[0].prefix_lines;
   print_script (script, find_change, print_sdiff_hunk);
 
-  print_sdiff_common_lines (files[0].valid_lines, files[1].valid_lines);
+  print_sdiff_common_lines (curr.file[0].valid_lines,
+			    curr.file[1].valid_lines);
 }
 
 /* Tab from column FROM to column TO, where FROM <= TO.  Yield TO.  */
 
-static size_t
-tab_from_to (size_t from, size_t to)
+static intmax_t
+tab_from_to (intmax_t from, intmax_t to)
 {
   FILE *out = outfile;
-  size_t tab;
-  size_t tab_size = tabsize;
 
   if (!expand_tabs)
-    for (tab = from + tab_size - from % tab_size;  tab <= to;  tab += tab_size)
-      {
-	putc ('\t', out);
-	from = tab;
-      }
+    {
+      intmax_t tab_size = tabsize;
+      for (intmax_t tab = from + tab_size - from % tab_size;
+	   tab <= to;  tab += tab_size)
+	{
+	  putc ('\t', out);
+	  from = tab;
+	}
+    }
   while (from++ < to)
     putc (' ', out);
   return to;
 }
 
 /* Print the text for half an sdiff line.  This means truncate to
-   width observing tabs, and trim a trailing newline.  Return the
-   last column written (not the number of chars).  */
+   OUT_BOUND columns, observing tabs, and trim a trailing newline.
+   Return the presumed column position on the output device after
+   the write (not the number of chars).  */
 
-static size_t
-print_half_line (char const *const *line, size_t indent, size_t out_bound)
+static intmax_t
+print_half_line (char const *const *line, intmax_t indent, intmax_t out_bound)
 {
   FILE *out = outfile;
-  register size_t in_position = 0;
-  register size_t out_position = 0;
-  register char const *text_pointer = line[0];
-  register char const *text_limit = line[1];
-  mbstate_t mbstate = { 0 };
+  /* IN_POSITION is the current column position if we were outputting the
+     entire line, i.e. ignoring OUT_BOUND.  */
+  intmax_t in_position = 0;
+  /* OUT_POSITION is the current column position.  It stays <= OUT_BOUND
+     at any moment.  */
+  intmax_t out_position = 0;
+  char const *text_pointer = line[0];
+  char const *text_limit = line[1];
 
   while (text_pointer < text_limit)
     {
       char const *tp0 = text_pointer;
-      register char c = *text_pointer++;
+      char c = *text_pointer++;
 
       switch (c)
-	{
-	case '\t':
-	  {
-	    size_t spaces = tabsize - in_position % tabsize;
-	    if (in_position == out_position)
+        {
+        case '\t':
+          {
+            intmax_t spaces = tabsize - in_position % tabsize;
+	    intmax_t tabstop;
+	    if (ckd_add (&tabstop, in_position, spaces))
+	      return out_position;
+            if (in_position == out_position)
+              {
+                if (expand_tabs)
+                  {
+                    if (out_bound < tabstop)
+                      tabstop = out_bound;
+                    for (;  out_position < tabstop;  out_position++)
+                      putc (' ', out);
+                  }
+                else
+                  if (tabstop < out_bound)
+                    {
+                      out_position = tabstop;
+                      putc (c, out);
+                    }
+              }
+	    in_position = tabstop;
+          }
+          break;
+
+        case '\r':
+          {
+            putc (c, out);
+            tab_from_to (0, indent);
+            in_position = out_position = 0;
+          }
+          break;
+
+        case '\b':
+          if (in_position != 0 && --in_position < out_bound)
+            {
+              if (out_position <= in_position)
+                /* Add spaces to make up for suppressed tab past out_bound.  */
+                for (;  out_position < in_position;  out_position++)
+                  putc (' ', out);
+              else
+                {
+                  out_position = in_position;
+                  putc (c, out);
+                }
+            }
+          break;
+
+        default:
+          {
+	    /* A byte that might start a multibyte character.
+	       Increase TEXT_POINTER, counting columns.
+	       Assume encoding errors have print width 1.  */
+	    mcel_t g = mcel_scan (tp0, text_limit);
+	    int width = g.err ? 1 : c32width (g.ch);
+	    if (0 < width && ckd_add (&in_position, in_position, width))
+	      return out_position;
+
+	    /* If there is room, output the bytes since TP0.  */
+	    if (in_position <= out_bound)
 	      {
-		size_t tabstop = out_position + spaces;
-		if (expand_tabs)
-		  {
-		    if (out_bound < tabstop)
-		      tabstop = out_bound;
-		    for (;  out_position < tabstop;  out_position++)
-		      putc (' ', out);
-		  }
-		else
-		  if (tabstop < out_bound)
-		    {
-		      out_position = tabstop;
-		      putc (c, out);
-		    }
+		out_position = in_position;
+		fwrite (tp0, 1, g.len, out);
 	      }
-	    in_position += spaces;
-	  }
-	  break;
 
-	case '\r':
-	  {
-	    putc (c, out);
-	    tab_from_to (0, indent);
-	    in_position = out_position = 0;
-	  }
-	  break;
+	    text_pointer = tp0 + g.len;
+          }
+          break;
 
-	case '\b':
-	  if (in_position != 0 && --in_position < out_bound)
-	    {
-	      if (out_position <= in_position)
-		/* Add spaces to make up for suppressed tab past out_bound.  */
-		for (;  out_position < in_position;  out_position++)
-		  putc (' ', out);
-	      else
-		{
-		  out_position = in_position;
-		  putc (c, out);
-		}
-	    }
-	  break;
-
-	default:
-	  {
-	    wchar_t wc;
-	    size_t bytes = mbrtowc (&wc, tp0, text_limit - tp0, &mbstate);
-
-	    if (0 < bytes && bytes < (size_t) -2)
-	      {
-		int width = wcwidth (wc);
-		if (0 < width)
-		  in_position += width;
-		if (in_position <= out_bound)
-		  {
-		    out_position = in_position;
-		    fwrite (tp0, 1, bytes, stdout);
-		  }
-		text_pointer = tp0 + bytes;
-		break;
-	      }
-	  }
-	  FALLTHROUGH;
-	case '\f':
-	case '\v':
-	  if (in_position < out_bound)
-	    putc (c, out);
-	  break;
-
-	case ' ': case '!': case '"': case '#': case '%':
-	case '&': case '\'': case '(': case ')': case '*':
-	case '+': case ',': case '-': case '.': case '/':
-	case '0': case '1': case '2': case '3': case '4':
-	case '5': case '6': case '7': case '8': case '9':
-	case ':': case ';': case '<': case '=': case '>':
-	case '?':
-	case 'A': case 'B': case 'C': case 'D': case 'E':
-	case 'F': case 'G': case 'H': case 'I': case 'J':
-	case 'K': case 'L': case 'M': case 'N': case 'O':
-	case 'P': case 'Q': case 'R': case 'S': case 'T':
-	case 'U': case 'V': case 'W': case 'X': case 'Y':
-	case 'Z':
-	case '[': case '\\': case ']': case '^': case '_':
-	case 'a': case 'b': case 'c': case 'd': case 'e':
-	case 'f': case 'g': case 'h': case 'i': case 'j':
-	case 'k': case 'l': case 'm': case 'n': case 'o':
-	case 'p': case 'q': case 'r': case 's': case 't':
-	case 'u': case 'v': case 'w': case 'x': case 'y':
-	case 'z': case '{': case '|': case '}': case '~':
-	  /* These characters are printable ASCII characters.  */
-	  if (in_position++ < out_bound)
+        /* Print width 1.  */
+        case ' ': case '!': case '"': case '#': case '$': case '%':
+        case '&': case '\'': case '(': case ')': case '*':
+        case '+': case ',': case '-': case '.': case '/':
+        case '0': case '1': case '2': case '3': case '4':
+        case '5': case '6': case '7': case '8': case '9':
+        case ':': case ';': case '<': case '=': case '>':
+        case '?': case '@':
+        case 'A': case 'B': case 'C': case 'D': case 'E':
+        case 'F': case 'G': case 'H': case 'I': case 'J':
+        case 'K': case 'L': case 'M': case 'N': case 'O':
+        case 'P': case 'Q': case 'R': case 'S': case 'T':
+        case 'U': case 'V': case 'W': case 'X': case 'Y':
+        case 'Z':
+        case '[': case '\\': case ']': case '^': case '_': case '`':
+        case 'a': case 'b': case 'c': case 'd': case 'e':
+        case 'f': case 'g': case 'h': case 'i': case 'j':
+        case 'k': case 'l': case 'm': case 'n': case 'o':
+        case 'p': case 'q': case 'r': case 's': case 't':
+        case 'u': case 'v': case 'w': case 'x': case 'y':
+        case 'z': case '{': case '|': case '}': case '~':
+	  if (ckd_add (&in_position, in_position, 1))
+	    return out_position;
+	  if (in_position <= out_bound)
 	    {
 	      out_position = in_position;
 	      putc (c, out);
 	    }
 	  break;
 
-	case '\n':
-	  return out_position;
-	}
+	/* Print width 0.  */
+	case '\0': case '\a': case '\f': case '\v':
+	  if (in_position <= out_bound)
+	    putc (c, out);
+	  break;
+
+        case '\n':
+          return out_position;
+        }
     }
 
   return out_position;
@@ -199,12 +211,12 @@ print_half_line (char const *const *line, size_t indent, size_t out_bound)
 
 static void
 print_1sdiff_line (char const *const *left, char sep,
-		   char const *const *right)
+                   char const *const *right)
 {
   FILE *out = outfile;
-  size_t hw = sdiff_half_width;
-  size_t c2o = sdiff_column2_offset;
-  size_t col = 0;
+  intmax_t hw = sdiff_half_width;
+  intmax_t c2o = sdiff_column2_offset;
+  intmax_t col = 0;
   bool put_newline = false;
   bool color_to_reset = false;
 
@@ -227,9 +239,9 @@ print_1sdiff_line (char const *const *left, char sep,
 
   if (sep != ' ')
     {
-      col = tab_from_to (col, (hw + c2o - 1) / 2) + 1;
+      col = tab_from_to (col, (hw + c2o - 1) >> 1) + 1;
       if (sep == '|' && put_newline != (right[1][-1] == '\n'))
-	sep = put_newline ? '/' : '\\';
+        sep = put_newline ? '/' : '\\';
       putc (sep, out);
     }
 
@@ -237,10 +249,10 @@ print_1sdiff_line (char const *const *left, char sep,
     {
       put_newline |= right[1][-1] == '\n';
       if (**right != '\n')
-	{
-	  col = tab_from_to (col, c2o);
-	  print_half_line (right, col, hw);
-	}
+        {
+          col = tab_from_to (col, c2o);
+          print_half_line (right, col, hw);
+        }
     }
 
   if (put_newline)
@@ -259,22 +271,18 @@ print_sdiff_common_lines (lin limit0, lin limit1)
   if (!suppress_common_lines && (i0 != limit0 || i1 != limit1))
     {
       if (sdiff_merge_assist)
-	{
-	  printint len0 = limit0 - i0;
-	  printint len1 = limit1 - i1;
-	  fprintf (outfile, "i%"pI"d,%"pI"d\n", len0, len1);
-	}
+	fprintf (outfile, "i%"pI"d,%"pI"d\n", limit0 - i0, limit1 - i1);
 
       if (!left_column)
-	{
-	  while (i0 != limit0 && i1 != limit1)
-	    print_1sdiff_line (&files[0].linbuf[i0++], ' ',
-			       &files[1].linbuf[i1++]);
-	  while (i1 != limit1)
-	    print_1sdiff_line (0, ')', &files[1].linbuf[i1++]);
-	}
+        {
+          while (i0 != limit0 && i1 != limit1)
+	    print_1sdiff_line (&curr.file[0].linbuf[i0++], ' ',
+			       &curr.file[1].linbuf[i1++]);
+          while (i1 != limit1)
+	    print_1sdiff_line (0, ')', &curr.file[1].linbuf[i1++]);
+        }
       while (i0 != limit0)
-	print_1sdiff_line (&files[0].linbuf[i0++], '(', 0);
+	print_1sdiff_line (&curr.file[0].linbuf[i0++], '(', 0);
     }
 
   next0 = limit0;
@@ -288,10 +296,8 @@ print_sdiff_common_lines (lin limit0, lin limit1)
 static void
 print_sdiff_hunk (struct change *hunk)
 {
-  lin first0, last0, first1, last1;
-  register lin i, j;
-
   /* Determine range of line numbers involved in each file.  */
+  lin first0, last0, first1, last1;
   enum changes changes =
     analyze_hunk (hunk, &first0, &last0, &first1, &last1);
   if (!changes)
@@ -301,17 +307,17 @@ print_sdiff_hunk (struct change *hunk)
   print_sdiff_common_lines (first0, first1);
 
   if (sdiff_merge_assist)
-    {
-      printint len0 = last0 - first0 + 1;
-      printint len1 = last1 - first1 + 1;
-      fprintf (outfile, "c%"pI"d,%"pI"d\n", len0, len1);
-    }
+    fprintf (outfile, "c%"pI"d,%"pI"d\n",
+	     last0 - first0 + 1,
+	     last1 - first1 + 1);
 
