From eabf8062472dd0ffebb09980b3afcbb2b8c49c86 Mon Sep 17 00:00:00 2001
From: "Thomas E. Dickey" <dickey@invisible-island.net>
Date: Sun, 21 Feb 2010 02:05:44 -0500
Subject: [PATCH] snapshot of project "mawk", label t20100220

---
 CHANGES           |  12 +++-
 MANIFEST          |   4 +-
 files.c           |   4 +-
 fin.c             |   4 +-
 man/Makefile      |  41 ++++++++++++++
 patchlev.h        |   4 +-
 rexp3.c           |  10 +++-
 test/mawktest     | 141 ++++++++++++++++++++++++++++++----------------
 test/mawktest.bat |   3 +-
 test/reg-awk.out  |   1 +
 test/reg3.awk     |   5 ++
 11 files changed, 169 insertions(+), 60 deletions(-)
 create mode 100644 man/Makefile
 create mode 100644 test/reg3.awk

diff --git a/CHANGES b/CHANGES
index 900f73f..7cafd99 100644
--- a/CHANGES
+++ b/CHANGES
@@ -1,8 +1,16 @@
--- $MawkId: CHANGES,v 1.100 2010/02/17 11:48:46 tom Exp $
+-- $MawkId: CHANGES,v 1.102 2010/02/21 00:19:40 tom Exp $
 
 Changes by Thomas E Dickey <dickey@invisible-island.net>
 
-20100217
+20100220
+	+ modify mawktest script to report results from all tests, rather than
+	  halting on the first failure.
+
+	+ add limit-check after processing match(test, "[^0-9A-Za-z]") to
+	  ensure the internal trailing null of the test-string is not mistaken
+	  for part of the string, i.e., RSTART, RLENGTH are computed correctly
+	  (report by Markus Gnam).
+
 	+ modify parsing of -W option to use comma-separated values, e.g.,
 	  "-Wi,e" for "-Winteractive" and "-Werror".
 
diff --git a/MANIFEST b/MANIFEST
index 83d403c..d55752d 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -1,4 +1,4 @@
-MANIFEST for mawk, version t20100217
+MANIFEST for mawk, version t20100220
 --------------------------------------------------------------------------------
 MANIFEST                        this file
 ACKNOWLEDGMENT                  acknowledgements
@@ -90,6 +90,7 @@ examples/nocomment.awk          remove C comments  from a list of files
 examples/primes.awk             find all primes between 2 and STOP (parameter)
 examples/qsort.awk              qsort text files
 man                             subdirectory
+man/Makefile                    makefile for alternate doc-formats
 man/TODO                        to-do list (update to conform with POSIX-1003.1-2004)
 man/mawk.1                      troff source for unix style man pages
 man/mawk.doc                    ascii man pages
@@ -139,6 +140,7 @@ test/reg-awk.out                reference for testing regular expressions
 test/reg0.awk                   test simple pattern matching
 test/reg1.awk                   test pattern with "|" OR
 test/reg2.awk                   test pattern with ranges
+test/reg3.awk                   testcase for pattern
 test/reg4.awk                   test-case for square-brackets and embedded "/" or "]"
 test/reg5.awk                   test-case for character-classes
 test/reg6.awk                   testcase for J2C machine state
diff --git a/files.c b/files.c
index 825e825..e0fc96a 100644
--- a/files.c
+++ b/files.c
@@ -10,7 +10,7 @@ the GNU General Public License, version 2, 1991.
 ********************************************/
 
 /*
- * $MawkId: files.c,v 1.12 2009/12/17 00:06:51 tom Exp $
+ * $MawkId: files.c,v 1.13 2010/02/21 01:32:50 tom Exp $
  * @Log: files.c,v @
  * Revision 1.9  1996/01/14  17:14:10  mike
  * flush_all_output()
@@ -506,7 +506,7 @@ remove_from_child_list(int pid)
 int
 wait_for(int pid)
 {
-    int exit_status;
+    int exit_status = 0;
     struct child *p;
     int id;
 
diff --git a/fin.c b/fin.c
index 4106ae0..5296964 100644
--- a/fin.c
+++ b/fin.c
@@ -10,7 +10,7 @@ the GNU General Public License, version 2, 1991.
 ********************************************/
 
 /*
- * $MawkId: fin.c,v 1.23 2010/01/29 20:03:43 Jan.Psota Exp $
+ * $MawkId: fin.c,v 1.24 2010/02/21 01:45:21 tom Exp $
  * @Log: fin.c,v @
  * Revision 1.10  1995/12/24  22:23:22  mike
  * remove errmsg() from inside FINopen
@@ -317,7 +317,7 @@ FINgets(FIN * fin, unsigned *len_p)
 	unsigned rr;
 	unsigned amount = (unsigned) (fin->limit - p);
 