   /* Print "xxx  |  xxx " lines.  */
   if (changes == CHANGED)
     {
+      lin i, j;
       for (i = first0, j = first1;  i <= last0 && j <= last1;  i++, j++)
-	print_1sdiff_line (&files[0].linbuf[i], '|', &files[1].linbuf[j]);
+	print_1sdiff_line (&curr.file[0].linbuf[i], '|',
+			   &curr.file[1].linbuf[j]);
       changes = (i <= last0 ? OLD : 0) + (j <= last1 ? NEW : 0);
       next0 = first0 = i;
       next1 = first1 = j;
@@ -320,16 +326,18 @@ print_sdiff_hunk (struct change *hunk)
   /* Print "     >  xxx " lines.  */
   if (changes & NEW)
     {
+      lin j;
       for (j = first1; j <= last1; ++j)
-	print_1sdiff_line (0, '>', &files[1].linbuf[j]);
+	print_1sdiff_line (0, '>', &curr.file[1].linbuf[j]);
       next1 = j;
     }
 
   /* Print "xxx  <     " lines.  */
   if (changes & OLD)
     {
+      lin i;
       for (i = first0; i <= last0; ++i)
-	print_1sdiff_line (&files[0].linbuf[i], '<', 0);
+	print_1sdiff_line (&curr.file[0].linbuf[i], '<', 0);
       next0 = i;
     }
 }

src/syncsig.c

@@ -0,0 +1,350 @@
+/* Synchronous signal handling
+
+   Copyright 2025-2026 Free Software Foundation, Inc.
+
+   This file is part of GNU DIFF.
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#include <config.h>
+
+#include "syncsig.h"
+
+#include "minmax.h"
+
+#include <signal.h>
+#include <stdcountof.h>
+
+/* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
+   present.  */
+#ifndef SA_NOCLDSTOP
+# define SA_NOCLDSTOP 0
+# define sigprocmask(how, set, oset) 0
+# if ! HAVE_SIGINTERRUPT
+#  define siginterrupt(sig, flag) 0
+# endif
+#endif
+
+#ifndef SA_RESTART
+# define SA_RESTART 0
+#endif
+#ifndef SIGSTOP
+# define SIGSTOP 0
+#endif
+
+/* The set of signals that are caught.  */
+static sigset_t caught_signals;
+
+/* The signals we can catch.
+   This includes all catchable GNU/Linux or POSIX signals that by
+   default are ignored, or that stop or terminate the process.
+   It also includes SIGQUIT since that can come from the terminal.
+   It excludes other signals that normally come from program failure.
+   If you modify this table, also modify signal_count's initializer.  */
+static int const catchable[] =
+  {
+    /* SIGABRT is normally from program failure.  */
+    SIGALRM,
+    /* SIGBUS is normally from program failure.  */
+#ifdef SIGCHLD
+    SIGCHLD,
+#else
+# define SIGCHLD 0
+#endif
+    /* SIGCLD (older platforms) is an alias for SIGCHLD.  */
+#ifdef SIGCONT
+    SIGCONT,
+#else
+# define SIGCONT 0
+#endif
+    /* SIGEMT is normally from program failure.  */
+    /* SIGFPE is normally from program failure.  */
+    SIGHUP,
+    /* SIGILL is normally from program failure.  */
+    /* SIGINFO is an alias for SIGPWR on Linux.  */
+    SIGINT,
+    /* SIGIO is an alias for SIGPOLL.  */
+    /* SIGKILL can't be caught.  */
+#ifdef SIGLOST
+    SIGLOST,
+#else
+# define SIGLOST 0
+#endif
+    SIGPIPE,
+#ifdef SIGPOLL /* Removed from POSIX.1-2024; still in Linux.  */
+    SIGPOLL,
+#else
+# define SIGPOLL 0
+#endif
+#ifdef SIGPROF /* Removed from POSIX.1-2024; still in Linux.  */
+    SIGPROF,
+#else
+# define SIGPROF 0
+#endif
+#ifdef SIGPWR
+    SIGPWR,
+#else
+# define SIGPWR 0
+#endif
+    SIGQUIT,
+    /* SIGSEGV is normally from program failure.  */
+    /* Linux SIGSTKFLT is unused.  */
+    /* SIGSTOP can't be caught.  */
+    /* SIGSYS is normally from program failure.  */
+    SIGTERM,
+    /* SIGTRAP is normally from program failure.  */
+#ifdef SIGTSTP
+    SIGTSTP,
+#else
+# define SIGTSTP 0
+#endif
+#ifdef SIGTTIN
+    SIGTTIN,
+#else
+# define SIGTTIN 0
+#endif
+#ifdef SIGTTOU
+    SIGTTOU,
+#else
+# define SIGTTOU 0
+#endif
+#ifdef SIGURG
+    SIGURG,
+#else
+# define SIGURG 0
+#endif
+#ifdef SIGUSR1
+    SIGUSR1,
+#else
+# define SIGUSR1 0
+#endif
+#ifdef SIGUSR2
+    SIGUSR2,
+#else
+# define SIGUSR2 0
+#endif
+#ifdef SIGVTALRM
+    SIGVTALRM,
+#else
+# define SIGVTALRM 0
+#endif
+#ifdef SIGWINCH
+    SIGWINCH,
+#else
+# define SIGWINCH 0
+#endif
+#ifdef SIGXCPU
+    SIGXCPU,
+#else
+# define SIGXCPU 0
+#endif
+#ifdef SIGXFSZ
+    SIGXFSZ,
+#else
+# define SIGXFSZ 0
+#endif
+  };
+
+/* Number of pending signals received, for each signal type.  */
+static sig_atomic_t volatile signal_count[] =
+  {
+    /* Explicitly initialize, so that the table is large enough.  */
+    [SIGALRM] = 0,
+    [SIGCHLD] = 0,
+    [SIGCONT] = 0,
+    [SIGHUP] = 0,
+    [SIGINT] = 0,
+    [SIGLOST] = 0,
+    [SIGPIPE] = 0,
+    [SIGPOLL] = 0,
+    [SIGPROF] = 0,
+    [SIGPWR] = 0,
+    [SIGQUIT] = 0,
+    [SIGTERM] = 0,
+    [SIGTSTP] = 0,
+    [SIGTTIN] = 0,
+    [SIGTTOU] = 0,
+    [SIGURG] = 0,
+    [SIGUSR1] = 0,
+    [SIGUSR2] = 0,
+    [SIGVTALRM] = 0,
+    [SIGWINCH] = 0,
+    [SIGXCPU] = 0,
+    [SIGXFSZ] = 0,
+  };
+
+/* This acts as bool though its type is sig_atomic_t.
+   If true, signal_count might contain nonzero entries.
+   If false, signal_count is all zero.
+   This is to speed up syncsig_poll when it returns 0.  */
+static sig_atomic_t volatile possible_signal_count;
+
+/* Actions before syncsig_install was called.  */
+#if SA_NOCLDSTOP
+static struct sigaction oldact[countof (signal_count)];
+#else
+static void (*oldact[countof (signal_count)]) (int);
+#endif
+
+/* Record an asynchronous signal.  This function is async-signal-safe.  */
+static void
+sighandler (int sig)
+{
+#if !SA_NOCLDSTOP
+  /* An unavoidable race here: the default action might mistakenly
+     be taken before 'signal' is called.  */
+  signal (sig, sighandler);
+#endif
+
+  possible_signal_count = true;
+  signal_count[sig]++;
+}
+
+void
+syncsig_install (int flags)
+{
+  for (int i = 0; i < countof (catchable); i++)
+    signal_count[catchable[i]] = 0;
+  possible_signal_count = false;
+
+  sigemptyset (&caught_signals);
+
+  for (int i = 0; i < countof (catchable); i++)
+    {
+      int sig = catchable[i];
+
+      if (((sig == SIGTSTP) & !(flags & SYNCSIG_TSTP))
+	  | ((sig == SIGTTIN) & !(flags & SYNCSIG_TTIN))
+	  | ((sig == SIGTTOU) & !(flags & SYNCSIG_TTOU)))
+	continue;
+
+#if SA_NOCLDSTOP
+      sigaction (sig, nullptr, &oldact[sig]);
+      bool catching_sig = oldact[sig].sa_handler != SIG_IGN;
+#else
+      oldact[i] = signal (sig, SIG_IGN);
+      bool catching_sig = oldact[i] != SIG_IGN;
+      if (catching_sig)
+	{
+	  /* An unavoidable race here: SIG might be mistakenly ignored
+	     before 'signal' is called.  */
+	  signal (sig, sighandler);
+	  siginterrupt (sig, 0);
+	}
+#endif
+      if (catching_sig)
+	sigaddset (&caught_signals, sig);
+    }
+
+#if SA_NOCLDSTOP
+  struct sigaction act;
+  act.sa_handler = sighandler;
+  act.sa_mask = caught_signals;
+  act.sa_flags = SA_RESTART;
+
+  for (int i = 0; i < countof (catchable); i++)
+    {
+      int sig = catchable[i];
+      if (sigismember (&caught_signals, sig))
+	sigaction (sig, &act, nullptr);
+    }
+#endif
+}
+
+void
+syncsig_uninstall (void)
+{
+  for (int i = 0; i < countof (catchable); i++)
+    {
+      int sig = catchable[i];
+      if (sigismember (&caught_signals, sig))
+	{
+#if SA_NOCLDSTOP
+	  sigaction (sig, &oldact[sig], nullptr);
+#else
+	  signal (sig, oldact[sig]);
+#endif
+	}
+    }
+}
+
+int
+syncsig_poll (void)
+{
+  int sig = 0;
+
+  if (possible_signal_count)
+    {
+      /* This module uses static rather than thread-local storage,
+	 so it is useful only in single-threaded programs,
+	 and it uses sigprocmask rather than pthread_sigmask.  */
+      sigset_t oldset;
+      sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
+
+      for (int i = 0; ; i++)
+	{
+	  if (i == countof (catchable))
+	    {
+	      possible_signal_count = false;
+	      break;
+	    }
+	  int s = catchable[i];
+	  int c = signal_count[s];
+	  if (c)
+	    {
+	      signal_count[s] = c - 1;
+	      sig = s;
+	      break;
+	    }
+	}
+
+      sigprocmask (SIG_SETMASK, &oldset, nullptr);
+    }
+
+  return sig;
+}
+
+void
+syncsig_deliver (int sig)
+{
+#if SA_NOCLDSTOP
+  struct sigaction act;
+#else
+  void (*act) (int);
+#endif
+
+  if (sig == SIGTSTP)
+    sig = SIGSTOP;
+  else
+    {
+#if SA_NOCLDSTOP
+      sigaction (sig, &oldact[sig], &act);
+#else
+      act = signal (sig, oldact[sig]);
+#endif
+
+    }
+
+  raise (sig);
+
+  if (sig != SIGSTOP)
+    {
+      /* The program did not exit due to the raised signal, so continue.  */
+#if SA_NOCLDSTOP
+      sigaction (sig, &act, nullptr);
+#else
+      signal (sig, act);
+#endif
+    }
+}

src/syncsig.h

@@ -0,0 +1,60 @@
+/* Synchronous signal handling
+
+   Copyright 2025-2026 Free Software Foundation, Inc.
+
+   This file is part of GNU DIFF.
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+/* Written by Paul Eggert.  */
+
+/* Flags for syncsig_install.  */
+enum
+  {
+    /* Also catch SIGTSTP, SIGTTIN, SIGTTOU, which by default stop the process.
+       These flags have no effect on platforms lacking these signals.  */
+    SYNCSIG_TSTP = (1 << 0),
+    SYNCSIG_TTIN = (1 << 1),
+    SYNCSIG_TTOU = (1 << 2),
+  };
+
+/* Set up asynchronous signal handling according to FLAGS.
+   syncsig_install fails only on unusual platforms where
+   valid calls to functions like sigaction can fail;
+   if it fails, signal handling is in a weird state
+   and neither syncsig_process nor syncsig_uninstall should be called.  */
+extern void syncsig_install (int flags);
+
+/* Return a signal number if a signal has arrived, zero otherwise.
+   After syncsig_install, there should not be an unbounded amount of
+   time between calls to this function, and its result should be dealt
+   with promptly.  */
+extern int syncsig_poll (void);
+
+/* Do the action for SIG that would have been done
+   had syncsig_install not been called.
+   SIG should have recently been returned by syncsig_poll.
+
+   For example, if SIG is SIGTSTP stop the process and return after
+   SIGCONT is delivered.  Another example: kill the process if SIG is
+   SIGINT and SIGINT handling is the default.
+
+   This function should be called only after syncsig_install.  */
+extern void syncsig_deliver (int sig);
+
+/* Stop doing asynchronous signal handling, undoing syncsig_install.
+   This function should be called only after syncsig_install.
+   To deal with signals arriving just before calling this function,
+   call syncsig_poll afterwards.  */
+extern void syncsig_uninstall (void);

src/system.c

@@ -0,0 +1,163 @@
+/* System dependent declarations.
+
+   Copyright (C) 1988-1989, 1992-1995, 1998, 2001-2002, 2004, 2006, 2009-2013,
+   2015-2026 Free Software Foundation, Inc.
+
+   This file is part of GNU DIFF.
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#define SYSTEM_INLINE _GL_EXTERN_INLINE
+#include "system.h"
+
+/* Do struct stat *S, *T describe the same file?  */
+bool
+same_file (struct stat const *s, struct stat const *t)
+{
+  if (! SAME_INODE (*s, *t))
+    {
+# if HAVE_STRUCT_STAT_ST_RDEV
+      /* Two character special files describe the same device if st_rdev
+	 is the same, and likewise for block special devices.
+	 They have the same contents, so treat them as the same.  */
+      if (((S_ISCHR (s->st_mode) && S_ISCHR (t->st_mode))
+	   || (S_ISBLK (s->st_mode) && S_ISBLK (t->st_mode)))
+	  && s->st_rdev == t->st_rdev)
+	return true;
+# endif
+      return false;
+    }
+
+  /* Although POSIX says that two files are identical if st_ino and st_dev
+     are the same, all too many file systems incorrectly assign the same
+     (device, inode) pair to two distinct files, including:
+
+     - GNU/Linux NFS servers that export all local file systems as a
+       single NFS file system, if a local (device, inode) pair collides
+       with another one after hashing.
+
+     - GNU/Linux NFS servers that export Btrfs file systems with subvolumes,
+       if the Btrfs (subvolume, inode) hashing function collides.
+       See <https://lwn.net/Articles/866709/>.
+
+     - Qemu virtio-fs before Qemu 5.2 (2020); see
+       <https://bugzilla.redhat.com/show_bug.cgi?id=1795362>.
+
+     - Network Appliance NFS servers in snapshot directories; see
+       Network Appliance bug #195.
+
+     - ClearCase MVFS; see bug id ATRia04618.
+
+     Check whether two files that purport to be the same have the same
+     attributes, to work around instances of this common bug.
+
+     Birthtime is special as st_birthtime is not portable, but when
+     either birthtime is available comparing them should be definitive.  */
+
+#if (defined HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC \
+     || defined HAVE_STRUCT_STAT_ST_BIRTHTIM_TV_NSEC \
+     || defined HAVE_STRUCT_STAT_ST_BIRTHTIMENSEC \
+     || (defined _WIN32 && ! defined __CYGWIN__))
+  /* If either file has a birth time, comparing them is definitive.  */
+  struct timespec sbirth = get_stat_birthtime (s);
+  struct timespec tbirth = get_stat_birthtime (t);
+  if (0 <= sbirth.tv_nsec || 0 <= tbirth.tv_nsec)
+    return timespec_cmp (sbirth, tbirth) == 0;
+#endif
+
+  /* Fall back on comparing other easily-obtainable attributes.
+     Do not inspect all attributes, only attributes useful in checking
+     for the bug.  Check attributes most likely to differ first.
+
+     It's possible for two distinct files on a buggy file system to have
+     the same attributes, but it's not worth slowing down all
+     implementations (or complicating the configuration) to cater to
+     these rare cases in buggy implementations.
+
+     It's also possible for the same file to appear to be two different
+     files, e.g., because its permissions were changed between the two
+     stat calls.  In that case cmp and diff will do extra work
+     to determine that the file contents are the same.  */
+
+  return (get_stat_ctime_ns (s) == get_stat_ctime_ns (t)
+	  && get_stat_mtime_ns (s) == get_stat_mtime_ns (t)
+	  && s->st_ctime == t->st_ctime
+	  && s->st_mtime == t->st_mtime
+	  && s->st_size == t->st_size
+	  && s->st_mode == t->st_mode
+	  && s->st_uid == t->st_uid
+	  && s->st_gid == t->st_gid
+	  && s->st_nlink == t->st_nlink);
+}
+
+/* Use this for code that could be used if diff ever cares about
+   st_size for symlinks, which it doesn't now.  */
+#define care_about_symlink_size false
+
+/* Return the number of bytes in the file described by *S,
+   or -1 if this cannot be determined reliably.  */
+off_t
+stat_size (struct stat const *s)
+{
+  mode_t mode = s->st_mode;
+  off_t size = s->st_size;
+  if (size < 0)
+    return -1;
+  if (! (S_ISREG (mode) || (care_about_symlink_size && S_ISLNK (mode))
+	 || S_TYPEISSHM (s) || S_TYPEISTMO (s)))
+    return -1;
+
+#if (defined __linux__ || defined __CYGWIN__ || defined __FreeBSD__ \
+     || defined __NetBSD__ || defined _AIX)
+  /* On some systems, /proc files with size zero are suspect.  */
+  if (size == 0)
+    {
+      static dev_t proc_dev;
+      if (!proc_dev)
+	{
+	  struct stat st;
+	  st.st_dev = 0;
+	  lstat ("/proc/self", &st);
+	  proc_dev = st.st_dev;
+	}
+      if (proc_dev && s->st_dev == proc_dev)
+	return -1;
+    }
+#endif
+#if care_about_symlink_size && (defined __linux__ || defined __ANDROID__)
+  /* Symlinks have suspect sizes on Linux kernels before 5.15,
+     due to bugs in fscrypt.  */
+  if (S_ISLNK (mode))
+    {
+      static signed char symlink_size_ok;
+      if (! symlink_size_ok)
+	{
+	  struct utsname name;
+	  uname (&name);
+	  char *p = name.release;
+	  symlink_size_ok = ((p[1] != '.' || '5' < p[0]
+			      || (p[0] == '5'
+				  && ('1' <= p[2] && p[2] <= '9')
+				  && ('0' <= p[3] && p[3] <= '9')
+				  && ('5' <= p[3]
+				      || ('0' <= p[4] && p[4] <= '9'))))
+			     ? 1 : -1);
+	}
+      if (symlink_size_ok < 0)
+	return -1;
+    }
+#endif
+
+  return size;
+}

src/system.h

@@ -1,7 +1,7 @@
 /* System dependent declarations.
 
    Copyright (C) 1988-1989, 1992-1995, 1998, 2001-2002, 2004, 2006, 2009-2013,
-   2015-2017 Free Software Foundation, Inc.
+   2015-2026 Free Software Foundation, Inc.
 
    This file is part of GNU DIFF.
 
@@ -20,33 +20,15 @@
 
 #include <config.h>
 
-/* Use this to suppress gcc's "...may be used before initialized" warnings. */
-#ifdef lint
-# define IF_LINT(Code) Code
-#else
-# define IF_LINT(Code) /* empty */
-#endif
-
-/* Define '__attribute__' and 'volatile' first
-   so that they're used consistently in all system includes.  */
-#if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 6) || __STRICT_ANSI__
-# define __attribute__(x)
-#endif
-
 #include <verify.h>
 
 #include <sys/types.h>
 
 #include <sys/stat.h>
-#include "stat-macros.h"
-
-#ifndef STAT_BLOCKSIZE
-# if HAVE_STRUCT_STAT_ST_BLKSIZE
-#  define STAT_BLOCKSIZE(s) ((s).st_blksize)
-# else
-#  define STAT_BLOCKSIZE(s) (8 * 1024)
-# endif
-#endif
+#include <stat-macros.h>
+#include <stat-size.h>
+#include <stat-time.h>
+#include <timespec.h>
 
 #include <unistd.h>
 
@@ -61,24 +43,14 @@
 #endif
 
 #include <stdlib.h>
-#define EXIT_TROUBLE 2
 
+#include <inttypes.h>
 #include <limits.h>
 #include <locale.h>
+#include <stdbit.h>
+#include <stdckdint.h>
 #include <stddef.h>
-#include <inttypes.h>
-
 #include <string.h>
-#if ! HAVE_STRCASECOLL
-# if HAVE_STRICOLL || defined stricoll
-#  define strcasecoll(a, b) stricoll (a, b)
-# else
-#  define strcasecoll(a, b) strcasecmp (a, b) /* best we can do */
-# endif
-#endif
-#if ! (HAVE_STRCASECMP || defined strcasecmp)
-int strcasecmp (char const *, char const *);
-#endif
 
 #include <gettext.h>
 #if ! ENABLE_NLS
@@ -91,17 +63,6 @@ int strcasecmp (char const *, char const *);
 #define _(msgid) gettext (msgid)
 #define N_(msgid) msgid
 
-#include <ctype.h>
-
-/* ISDIGIT differs from isdigit, as follows:
-   - Its arg may be any int or unsigned int; it need not be an unsigned char.
-   - It's guaranteed to evaluate its argument exactly once.
-   - It's typically faster.
-   POSIX 1003.1-2001 says that only '0' through '9' are digits.
-   Prefer ISDIGIT to isdigit unless it's important to use the locale's
-   definition of 'digit' even when the host does not conform to POSIX.  */
-#define ISDIGIT(c) ((unsigned int) (c) - '0' <= 9)
-
 #include <errno.h>
 
 #include <signal.h>
@@ -109,55 +70,50 @@ int strcasecmp (char const *, char const *);
 # define SIGCHLD SIGCLD
 #endif
 
-#undef MIN
-#undef MAX
-#define MIN(a, b) ((a) <= (b) ? (a) : (b))
-#define MAX(a, b) ((a) >= (b) ? (a) : (b))
-
-#include <stdbool.h>
+#include <attribute.h>
+#include <idx.h>
 #include <intprops.h>
-#include "propername.h"
+#include <minmax.h>
+#include <propername.h>
+#include <same-inode.h>
+
 #include "version.h"
 
+/* Evaluate an assertion E that is guaranteed to be true.
+   E should not crash, loop forever, or have side effects.  */
+#if defined DDEBUG && !defined NDEBUG
+/* Abort the program if E is false.  */
+# include <assert.h>
+# define dassert(e) assert (e)
+#else
+/* The compiler can assume E, as behavior is undefined otherwise.  */
+# define dassert(e) assume (e)
+#endif
+
+#ifndef SYSTEM_INLINE
+# define SYSTEM_INLINE _GL_INLINE
+#endif
+
+_GL_INLINE_HEADER_BEGIN
+
 /* Type used for fast comparison of several bytes at a time.
-   This used to be uintmax_t, but changing it to size_t
+   The type is a pointer to an incomplete struct,
+   so that its values are less likely to be misused.
+   This used to be uintmax_t, but changing it to the size of a pointer
    made plain 'cmp' 90% faster (GCC 4.8.1, x86).  */
 
 #ifndef word
-# define word size_t
+typedef struct incomplete *word;
 #endif
 
 /* The signed integer type of a line number.  Since files are read
-   into main memory, ptrdiff_t should be wide enough.  */
+   into main memory, ptrdiff_t should be wide enough.  pI is for
+   printing line numbers.  */
 
 typedef ptrdiff_t lin;
 #define LIN_MAX PTRDIFF_MAX
-
-/* The signed integer type for printing line numbers, and its printf
-   length modifier.  This is not simply ptrdiff_t, to cater to older
-   and/or nonstandard C libraries where "l" works but "ll" and "t" do
-   not, or where 'long' is too narrow and "ll" works but "t" does not.  */
-
-#if LIN_MAX <= LONG_MAX
-typedef long int printint;
-# define pI "l"
-#elif LIN_MAX <= LLONG_MAX
-typedef long long int printint;
-# define pI "ll"
-#else
-typedef ptrdiff_t printint;
-# define pI "t"
-#endif
-
-verify (TYPE_SIGNED (lin));
-verify (TYPE_SIGNED (printint));
-verify (LIN_MAX == TYPE_MAXIMUM (lin));
-verify (LIN_MAX <= TYPE_MAXIMUM (printint));
-
-/* Limit so that 2 * CONTEXT + 1 does not overflow.  */
-
-#define CONTEXT_MAX ((LIN_MAX - 1) / 2)
-
+#define pI "t"
+static_assert (LIN_MAX == IDX_MAX);
 
 /* This section contains POSIX-compliant defaults for macros
    that are meant to be overridden by hand in config.h as needed.  */
@@ -174,67 +130,20 @@ verify (LIN_MAX <= TYPE_MAXIMUM (printint));
 # define NULL_DEVICE "/dev/null"
 #endif
 
-/* Do struct stat *S, *T describe the same special file?  */
-#ifndef same_special_file
-# if HAVE_STRUCT_STAT_ST_RDEV && defined S_ISBLK && defined S_ISCHR
-#  define same_special_file(s, t) \
-     (((S_ISBLK ((s)->st_mode) && S_ISBLK ((t)->st_mode)) \
-       || (S_ISCHR ((s)->st_mode) && S_ISCHR ((t)->st_mode))) \
-      && (s)->st_rdev == (t)->st_rdev)
-# else
-#  define same_special_file(s, t) 0
-# endif
-#endif
-
-/* Do struct stat *S, *T describe the same file?  Answer -1 if unknown.  */
-#ifndef same_file
-# define same_file(s, t) \
-    ((((s)->st_ino == (t)->st_ino) && ((s)->st_dev == (t)->st_dev)) \
-     || same_special_file (s, t))
-#endif
-
-/* Do struct stat *S, *T have the same file attributes?
-
-   POSIX says that two files are identical if st_ino and st_dev are
-   the same, but many file systems incorrectly assign the same (device,
-   inode) pair to two distinct files, including:
-
-   - GNU/Linux NFS servers that export all local file systems as a
-     single NFS file system, if a local device number (st_dev) exceeds
-     255, or if a local inode number (st_ino) exceeds 16777215.
-
-   - Network Appliance NFS servers in snapshot directories; see
-     Network Appliance bug #195.
-
-   - ClearCase MVFS; see bug id ATRia04618.
-
-   Check whether two files that purport to be the same have the same
-   attributes, to work around instances of this common bug.  Do not
-   inspect all attributes, only attributes useful in checking for this
-   bug.
-
-   It's possible for two distinct files on a buggy file system to have
-   the same attributes, but it's not worth slowing down all
-   implementations (or complicating the configuration) to cater to
-   these rare cases in buggy implementations.  */
-
-#ifndef same_file_attributes
-# define same_file_attributes(s, t) \
-   ((s)->st_mode == (t)->st_mode \
-    && (s)->st_nlink == (t)->st_nlink \
-    && (s)->st_uid == (t)->st_uid \
-    && (s)->st_gid == (t)->st_gid \
-    && (s)->st_size == (t)->st_size \
-    && (s)->st_mtime == (t)->st_mtime \
-    && (s)->st_ctime == (t)->st_ctime)
-#endif
-
 #define STREQ(a, b) (strcmp (a, b) == 0)
 
-#ifndef FALLTHROUGH
-# if __GNUC__ < 7
-#  define FALLTHROUGH ((void) 0)
-# else
-#  define FALLTHROUGH __attribute__ ((__fallthrough__))
-# endif
-#endif
+/* Return the floor of the log base 2 of N.  Return -1 if N is zero.  */
+SYSTEM_INLINE int floor_log2 (idx_t n)
+{
+  static_assert (IDX_MAX <= SIZE_MAX);
+  size_t s = n;
+  int w = stdc_bit_width (s);
+  return w - 1;
+}
+
+_GL_INLINE_HEADER_END
+
+extern bool same_file (struct stat const *, struct stat const *)
+  ATTRIBUTE_PURE;
+extern off_t stat_size (struct stat const *)
+  ATTRIBUTE_PURE;

tests/Makefile.am

@@ -5,25 +5,42 @@ TESTS = \
   bignum \
   binary \
   brief-vs-stat-zero-kernel-lies \
+  bug-64316 \
   cmp \
   colliding-file-names \
   diff3 \
+  empty-file \
   excess-slash \
+  expand-tabs \
   help-version	\
+  ifdef \
   invalid-re	\
   function-line-vs-leading-space \
+  ignore-case \
   ignore-matching-lines \
+  ignore-tab-expansion \
   label-vs-func	\
+  large-subopt \
   new-file \
   no-dereference \
   no-newline-at-eof \
+  side-by-side \
+  side-by-side-seq \
+  starting-file \
   stdin \
   strcoll-0-names \
   filename-quoting \
-  colors
+  strip-trailing-cr \
+  timezone \
+  colors \
+  y2038-vs-32bit
+
+XFAIL_TESTS = large-subopt
 
 EXTRA_DIST = \
-  $(TESTS) init.sh t-local.sh envvar-check
+  $(TESTS) init.cfg init.sh t-local.sh envvar-check \
+  large-subopt.in1 \
+  large-subopt.in2
 
 # Note that the first lines are statements.  They ensure that environment
 # variables that can perturb tests are unset or set to expected values.
@@ -75,7 +92,6 @@ TESTS_ENVIRONMENT =					\
   PREFERABLY_POSIX_SHELL='$(PREFERABLY_POSIX_SHELL)'	\
   REPLACE_GETCWD=$(REPLACE_GETCWD)			\
   PATH='$(abs_top_builddir)/src$(PATH_SEPARATOR)'"$$PATH"	\
-  stderr_fileno_=9					\
   ; 9>&2
 
 LOG_COMPILER= $(SHELL)

tests/brief-vs-stat-zero-kernel-lies

@@ -35,4 +35,9 @@ printf 'diff\0--brief\0/proc/self/cmdline\0bin\0' > bin || framework_failure_
 diff --brief /proc/self/cmdline bin > out 2>&1 || fail=1
 compare /dev/null out || fail=1
 
+# Similarly for cmp -s.
+printf 'cmp\0-s\0/proc/self/cmdline\0bin\0' > bin || framework_failure_
+cmp -s /proc/self/cmdline bin > out 2>&1 || fail=1
+compare /dev/null out || fail=1
+
 Exit $fail

tests/bug-64316

@@ -0,0 +1,31 @@
+#!/bin/sh
+# Test for signed integer overflow bug within diff.
+# Bug reported by Gisele Vanem <http://bugs.gnu.org/64316>.
+# Compile with gcc -fsanitize=undefined to test for this bug.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+for f in a b; do
+  printf 'c\nd\ne\nf\ng\n%s\nh\ni\nj\nk\nl\n' $f >$f || framework_failure_
+done
+
+cat >exp <<'EOF' || framework_failure_
+@@ -3,7 +3,7 @@
+ e
+ f
+ g
+-a
++b
+ h
+ i
+ j
+EOF
+
+returns_ 1 diff -u a b >out 2>err || fail=1
+sed '1,2d' out >out1 || framework_failure_
+compare exp out1 || fail=1
+compare /dev/null err || fail=1
+
+Exit $fail

tests/cmp

@@ -1,7 +1,7 @@
 #!/bin/sh
 # Test 'cmp'.
 
-# Copyright 2017 Free Software Foundation, Inc.
+# Copyright 2017-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -27,7 +27,7 @@ cmp a b
 a b differ: char 1, line 1
 1
 cmp a c
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp a d
 cmp: d: No such file or directory
@@ -38,16 +38,16 @@ b a differ: char 1, line 1
 cmp b b
 0
 cmp b c
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp b d
 cmp: d: No such file or directory
 2
 cmp c a
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp c b
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp c c
 0
@@ -72,7 +72,7 @@ cmp -l a b
 1 141 142
 1
 cmp -l a c
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp -l a d
 cmp: d: No such file or directory
@@ -83,16 +83,16 @@ cmp -l b a
 cmp -l b b
 0
 cmp -l b c
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp -l b d
 cmp: d: No such file or directory
 2
 cmp -l c a
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp -l c b
-cmp: EOF on c which is empty
+cmp: EOF on 'c' which is empty
 1
 cmp -l c c
 0
@@ -166,22 +166,22 @@ compare exp out || fail=1
 
 cat <<'EOF' > exp1 || fail=1
 cmp a0 a1
-cmp: EOF on a0 which is empty
+cmp: EOF on 'a0' which is empty
 1
 cmp a1 a2
-cmp: EOF on a1 after byte 2, line 1
+cmp: EOF on 'a1' after byte 2, line 1
 1
 cmp a2 a3
-cmp: EOF on a2 after byte 5, in line 2
+cmp: EOF on 'a2' after byte 5, in line 2
 1
 cmp -l a0 a1
-cmp: EOF on a0 which is empty
+cmp: EOF on 'a0' which is empty
 1
 cmp -l a1 a2
-cmp: EOF on a1 after byte 2
+cmp: EOF on 'a1' after byte 2
 1
 cmp -l a2 a3
-cmp: EOF on a2 after byte 5
+cmp: EOF on 'a2' after byte 5
 1
 cmp -s a0 a1
 1
@@ -208,4 +208,47 @@ done >out1
 
 compare exp1 out1 || fail=1
 
+printf 'bad\n' >bad
+printf 'bug\n' >bug
+echo LC_ALL=C cmp -b bad bug
+LC_ALL=C cmp -b bad bug
+test $? -eq 1 || fail=1
+case `LC_ALL=C cmp -b bad bug` in
+  'bad bug differ: byte 2, line 1 is '*' a '*' u') ;;
+  *) echo 'expected cmp -b to report a and u'; fail=1;;
+esac
+
+printf 'Jackdaws love my big sphinx of quartz!' > j1
+printf 'jackdaws love my big sphinx of quartz.' > j2
+cat <<'EOF' > exp2 || fail=1
+ 1 112 J    152 j
+38  41 !     56 .
+EOF
+cmp -bl j1 j2 > out2
+test $? -eq 1 || fail=1
+compare exp2 out2 || fail=1
+
+if (LC_ALL=en_US.iso8859-1 locale charmap) >/dev/null 2>&1; then
+  printf '\200\300\377\n' >j3
+  printf '\0@\177\n' >j4
+  cat <<'EOF' >exp3 || fail=1
+1 200 M-^@   0 ^@
+2 300 M-@  100 @
+3 377 M-^? 177 ^?
+EOF
+  LC_ALL=en_US.iso8859-1 cmp -bl j3 j4 >out3
+  test $? -eq 1 || fail=1
+  compare exp3 out3 || fail=1
+fi
+
+big=99999999999999999999999999999999999999999999999999999999999
+cmp -i $big j1 j2 || fail=1
+cmp -i 1000 -n $big j1 j2 || fail=1
+
+rm -f a b
+if timeout 0.1 true && truncate -s 14T a && truncate -s 15T b; then
+  returns_ 1 timeout 0.4 cmp a b >/dev/null || fail=1
+fi
+rm -f a b
+
 Exit $fail

tests/colors

@@ -10,6 +10,10 @@ fail=0
 echo a > a
 echo b > b
 
+# On systems lacking fractional timestamp support, diff -u does not print
+# a decimal point and 9-digit nanosecond suffix.
+nanosecond_zeros=$(diff -u a b|grep '\.' > /dev/null && echo .000000000)
+
 epoch='1970-01-01 00:00:00'
 touch -t 197001010000.00 a b
 
@@ -18,16 +22,16 @@ tab=$(printf '\t')
 
 gen_exp_u()
 {
-    local epoch_plus="$epoch.000000000 +0000"
+    local epoch_plus="$epoch$nanosecond_zeros +0000"
     local rs=$(printf "$e[${rs}m")
     local hd=$(printf "$e[${hd}m")
     local ad=$(printf "$e[${ad}m")
     local de=$(printf "$e[${de}m")
     local ln=$(printf "$e[${ln}m")
     printf '%s' \
-"$hd--- a$tab$epoch_plus
-+++ b$tab$epoch_plus
-$rs${ln}@@ -1 +1 @@$rs
+"$hd--- a$tab$epoch_plus$rs
+$hd+++ b$tab$epoch_plus$rs
+${ln}@@ -1 +1 @@$rs
 $de-a$rs
 $ad+b$rs
 "
@@ -42,9 +46,9 @@ gen_exp_c()
     local de=$(printf "$e[${de}m")
     local ln=$(printf "$e[${ln}m")
     printf '%s' \
-"$hd*** a$tab$epoch_posix_1003_1_2001
---- b$tab$epoch_posix_1003_1_2001
-$rs***************
+"$hd*** a$tab$epoch_posix_1003_1_2001$rs
+$hd--- b$tab$epoch_posix_1003_1_2001$rs
+***************
 $ln*** 1 ****$rs
 $de! a$rs
 $ln--- 1 ----$rs
@@ -85,7 +89,8 @@ returns_ 1 diff --color=auto a b > out || fail=1
 gen_exp_default > exp || framework_failure_
 compare exp out || fail=1
 
-returns_ 1 env TERM=dumb diff ---presume-output-tty --color=auto a b > out \
+returns_ 1 env PATH="$PATH" TERM=dumb \
+         diff ---presume-output-tty --color=auto a b > out \
   || fail=1
 gen_exp_default > exp || framework_failure_
 compare exp out || fail=1
@@ -119,10 +124,13 @@ compare exp out || fail=1
 # Before the fix in http://debbugs.gnu.org/22067,
 # this test would trigger an infinite loop bug.
 mkfifo fifo
-printf '%*s-a' 1000000 > a
-printf '%*s-b' 1000000 > b
+printf '%1000000s-a' > a
+printf '%1000000s-b' > b
 head -c 10 < fifo > /dev/null &
 diff --color=always ---presume-output-tty a b > fifo
-test $? = 141 || fail=1
+
+# Depending on version of GNU make (4.3.92-4.4 set SIGPIPE to "ignore"),
+# either of these is acceptable.
+case $? in 2|141) ;; *) fail=1 ;; esac
 
 Exit $fail

tests/diff3

@@ -3,18 +3,6 @@
 
 . "${srcdir=.}/init.sh"; path_prepend_ ../src
 
-# Some systems lack seq.
-# A limited replacement for seq: handle 1 or 2 args; increment must be 1
-seq()
-{
-  case $# in
-    1) start=1  final=$1;;
-    2) start=$1 final=$2;;
-    *) echo you lose 1>&2; exit 1;;
-  esac
-  awk 'BEGIN{for(i='$start';i<='$final';i++) print i}' < /dev/null
-}
-
 echo a > a || framework_failure_
 echo b > b || framework_failure_
 echo c > c || framework_failure_
@@ -30,12 +18,14 @@ EOF
 
 fail=0
 
-diff3 a b c > out 2> err || fail=1
+# Use --diff-program=diff since --program-transform-name and similar are
+# applied after 'make install'.
+diff3 --diff-program=diff a b c > out 2> err || fail=1
 compare exp out || fail=1
 compare /dev/null err || fail=1
 
 # Repeat, but with all three files the same:
-diff3 a a a > out 2> err || fail=1
+diff3 --diff-program=diff a a a > out 2> err || fail=1
 compare /dev/null out || fail=1
 compare /dev/null err || fail=1
 
@@ -101,7 +91,7 @@ cat <<'EOF' > exp40 || framework_failure_
   3z
 EOF
 
-diff3 d e f > out 2> err
+diff3 --diff-program=diff d e f > out 2> err
 compare exp40 out || fail=1
 compare /dev/null err || fail=1
 

tests/empty-file

@@ -0,0 +1,23 @@
+#!/bin/sh
+# Test empty files
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+mkdir a b
+touch a/test
+echo 'content' > b/test
+
+returns_ 1 diff -Naur a b >out || fail=1
+cat <<EOF > expected || framework_failure_
+diff -Naur a/test b/test
+--- a/test
++++ b/test
+@@ -0,0 +1 @@
++content
+EOF
+# Remove date and time.
+sed -e 's/^\([-+*][-+*][-+*] [^	]*\)	.*/\1/' out > k; mv k out
+compare expected out || fail=1
+
+Exit $fail

tests/envvar-check

@@ -1,7 +1,7 @@
 # -*- sh -*-
 # Check environment variables for sane values while testing.
 
-# Copyright (C) 2000-2017 Free Software Foundation, Inc.
+# Copyright (C) 2000-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by

tests/expand-tabs

@@ -0,0 +1,19 @@
+#!/bin/sh
+# Test tab expansion.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+cat >exp <<'EOF' || framework_failure_
+0a1
+>         x
+EOF
+
+for p in '\b\tx\n' '\b        x\n' '\b \tx\n'; do
+  printf "$p" | returns_ 1 diff -t /dev/null - >out || fail=1
+done
+
+compare exp out || fail=1
+
+Exit $fail

tests/help-version

@@ -2,7 +2,7 @@
 # Make sure all these programs work properly
 # when invoked with --help or --version.
 
-# Copyright (C) 2000-2013, 2015-2017 Free Software Foundation, Inc.
+# Copyright (C) 2000-2013, 2015-2026 Free Software Foundation, Inc.
 
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -170,8 +170,12 @@ egrep_setup () { args="0 $tmp_in"; }
 fgrep_setup () { args="0 $tmp_in"; }
 
 diff_setup () { args="$tmp_in $tmp_in2"; }
-sdiff_setup () { args="$tmp_in $tmp_in2"; }
-diff3_setup () { args="$tmp_in $tmp_in2 $tmp_in2"; }
+
+# Use --diff-program=diff since --program-transform-name and similar are
+# applied after 'make install'.
+sdiff_setup () { args="--diff-program=diff $tmp_in $tmp_in2"; }
+diff3_setup () { args="--diff-program=diff $tmp_in $tmp_in2 $tmp_in2"; }
+
 cp_setup () { args="$tmp_in $tmp_in2"; }
 ln_setup () { args="$tmp_in ln-target"; }
 ginstall_setup () { args="$tmp_in $tmp_in2"; }

tests/ifdef

@@ -0,0 +1,37 @@
+#!/bin/sh
+# --ifdef
+
+# Bug reported by Robert Webb in <http://bugs.gnu.org/61193>.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+cat <<'EOF' >a
+1
+2
+3
+4
+5
+EOF
+
+cat <<'EOF' >b
+1
+4
+5
+EOF
+
+cat <<'EOF' >exp
+1
+#ifndef ZZZ
+2
+3
+#endif /* ! ZZZ */
+4
+5
+EOF
+
+returns_ 1 diff -D ZZZ a b >out 2>err || fail=1
+compare exp out || fail=1
+
+Exit $fail

tests/ignore-case

@@ -0,0 +1,26 @@
+#!/bin/sh
+# Test ignoring case
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+printf 'Funky Stuff\n' >a
+printf 'fUNKy stuFf\n' >b
+
+diff -i a b >out || fail=1
+compare /dev/null out || fail=1
+
+require_utf8_locale_
+
+echo 'AĀȀΆΑАӐḀἈＡ𐐀-Δ' >a
+echo 'aāȁάαаӑḁἀａ𐐨-δ' >b
+
+diff -i a b >out || fail=1
+compare /dev/null out || fail=1
+
+mkdir d || framework_failure_
+touch f d/F || framework_failure_
+diff --ignore-file-name-case f d || fail=1
+
+Exit $fail

tests/ignore-tab-expansion

@@ -0,0 +1,28 @@
+#!/bin/sh
+# Test ignoring tab expansion.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+for p in '\b' '\r' '\t ' '\n'; do
+  printf "$p"'\b\tx\n' >a || framework_failure_
+  printf "$p"'\b        x\n' >b || framework_failure
+
+  diff -E a b >out || fail=1
+  compare /dev/null out || fail=1
+done
+
+require_utf8_locale_
+
+tr '_@' ' \t' >a <<\EOF
+@字字字xx@x_@字_@xxx_x@@x
+EOF
+tr '_@' ' \t' >b <<\EOF
+_@字字字xx________x_@字_@xxx_x@@x
+EOF
+
+diff -E a b >out || fail=1
+compare /dev/null out || fail=1
+
+Exit $fail

tests/init.cfg

@@ -0,0 +1,135 @@
+# This file is sourced by init.sh, *before* its initialization.
+
+# Copyright (C) 2010-2026 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+# Skip the current test if valgrind doesn't work,
+# which could happen if not installed,
+# or hasn't support for the built architecture,
+# or hasn't appropriate error suppressions installed etc.
+
+# This goes hand in hand with the "exec 9>&2;" in tests/Makefile.am's
+# TESTS_ENVIRONMENT definition.
+stderr_fileno_=9
+
+# Having an unsearchable directory in PATH causes execve to fail with EACCES
+# when applied to an unresolvable program name, contrary to the desired ENOENT.
+# Avoid the problem by rewriting PATH to exclude unsearchable directories.
+# Also, if PATH lacks /sbin and/or /usr/sbin, append it/them.
+sanitize_path_()
+{
+  # FIXME: remove double quotes around $IFS when all tests use init.sh.
+  # They constitute a work-around for a bug in FreeBSD 8.1's /bin/sh.
+  local saved_IFS="$IFS"
+    IFS=$PATH_SEPARATOR
+    set -- $PATH
+  IFS=$saved_IFS
+
+  local d d1
+  local colon=
+  local new_path=
+  for d in "$@"; do
+    test -z "$d" && d1=. || d1=$d
+    if ls -d "$d1/." > /dev/null 2>&1; then
+      new_path="$new_path$colon$d"
+      colon=$PATH_SEPARATOR
+    fi
+  done
+
+  for d in /sbin /usr/sbin ; do
+    case "$PATH_SEPARATOR$new_path$PATH_SEPARATOR" in
+      *$PATH_SEPARATOR$d$PATH_SEPARATOR*) ;;
+      *) new_path="$new_path$PATH_SEPARATOR$d" ;;
+    esac
+  done
+
+  PATH=$new_path
+  export PATH
+}
+
+require_timeout_()
+{
+  ( timeout 10s true ) > /dev/null 2>&1 \
+    || skip_ your system lacks the timeout program
+  returns_ 1 timeout 10s false \
+    || skip_ your system has a non-GNU timeout program
+  returns_ 124 timeout 0.01 sleep 0.02 \
+    || skip_ "'timeout 0.01 sleep 0.02' did not time out"
+}
+
+require_valgrind_()
+{
+  require_timeout_
+  local errout; errout=$(
+    LC_ALL=C timeout --signal=9 3 valgrind --error-exitcode=1 diff /dev/null /dev/null 2>&1
+  ) ||
+    skip_ "requires a working valgrind"
+  case $errout in
+    *'Serious error'*)
+      skip_ "requires a valgrind without serious errors";;
+  esac
+}
+
+# Skip the current test if we lack Perl.
+require_perl_()
+{
+  : ${PERL=perl}
+  $PERL -e 'use warnings' > /dev/null 2>&1 \
+    || skip_ 'configure did not find a usable version of Perl'
+}
+
+# Run this test in a UTF-8 locale if possible, and skip the test otherwise.
+# Prefer en_US for its diagnostics.
+require_utf8_locale_()
+{
+  local locale
+
+  if test "`(locale charmap) 2>/dev/null`" != UTF-8; then
+    for locale in en_US.UTF-8 `(locale -a) 2>/dev/null` not-found; do
+      case $locale in
+        *.[Uu][Tt][Ff]*8)
+	  if test "`(LC_ALL=$locale locale charmap) 2>/dev/null`" = UTF-8; then
+	    LC_ALL=$locale
+	    export LC_ALL
+	    break
+	  fi;;
+        not-found)
+	  skip_ "No UTF-8 locale found";;
+      esac
+    done
+  fi
+
+  # Solaris 10's /usr/bin/tr can silently malfunction. Try to find one that works.
+  found_working_tr=0
+  for i in tr /usr/xpg4/bin/tr gtr; do
+    tr() { env $i "$@"; }
+    test '一' = "$(printf 一|tr a b)" && { found_working_tr=1; break; }
+  done
+  test $found_working_tr = 1 || skip_ "failed to find a working tr program"
+}
+
+# Some systems lack seq.
+# A limited replacement for seq: handle 1 or 2 args; increment must be 1
+seq()
+{
+  case $# in
+    1) start=1  final=$1;;
+    2) start=$1 final=$2;;
+    *) echo you lose 1>&2; exit 1;;
+  esac
+  awk 'BEGIN{for(i='$start';i<='$final';i++) print i}' < /dev/null
+}
+
+sanitize_path_

tests/init.sh

@@ -1,605 +0,0 @@
-# source this file; set up for tests
-
-# Copyright (C) 2009-2017 Free Software Foundation, Inc.
-
-# This program is free software: you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation, either version 3 of the License, or
-# (at your option) any later version.
-
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-# GNU General Public License for more details.
-
-# You should have received a copy of the GNU General Public License
-# along with this program.  If not, see <http://www.gnu.org/licenses/>.
-
-# Using this file in a test
-# =========================
-#
-# The typical skeleton of a test looks like this:
-#
-#   #!/bin/sh
-#   . "${srcdir=.}/init.sh"; path_prepend_ .
-#   Execute some commands.
-#   Note that these commands are executed in a subdirectory, therefore you
-#   need to prepend "../" to relative filenames in the build directory.
-#   Note that the "path_prepend_ ." is useful only if the body of your
-#   test invokes programs residing in the initial directory.
-#   For example, if the programs you want to test are in src/, and this test
-#   script is named tests/test-1, then you would use "path_prepend_ ../src",
-#   or perhaps export PATH='$(abs_top_builddir)/src$(PATH_SEPARATOR)'"$$PATH"
-#   to all tests via automake's TESTS_ENVIRONMENT.
-#   Set the exit code 0 for success, 77 for skipped, or 1 or other for failure.
-#   Use the skip_ and fail_ functions to print a diagnostic and then exit
-#   with the corresponding exit code.
-#   Exit $?
-
-# Executing a test that uses this file
-# ====================================
-#
-# Running a single test:
-#   $ make check TESTS=test-foo.sh
-#
-# Running a single test, with verbose output:
-#   $ make check TESTS=test-foo.sh VERBOSE=yes
-#
-# Running a single test, keeping the temporary directory:
-#   $ make check TESTS=test-foo.sh KEEP=yes
-#
-# Running a single test, with single-stepping:
-#   1. Go into a sub-shell:
-#   $ bash
-#   2. Set relevant environment variables from TESTS_ENVIRONMENT in the
-#      Makefile:
-#   $ export srcdir=../../tests # this is an example
-#   3. Execute the commands from the test, copy&pasting them one by one:
-#   $ . "$srcdir/init.sh"; path_prepend_ .
-#   ...
-#   4. Finally
-#   $ exit
-
-ME_=`expr "./$0" : '.*/\(.*\)$'`
-
-# We use a trap below for cleanup.  This requires us to go through
-# hoops to get the right exit status transported through the handler.
-# So use 'Exit STATUS' instead of 'exit STATUS' inside of the tests.
-# Turn off errexit here so that we don't trip the bug with OSF1/Tru64
-# sh inside this function.
-Exit () { set +e; (exit $1); exit $1; }
-
-# Print warnings (e.g., about skipped and failed tests) to this file number.
-# Override by defining to say, 9, in init.cfg, and putting say,
-#   export ...ENVVAR_SETTINGS...; $(SHELL) 9>&2
-# in the definition of TESTS_ENVIRONMENT in your tests/Makefile.am file.
-# This is useful when using automake's parallel tests mode, to print
-# the reason for skip/failure to console, rather than to the .log files.
-: ${stderr_fileno_=2}
-
-# Note that correct expansion of "$*" depends on IFS starting with ' '.
-# Always write the full diagnostic to stderr.
-# When stderr_fileno_ is not 2, also emit the first line of the
-# diagnostic to that file descriptor.
-warn_ ()
-{
-  # If IFS does not start with ' ', set it and emit the warning in a subshell.
-  case $IFS in
-    ' '*) printf '%s\n' "$*" >&2
-          test $stderr_fileno_ = 2 \
-            || { printf '%s\n' "$*" | sed 1q >&$stderr_fileno_ ; } ;;
-    *) (IFS=' '; warn_ "$@");;
-  esac
-}
-fail_ () { warn_ "$ME_: failed test: $@"; Exit 1; }
-skip_ () { warn_ "$ME_: skipped test: $@"; Exit 77; }
-fatal_ () { warn_ "$ME_: hard error: $@"; Exit 99; }
-framework_failure_ () { warn_ "$ME_: set-up failure: $@"; Exit 99; }
-
-# This is used to simplify checking of the return value
-# which is useful when ensuring a command fails as desired.
-# I.e., just doing `command ... &&fail=1` will not catch
-# a segfault in command for example.  With this helper you
-# instead check an explicit exit code like
-#   returns_ 1 command ... || fail
-returns_ () {
-  # Disable tracing so it doesn't interfere with stderr of the wrapped command
-  { set +x; } 2>/dev/null
-
-  local exp_exit="$1"
-  shift
-  "$@"
-  test $? -eq $exp_exit && ret_=0 || ret_=1
-
-  if test "$VERBOSE" = yes && test "$gl_set_x_corrupts_stderr_" = false; then
-    set -x
-  fi
-  { return $ret_; } 2>/dev/null
-}
-
-# Sanitize this shell to POSIX mode, if possible.
-DUALCASE=1; export DUALCASE
-if test -n "${ZSH_VERSION+set}" && (emulate sh) >/dev/null 2>&1; then
-  emulate sh
-  NULLCMD=:
-  alias -g '${1+"$@"}'='"$@"'
-  setopt NO_GLOB_SUBST
-else
-  case `(set -o) 2>/dev/null` in
-    *posix*) set -o posix ;;
-  esac
-fi
-
-# We require $(...) support unconditionally.
-# We require non-surprising "local" semantics (this eliminates dash).
-# This takes the admittedly draconian step of eliminating dash, because the
-# assignment tab=$(printf '\t') works fine, yet preceding it with "local "
-# transforms it into an assignment that sets the variable to the empty string.
-# That is too counter-intuitive, and can lead to subtle run-time malfunction.
-# The example below is less subtle in that with dash, it evokes the run-time
-# exception "dash: 1: local: 1: bad variable name".
-# We require a few additional shell features only when $EXEEXT is nonempty,
-# in order to support automatic $EXEEXT emulation:
-# - hyphen-containing alias names
-# - we prefer to use ${var#...} substitution, rather than having
-#   to work around lack of support for that feature.
-# The following code attempts to find a shell with support for these features.
-# If the current shell passes the test, we're done.  Otherwise, test other
-# shells until we find one that passes.  If one is found, re-exec it.
-# If no acceptable shell is found, skip the current test.
-#
-# The "...set -x; P=1 true 2>err..." test is to disqualify any shell that
-# emits "P=1" into err, as /bin/sh from SunOS 5.11 and OpenBSD 4.7 do.
-#
-# Use "9" to indicate success (rather than 0), in case some shell acts
-# like Solaris 10's /bin/sh but exits successfully instead of with status 2.
-
-# Eval this code in a subshell to determine a shell's suitability.
-# 10 - passes all tests; ok to use
-#  9 - ok, but enabling "set -x" corrupts app stderr; prefer higher score
-#  ? - not ok
-gl_shell_test_script_='
-test $(echo y) = y || exit 1
-f_local_() { local v=1; }; f_local_ || exit 1
-f_dash_local_fail_() { local t=$(printf " 1"); }; f_dash_local_fail_
-score_=10
-if test "$VERBOSE" = yes; then
-  test -n "$( (exec 3>&1; set -x; P=1 true 2>&3) 2> /dev/null)" && score_=9
-fi
-test -z "$EXEEXT" && exit $score_
-shopt -s expand_aliases
-alias a-b="echo zoo"
-v=abx
-     test ${v%x} = ab \
-  && test ${v#a} = bx \
-  && test $(a-b) = zoo \
-  && exit $score_
-'
-
-if test "x$1" = "x--no-reexec"; then
-  shift
-else
-  # Assume a working shell.  Export to subshells (setup_ needs this).
-  gl_set_x_corrupts_stderr_=false
-  export gl_set_x_corrupts_stderr_
-
-  # Record the first marginally acceptable shell.
-  marginal_=
-
-  # Search for a shell that meets our requirements.
-  for re_shell_ in __current__ "${CONFIG_SHELL:-no_shell}" \
-      /bin/sh bash dash zsh pdksh fail
-  do
-    test "$re_shell_" = no_shell && continue
-
-    # If we've made it all the way to the sentinel, "fail" without
-    # finding even a marginal shell, skip this test.
-    if test "$re_shell_" = fail; then
-      test -z "$marginal_" && skip_ failed to find an adequate shell
-      re_shell_=$marginal_
-      break
-    fi
-
-    # When testing the current shell, simply "eval" the test code.
-    # Otherwise, run it via $re_shell_ -c ...
-    if test "$re_shell_" = __current__; then
-      # 'eval'ing this code makes Solaris 10's /bin/sh exit with
-      # $? set to 2.  It does not evaluate any of the code after the
-      # "unexpected" first '('.  Thus, we must run it in a subshell.
-      ( eval "$gl_shell_test_script_" ) > /dev/null 2>&1
-    else
-      "$re_shell_" -c "$gl_shell_test_script_" 2>/dev/null
-    fi
-
-    st_=$?
-
-    # $re_shell_ works just fine.  Use it.
-    if test $st_ = 10; then
-      gl_set_x_corrupts_stderr_=false
-      break
-    fi
-
-    # If this is our first marginally acceptable shell, remember it.
-    if test "$st_:$marginal_" = 9: ; then
-      marginal_="$re_shell_"
-      gl_set_x_corrupts_stderr_=true
-    fi
-  done
-
-  if test "$re_shell_" != __current__; then
-    # Found a usable shell.  Preserve -v and -x.
-    case $- in
-      *v*x* | *x*v*) opts_=-vx ;;
-      *v*) opts_=-v ;;
-      *x*) opts_=-x ;;
-      *) opts_= ;;
-    esac
-    re_shell=$re_shell_
-    export re_shell
-    exec "$re_shell_" $opts_ "$0" --no-reexec "$@"
-    echo "$ME_: exec failed" 1>&2
-    exit 127
-  fi
-fi
-
-# If this is bash, turn off all aliases.
-test -n "$BASH_VERSION" && unalias -a
-
-# Note that when supporting $EXEEXT (transparently mapping from PROG_NAME to
-# PROG_NAME.exe), we want to support hyphen-containing names like test-acos.
-# That is part of the shell-selection test above.  Why use aliases rather
-# than functions?  Because support for hyphen-containing aliases is more
-# widespread than that for hyphen-containing function names.
-test -n "$EXEEXT" && shopt -s expand_aliases
-
-# Enable glibc's malloc-perturbing option.
-# This is useful for exposing code that depends on the fact that
-# malloc-related functions often return memory that is mostly zeroed.
-# If you have the time and cycles, use valgrind to do an even better job.
-: ${MALLOC_PERTURB_=87}
-export MALLOC_PERTURB_
-
-# This is a stub function that is run upon trap (upon regular exit and
-# interrupt).  Override it with a per-test function, e.g., to unmount
-# a partition, or to undo any other global state changes.
-cleanup_ () { :; }
-
-# Emit a header similar to that from diff -u;  Print the simulated "diff"
-# command so that the order of arguments is clear.  Don't bother with @@ lines.
-emit_diff_u_header_ ()
-{
-  printf '%s\n' "diff -u $*" \
-    "--- $1	1970-01-01" \
-    "+++ $2	1970-01-01"
-}
-
-# Arrange not to let diff or cmp operate on /dev/null,
-# since on some systems (at least OSF/1 5.1), that doesn't work.
-# When there are not two arguments, or no argument is /dev/null, return 2.
-# When one argument is /dev/null and the other is not empty,
-# cat the nonempty file to stderr and return 1.
-# Otherwise, return 0.
-compare_dev_null_ ()
-{
-  test $# = 2 || return 2
-
-  if test "x$1" = x/dev/null; then
-    test -s "$2" || return 0
-    emit_diff_u_header_ "$@"; sed 's/^/+/' "$2"
-    return 1
-  fi
-
-  if test "x$2" = x/dev/null; then
-    test -s "$1" || return 0
-    emit_diff_u_header_ "$@"; sed 's/^/-/' "$1"
-    return 1
-  fi
-
-  return 2
-}
-
-for diff_opt_ in -u -U3 -c '' no; do
-  test "$diff_opt_" != no &&
-    diff_out_=`exec 2>/dev/null; diff $diff_opt_ "$0" "$0" < /dev/null` &&
-    break
-done
-if test "$diff_opt_" != no; then
-  if test -z "$diff_out_"; then
-    compare_ () { diff $diff_opt_ "$@"; }
-  else
-    compare_ ()
-    {
-      # If no differences were found, AIX and HP-UX 'diff' produce output
-      # like "No differences encountered".  Hide this output.
-      diff $diff_opt_ "$@" > diff.out
-      diff_status_=$?
-      test $diff_status_ -eq 0 || cat diff.out || diff_status_=2
-      rm -f diff.out || diff_status_=2
-      return $diff_status_
-    }
-  fi
-elif cmp -s /dev/null /dev/null 2>/dev/null; then
-  compare_ () { cmp -s "$@"; }
-else
-  compare_ () { cmp "$@"; }
-fi
-
-# Usage: compare EXPECTED ACTUAL
-#
-# Given compare_dev_null_'s preprocessing, defer to compare_ if 2 or more.
-# Otherwise, propagate $? to caller: any diffs have already been printed.
-compare ()
-{
-  # This looks like it can be factored to use a simple "case $?"
-  # after unchecked compare_dev_null_ invocation, but that would
-  # fail in a "set -e" environment.
-  if compare_dev_null_ "$@"; then
-    return 0
-  else
-    case $? in
-      1) return 1;;
-      *) compare_ "$@";;
-    esac
-  fi
-}
-
-# An arbitrary prefix to help distinguish test directories.
-testdir_prefix_ () { printf gt; }
-
-# Run the user-overridable cleanup_ function, remove the temporary
-# directory and exit with the incoming value of $?.
-remove_tmp_ ()
-{
-  __st=$?
-  cleanup_
-  if test "$KEEP" = yes; then
-    echo "Not removing temporary directory $test_dir_"
-  else
-    # cd out of the directory we're about to remove
-    cd "$initial_cwd_" || cd / || cd /tmp
-    chmod -R u+rwx "$test_dir_"
-    # If removal fails and exit status was to be 0, then change it to 1.
-    rm -rf "$test_dir_" || { test $__st = 0 && __st=1; }
-  fi
-  exit $__st
-}
-
-# Given a directory name, DIR, if every entry in it that matches *.exe
-# contains only the specified bytes (see the case stmt below), then print
-# a space-separated list of those names and return 0.  Otherwise, don't
-# print anything and return 1.  Naming constraints apply also to DIR.
-find_exe_basenames_ ()
-{
-  feb_dir_=$1
-  feb_fail_=0
-  feb_result_=
-  feb_sp_=
-  for feb_file_ in $feb_dir_/*.exe; do
-    # If there was no *.exe file, or there existed a file named "*.exe" that
-    # was deleted between the above glob expansion and the existence test
-    # below, just skip it.
-    test "x$feb_file_" = "x$feb_dir_/*.exe" && test ! -f "$feb_file_" \
-      && continue
-    # Exempt [.exe, since we can't create a function by that name, yet
-    # we can't invoke [ by PATH search anyways due to shell builtins.
-    test "x$feb_file_" = "x$feb_dir_/[.exe" && continue
-    case $feb_file_ in
-      *[!-a-zA-Z/0-9_.+]*) feb_fail_=1; break;;
-      *) # Remove leading file name components as well as the .exe suffix.
-         feb_file_=${feb_file_##*/}
-         feb_file_=${feb_file_%.exe}
-         feb_result_="$feb_result_$feb_sp_$feb_file_";;
-    esac
-    feb_sp_=' '
-  done
-  test $feb_fail_ = 0 && printf %s "$feb_result_"
-  return $feb_fail_
-}
-
-# Consider the files in directory, $1.
-# For each file name of the form PROG.exe, create an alias named
-# PROG that simply invokes PROG.exe, then return 0.  If any selected
-# file name or the directory name, $1, contains an unexpected character,
-# define no alias and return 1.
-create_exe_shims_ ()
-{
-  case $EXEEXT in
-    '') return 0 ;;
-    .exe) ;;
-    *) echo "$0: unexpected \$EXEEXT value: $EXEEXT" 1>&2; return 1 ;;
-  esac
-
-  base_names_=`find_exe_basenames_ $1` \
-    || { echo "$0 (exe_shim): skipping directory: $1" 1>&2; return 0; }
-
-  if test -n "$base_names_"; then
-    for base_ in $base_names_; do
-      alias "$base_"="$base_$EXEEXT"
-    done
-  fi
-
-  return 0
-}
-
-# Use this function to prepend to PATH an absolute name for each
-# specified, possibly-$initial_cwd_-relative, directory.
-path_prepend_ ()
-{
-  while test $# != 0; do
-    path_dir_=$1
-    case $path_dir_ in
-      '') fail_ "invalid path dir: '$1'";;
-      /*) abs_path_dir_=$path_dir_;;
-      *) abs_path_dir_=$initial_cwd_/$path_dir_;;
-    esac
-    case $abs_path_dir_ in
-      *:*) fail_ "invalid path dir: '$abs_path_dir_'";;
-    esac
-    PATH="$abs_path_dir_:$PATH"
-
-    # Create an alias, FOO, for each FOO.exe in this directory.
-    create_exe_shims_ "$abs_path_dir_" \
-      || fail_ "something failed (above): $abs_path_dir_"
-    shift
-  done
-  export PATH
-}
-
-setup_ ()
-{
-  if test "$VERBOSE" = yes; then
-    # Test whether set -x may cause the selected shell to corrupt an
-    # application's stderr.  Many do, including zsh-4.3.10 and the /bin/sh
-    # from SunOS 5.11, OpenBSD 4.7 and Irix 5.x and 6.5.
-    # If enabling verbose output this way would cause trouble, simply
-    # issue a warning and refrain.
-    if $gl_set_x_corrupts_stderr_; then
-      warn_ "using SHELL=$SHELL with 'set -x' corrupts stderr"
-    else
-      set -x
-    fi
-  fi
-
-  initial_cwd_=$PWD
-
-  pfx_=`testdir_prefix_`
-  test_dir_=`mktempd_ "$initial_cwd_" "$pfx_-$ME_.XXXX"` \
-    || fail_ "failed to create temporary directory in $initial_cwd_"
-  cd "$test_dir_" || fail_ "failed to cd to temporary directory"
-
-  # As autoconf-generated configure scripts do, ensure that IFS
-  # is defined initially, so that saving and restoring $IFS works.
-  gl_init_sh_nl_='
-'
-  IFS=" ""	$gl_init_sh_nl_"
-
-  # This trap statement, along with a trap on 0 below, ensure that the
-  # temporary directory, $test_dir_, is removed upon exit as well as
-  # upon receipt of any of the listed signals.
-  for sig_ in 1 2 3 13 15; do
-    eval "trap 'Exit $(expr $sig_ + 128)' $sig_"
-  done
-}
-
-# Create a temporary directory, much like mktemp -d does.
-# Written by Jim Meyering.
-#
-# Usage: mktempd_ /tmp phoey.XXXXXXXXXX
-#
-# First, try to use the mktemp program.
-# Failing that, we'll roll our own mktemp-like function:
-#  - try to get random bytes from /dev/urandom
-#  - failing that, generate output from a combination of quickly-varying
-#      sources and gzip.  Ignore non-varying gzip header, and extract
-#      "random" bits from there.
-#  - given those bits, map to file-name bytes using tr, and try to create
-#      the desired directory.
-#  - make only $MAX_TRIES_ attempts
-
-# Helper function.  Print $N pseudo-random bytes from a-zA-Z0-9.
-rand_bytes_ ()
-{
-  n_=$1
-
-  # Maybe try openssl rand -base64 $n_prime_|tr '+/=\012' abcd first?
-  # But if they have openssl, they probably have mktemp, too.
-
-  chars_=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789
-  dev_rand_=/dev/urandom
-  if test -r "$dev_rand_"; then
-    # Note: 256-length($chars_) == 194; 3 copies of $chars_ is 186 + 8 = 194.
-    dd ibs=$n_ count=1 if=$dev_rand_ 2>/dev/null \
-      | LC_ALL=C tr -c $chars_ 01234567$chars_$chars_$chars_
-    return
-  fi
-
-  n_plus_50_=`expr $n_ + 50`
-  cmds_='date; date +%N; free; who -a; w; ps auxww; ps ef; netstat -n'
-  data_=` (eval "$cmds_") 2>&1 | gzip `
-
-  # Ensure that $data_ has length at least 50+$n_
-  while :; do
-    len_=`echo "$data_"|wc -c`
-    test $n_plus_50_ -le $len_ && break;
-    data_=` (echo "$data_"; eval "$cmds_") 2>&1 | gzip `
-  done
-
-  echo "$data_" \
-    | dd bs=1 skip=50 count=$n_ 2>/dev/null \
-    | LC_ALL=C tr -c $chars_ 01234567$chars_$chars_$chars_
-}
-
-mktempd_ ()
-{
-  case $# in
-  2);;
-  *) fail_ "Usage: mktempd_ DIR TEMPLATE";;
-  esac
-
-  destdir_=$1
-  template_=$2
-
-  MAX_TRIES_=4
-
-  # Disallow any trailing slash on specified destdir:
-  # it would subvert the post-mktemp "case"-based destdir test.
-  case $destdir_ in
-  / | //) destdir_slash_=$destdir;;
-  */) fail_ "invalid destination dir: remove trailing slash(es)";;
-  *) destdir_slash_=$destdir_/;;
-  esac
-
-  case $template_ in
-  *XXXX) ;;
-  *) fail_ \
-       "invalid template: $template_ (must have a suffix of at least 4 X's)";;
-  esac
-
-  # First, try to use mktemp.
-  d=`unset TMPDIR; { mktemp -d -t -p "$destdir_" "$template_"; } 2>/dev/null` &&
-
-  # The resulting name must be in the specified directory.
-  case $d in "$destdir_slash_"*) :;; *) false;; esac &&
-
-  # It must have created the directory.
-  test -d "$d" &&
-
-  # It must have 0700 permissions.  Handle sticky "S" bits.
-  perms=`ls -dgo "$d" 2>/dev/null` &&
-  case $perms in drwx--[-S]---*) :;; *) false;; esac && {
-    echo "$d"
-    return
-  }
-
-  # If we reach this point, we'll have to create a directory manually.
-
-  # Get a copy of the template without its suffix of X's.
-  base_template_=`echo "$template_"|sed 's/XX*$//'`
-
-  # Calculate how many X's we've just removed.
-  template_length_=`echo "$template_" | wc -c`
-  nx_=`echo "$base_template_" | wc -c`
-  nx_=`expr $template_length_ - $nx_`
-
-  err_=
-  i_=1
-  while :; do
-    X_=`rand_bytes_ $nx_`
-    candidate_dir_="$destdir_slash_$base_template_$X_"
-    err_=`mkdir -m 0700 "$candidate_dir_" 2>&1` \
-      && { echo "$candidate_dir_"; return; }
-    test $MAX_TRIES_ -le $i_ && break;
-    i_=`expr $i_ + 1`
-  done
-  fail_ "$err_"
-}
-
-# If you want to override the testdir_prefix_ function,
-# or to add more utility functions, use this file.
-test -f "$srcdir/init.cfg" \
-  && . "$srcdir/init.cfg"
-
-setup_ "$@"
-# This trap is here, rather than in the setup_ function, because some
-# shells run the exit trap at shell function exit, rather than script exit.
-trap remove_tmp_ 0

tests/invalid-re

@@ -7,7 +7,7 @@ fail=0
 
 echo > a || framework_failure_
 echo b > b || framework_failure_
-echo 'diff: \: Trailing backslash' > exp-err || framework_failure_
+echo "diff: '\': Trailing backslash" > exp-err || framework_failure_
 
 # This must fail with an exit status of 2:
 returns_ 2 diff -Ix -I\\ a b > out 2> err || fail=1

tests/large-subopt

@@ -0,0 +1,27 @@
+#!/bin/sh
+# This test is expected to fail at least with diffutils-3.6.
+# Demonstrate how diff can produce suboptimal output.
+# With these two files, diff -u prints output including this:
+# -2
+# +L: 361
+# +L: 361
+# +2
+# The trouble is that "2" line that is both added and removed.
+# This smaller patch could induce the same change:
+# +L: 361
+# +L: 361
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+require_perl_
+
+fail=0
+
+diff -u \
+  "$abs_top_srcdir/tests/large-subopt.in1" \
+  "$abs_top_srcdir/tests/large-subopt.in2" \
+  | $PERL -n0 -e \
+     '/\n-2\n(\+L: 361\n){2}\+2\n/ and do {$e=1; last}; END{exit !$e}' \
+     && fail=1
+
+Exit $fail

tests/new-file

@@ -10,10 +10,7 @@ echo a > a || fail=1
 echo '0a1
 > a' > exp || fail=1
 
-returns_ 1 diff -N - a <&- > out || fail=1
-compare exp out || fail=1
-
-returns_ 1 diff --unidirectional-new-file - a <&- > out || fail=1
+returns_ 1 diff -N b a > out || fail=1
 compare exp out || fail=1
 
 returns_ 1 diff -N b - < a > out || fail=1
@@ -25,14 +22,6 @@ compare exp out || fail=1
 echo '1d0
 < a' > exp || fail=1
 
-returns_ 1 diff -N a - <&- > out || fail=1
-compare exp out || fail=1
-
-# With closed standard input, require an exit status of 2
-# and empty stdout.
-returns_ 2 diff --unidirectional-new-file a - <&- > out || fail=1
-compare /dev/null out || fail=1
-
 returns_ 1 diff -N - b < a > out || fail=1
 compare exp out || fail=1
 

tests/no-dereference

@@ -13,7 +13,7 @@ ln -s regular3 symlink3
 
 # Non-recursive comparisons.
 
-# Test case 3: Compare regular file with regular file.
+# Compare regular file with regular file.
 returns_ 1 diff --no-dereference regular1 regular2 > out || fail=1
 cat <<EOF > expected || framework_failure_
 1c1
@@ -23,43 +23,47 @@ cat <<EOF > expected || framework_failure_
 EOF
 compare expected out || fail=1
 
-# Test case 4: Compare regular file with symbolic link.
+# Compare regular file with symbolic link.
 returns_ 1 diff --no-dereference regular1 symlink1 > out || fail=1
 cat <<EOF > expected || framework_failure_
 File regular1 is a regular file while file symlink1 is a symbolic link
 EOF
 compare expected out || fail=1
 
-# Test case 5: Compare symbolic link with regular file.
+# Compare symbolic link with regular file.
 returns_ 1 diff --no-dereference symlink1 regular1 > out || fail=1
 cat <<EOF > expected || framework_failure_
 File symlink1 is a symbolic link while file regular1 is a regular file
 EOF
 compare expected out || fail=1
 
-# Test case 6: Compare symbolic links with same value.
+# Compare symbolic links with same value.
 diff --no-dereference symlink1 symlink1bis > out || fail=1
 compare /dev/null out || fail=1
 
-# Test case 7: Compare symbolic links with different value and different target
-# contents.
-returns_ 1 diff --no-dereference symlink1 symlink2 > out || fail=1
+# Compare symbolic links with different value and different target contents.
+LC_ALL=C returns_ 1 diff --no-dereference symlink1 symlink2 > out || fail=1
 cat <<EOF > expected || framework_failure_
-Symbolic links symlink1 and symlink2 differ
+Symbolic links 'symlink1' -> 'regular1' and 'symlink2' -> 'regular2' differ
 EOF
 compare expected out || fail=1
 
-# Test case 8: Compare symbolic links with different value and same target
-# contents.
-returns_ 1 diff --no-dereference symlink2 symlink3 > out || fail=1
+# Compare symbolic links with different value and same target contents.
+LC_ALL=C returns_ 1 diff --no-dereference symlink2 symlink3 > out || fail=1
 cat <<EOF > expected || framework_failure_
-Symbolic links symlink2 and symlink3 differ
+Symbolic links 'symlink2' -> 'regular2' and 'symlink3' -> 'regular3' differ
 EOF
 compare expected out || fail=1
 
+mkdir subdir &&
+ln -s loop loop &&
+ln -s loop subdir/loop || framework_failure_
+diff --no-dereference loop subdir > out || fail=1
+compare /dev/null out || fail=1
+
 # Recursive comparisons.
 
-# Test case 1: Compare symbolic link with nonexistent file.
+# Compare symbolic link with nonexistent file.
 mkdir subdir1a
 mkdir subdir1b
 ln -s nonexistent subdir1a/foo
@@ -71,7 +75,7 @@ Only in subdir1a: foo
 EOF
 compare expected out || fail=1
 
-# Test case 1: Compare nonexistent file with symbolic link.
+# Compare nonexistent file with symbolic link.
 mkdir subdir2a
 mkdir subdir2b
 ln -s nonexistent subdir2b/foo
@@ -83,7 +87,7 @@ Only in subdir2b: foo
 EOF
 compare expected out || fail=1
 
-# Test case 3: Compare regular file with regular file.
+# Compare regular file with regular file.
 mkdir subdir3a
 mkdir subdir3b
 cp regular1 subdir3a/foo
@@ -98,7 +102,7 @@ diff -r --no-dereference subdir3a/foo subdir3b/foo
 EOF
 compare expected out || fail=1
 
-# Test case 4: Compare regular file with symbolic link.
+# Compare regular file with symbolic link.
 mkdir subdir4a
 mkdir subdir4b
 cp regular1 subdir4a/foo
@@ -109,7 +113,7 @@ File subdir4a/foo is a regular file while file subdir4b/foo is a symbolic link
 EOF
 compare expected out || fail=1
 
-# Test case 5: Compare symbolic link with regular file.
+# Compare symbolic link with regular file.
 mkdir subdir5a
 mkdir subdir5b
 ln -s ../regular1 subdir5a/foo
@@ -120,7 +124,7 @@ File subdir5a/foo is a symbolic link while file subdir5b/foo is a regular file
 EOF
 compare expected out || fail=1
 
-# Test case 6: Compare symbolic links with same value.
+# Compare symbolic links with same value.
 mkdir subdir6a
 mkdir subdir6b
 ln -s ../regular1 subdir6a/foo
@@ -128,27 +132,25 @@ ln -s ../regular1 subdir6b/foo
 diff -r --no-dereference subdir6a subdir6b > out || fail=1
 compare /dev/null out || fail=1
 
-# Test case 7: Compare symbolic links with different value and different target
-# contents.
+# Compare symbolic links with different value and different target contents.
 mkdir subdir7a
 mkdir subdir7b
 ln -s ../regular1 subdir7a/foo
 ln -s ../regular2 subdir7b/foo
-returns_ 1 diff -r --no-dereference subdir7a subdir7b > out || fail=1
+LC_ALL=C returns_ 1 diff -r --no-dereference subdir7a subdir7b > out || fail=1
 cat <<EOF > expected || framework_failure_
-Symbolic links subdir7a/foo and subdir7b/foo differ
+Symbolic links 'subdir7a/foo' -> '../regular1' and 'subdir7b/foo' -> '../regular2' differ
 EOF
 compare expected out || fail=1
 
-# Test case 8: Compare symbolic links with different value and same target
-# contents.
+# Compare symbolic links with different value and same target contents.
 mkdir subdir8a
 mkdir subdir8b
 ln -s ../regular2 subdir8a/foo
 ln -s ../regular3 subdir8b/foo
-returns_ 1 diff -r --no-dereference subdir8a subdir8b > out || fail=1
+LC_ALL=C returns_ 1 diff -r --no-dereference subdir8a subdir8b > out || fail=1
 cat <<EOF > expected || framework_failure_
-Symbolic links subdir8a/foo and subdir8b/foo differ
+Symbolic links 'subdir8a/foo' -> '../regular2' and 'subdir8b/foo' -> '../regular3' differ
 EOF
 compare expected out || fail=1
 

tests/side-by-side

@@ -0,0 +1,52 @@
+#!/bin/sh
+# Test side-by-side output with non-ASCII characters.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+require_utf8_locale_
+
+fail=0
+
+cat >in1 <<\EOF || framework_failure_
+a
+ab
+abc
+[一天早上，當格雷戈爾·薩姆沙從不安的夢中醒來時，
+ 他發現自己在床上變成了一隻可怕的害蟲。
+]
+EOF
+
+cat >in2 <<\EOF || framework_failure_
+ab
+abcd
+<一天早上，當格雷戈爾·薩姆沙從不安的夢中醒來時，
+ 他發現自己在床上變成了一隻可怕的害蟲。
+>
+EOF
+
+tr '_@' ' \t' >exp <<'EOF' || framework_failure_
+a@@_____<
+ab@@@ab
+abc@@_____|@abcd
+[一天早上，當格雷戈__|@<一天早上，當格雷戈
+_他發現自己在床上變@_他發現自己在床上變
+]@@_____|@>
+EOF
+
+returns_ 1 diff -y -W 44 in1 in2 >out 2>err || fail=1
+compare exp out || fail=1
+compare /dev/null err || fail=1
+
+cat >exp <<'EOF' || framework_failure_
+a                    <
+ab                      ab
+abc                  |  abcd
+[一天早上，當格雷戈  |  <一天早上，當格雷戈
+ 他發現自己在床上變      他發現自己在床上變
+]                    |  >
+EOF
+
+returns_ 1 diff --expand-tabs -y -W 44 in1 in2 >out 2>err || fail=1
+compare exp out || fail=1
+compare /dev/null err || fail=1
+
+Exit $fail

tests/side-by-side-seq

@@ -0,0 +1,26 @@
+#!/bin/sh
+# Test side-by-side output on sequences.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+seq 1 100 >in1 || framework_failure_
+(seq 1 49 && seq 51 100) >in2 || framework_failure_
+awk '
+  BEGIN {
+    for (i = 1; i <= 100; i++) {
+      if (i == 50) {
+        print "50   <"
+      } else {
+        printf "%d\t%d\n", i, i
+      }
+    }
+  }
+' </dev/null >exp || framework_failure_
+
+returns_ 1 diff -yW 11 in1 in2 >out 2>err || fail=1
+compare exp out || fail=1
+compare /dev/null err || fail=1
+
+Exit $fail

tests/starting-file

@@ -0,0 +1,20 @@
+#!/bin/sh
+# Test diff -S.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+mkdir d1 d2 || framework_failure_
+
+for i in a abc abC aBc aBC z; do
+ echo $i >d1/$i || framework_failure_
+done
+
+for i in a ABC ABc AbC Abc; do
+ echo x$i >d2/$i || framework_failure_
+done
+echo z >d2/z || framework_failure_
+
+LC_ALL=C diff -S z d1 d2 || fail=1
+LC_ALL=C diff -S z --ignore-file-name-case d1 d2 || fail=1
+
+Exit $fail

tests/stdin

@@ -5,7 +5,9 @@
 
 fail=0
 
-cat <<EOF > exp || fail=1
+echo a > a || framework_failure_
+echo b > b || framework_failure_
+cat <<'EOF' > exp || framework_failure_
 --- -
 +++ b
 @@ -1 +1 @@
@@ -13,12 +15,17 @@ cat <<EOF > exp || fail=1
 +b
 EOF
 
-echo a > a
-echo b > b
-
 returns_ 1 diff -u - b < a > out 2> err || fail=1
 # Remove date and time.
-sed -e 's/^\([-+*][-+*][-+*] [^	]*\)	.*/\1/' out > k; mv k out
-compare exp out || fail=1
+sed -e 's/^\([-+*][-+*][-+*] [^	]*\)	.*/\1/' out >outk || framework_failure_
+compare exp outk || fail=1
+compare /dev/null err || fail=1
+
+mkdir d || framework_failure_
+echo a >d/a || framework_failure_
+
+diff -u - a <d >out 2>err || fail=1
+compare /dev/null out || fail=1
+compare /dev/null err || fail=1
 
 Exit $fail

tests/strip-trailing-cr

@@ -0,0 +1,25 @@
+#!/bin/sh
+# Before diff-3.7, this would provoke a UMR
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+require_valgrind_
+
+printf '\r' > r || framework_failure_
+echo b > b || framework_failure_
+
+# Use valgrind only when no Sanitizer is in use:
+ASAN_OPTIONS=help=true diff --help 2>&1 |grep Sanitizer > /dev/null \
+  && eval 'valgrind() { "$@"; }' \
+  || eval 'valgrind() { env valgrind --quiet --error-exitcode=3 "$@"; }'
+
+for opt in '' u c Dfoo; do
+  returns_ 1 valgrind \
+    diff -a$opt --strip-trailing-cr r b > out 2> err || fail=1
+done
+
+compare /dev/null err || fail=1
+
+Exit $fail

tests/timezone

@@ -0,0 +1,14 @@
+#!/bin/sh
+# In diff 3.4 through 3.8, this would output the wrong timezone on Solaris.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+echo a >a || fail=1
+case $(LC_ALL=C TZ=EST5 diff -u /dev/null a) in
+  *' -0500'*) ;;
+  *) fail=1 ;;
+esac
+
+Exit $fail

tests/y2038-vs-32bit

@@ -0,0 +1,13 @@
+#!/bin/sh
+# dates after Y2K38 are safe
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+
+fail=0
+
+touch -t 203901010000 in || skip_ 'touch -t 2039... failed'
+
+cmp in in || fail=1
+diff in in || fail=1
+
+Exit $fail