-	p = (char *) memcpy(fin->buff, p, r = (unsigned) (fin->limit - p));
+	p = (char *) memmove(fin->buff, p, r = (unsigned) (fin->limit - p));
 	q = p + r;
 	rr = fin->nbuffs * BUFFSZ - r;
 
diff --git a/man/Makefile b/man/Makefile
new file mode 100644
index 0000000..ba287e9
--- /dev/null
+++ b/man/Makefile
@@ -0,0 +1,41 @@
+# $MawkId: Makefile,v 1.1 2010/02/17 21:49:26 tom Exp $
+# produce alternate forms from mawk's documentation.
+
+SHELL=/bin/sh
+
+PROG	= mawk
+manext	= 1
+
+.SUFFIXES : .html .$(manext) .man .ps .pdf .doc .txt
+
+.$(manext).html :
+	GROFF_NO_SGR=stupid $(SHELL) -c "tbl $*.$(manext) | groff -Thtml -man" >$@
+
+.$(manext).ps :
+	$(SHELL) -c "tbl $*.$(manext) | groff -man" >$@
+
+.$(manext).doc :
+	GROFF_NO_SGR=stupid $(SHELL) -c "tbl $*.$(manext) | nroff -Tascii -man" >$@
+
+.$(manext).txt :
+	GROFF_NO_SGR=stupid $(SHELL) -c "tbl $*.$(manext) | nroff -Tascii -man | col -bx" >$@
+
+.ps.pdf :
+	ps2pdf $*.ps
+
+ALL	= \
+	$(PROG).html \
+	$(PROG).pdf \
+	$(PROG).ps \
+	$(PROG).txt \
+
+all:	$(PROG).doc $(ALL)
+
+clean:
+	rm -f $(ALL)
+
+distclean: clean
+
+maintainer-clean:
+
+$(PROG).pdf : $(PROG).ps
diff --git a/patchlev.h b/patchlev.h
index 0acaef2..1e93aeb 100644
--- a/patchlev.h
+++ b/patchlev.h
@@ -1,6 +1,6 @@
 /*
- * $MawkId: patchlev.h,v 1.17 2010/02/17 11:41:46 tom Exp $
+ * $MawkId: patchlev.h,v 1.18 2010/02/21 00:18:47 tom Exp $
  */
 #define  PATCHLEVEL	3
 #define  PATCH_STRING	".4"
-#define  DATE_STRING    "20100217"
+#define  DATE_STRING    "20100220"
diff --git a/rexp3.c b/rexp3.c
index 211f471..6003d11 100644
--- a/rexp3.c
+++ b/rexp3.c
@@ -10,7 +10,7 @@ the GNU General Public License, version 2, 1991.
 ********************************************/
 
 /*
- * $MawkId: rexp3.c,v 1.19 2010/02/17 11:38:24 Jonathan.Nieder Exp $
+ * $MawkId: rexp3.c,v 1.22 2010/02/21 02:05:44 tom Exp $
  * @Log: rexp3.c,v @
  * Revision 1.3  1993/07/24  17:55:15  mike
  * more cleanup
@@ -157,7 +157,10 @@ REmatch(char *str,		/* string to test */
 	goto reswitch;
 
     case M_STR + U_ON + END_OFF:
-	if (!(s = str_str(s, str_len, m->s_data.str, m->s_len))) {
+	if (s >= str_end) {
+	    goto refill;
+	}
+	if (!(s = str_str(s, (unsigned) (str_end - s), m->s_data.str, m->s_len))) {
 	    goto refill;
 	}
 	if (s >= str + strlen(str)) {
@@ -237,6 +240,9 @@ REmatch(char *str,		/* string to test */
 		s++;
 	    }
 	}
+	if (s >= str_end) {
+	    goto refill;
+	}
 	s++;
 	push(m, s, sp, ss, U_ON);
 	if (!ss) {
diff --git a/test/mawktest b/test/mawktest
index e42f043..beb9d5e 100755
--- a/test/mawktest
+++ b/test/mawktest
@@ -1,5 +1,5 @@
 #!/bin/sh
-# $MawkId: mawktest,v 1.24 2010/02/17 11:38:24 Jonathan.Nieder Exp $
+# $MawkId: mawktest,v 1.27 2010/02/21 01:06:26 tom Exp $
 
 # This is a simple test that a new made mawk seems to
 # be working OK.
@@ -9,6 +9,46 @@
 # It needs to be run from  mawk/test
 # and mawk needs to be in mawk/test or in PATH
 
+# POSIX shells have functions...
+Fail() {
+	echo "?? fail $*"
+	FAIL=`expr $FAIL + 1`
+	ERRS=`expr $ERRS + 1`
+}
+
+Begin() {
+	echo
+	echo "$*"
+	OK=OK
+	FAIL=0
+}
+
+Finish() {
+	if test $FAIL = 1
+	then
+		echo "$* had one failure"
+	elif test $FAIL != 0
+	then
+		echo "$* had $FAIL failures"
+	else
+		echo "$* OK"
+	fi
+}
+
+ERRS=0
+
+Summary() {
+	if test $ERRS = 1
+	then
+		echo "$* had one failure"
+	elif test $ERRS != 0
+	then
+		echo "$* had $ERRS failures"
+	else
+		echo "$* OK"
+	fi
+}
+
 PROG=mawk
 
 MAWKBINMODE=7
@@ -26,7 +66,9 @@ nulldat=mawknull.dat
 STDOUT=${TMPDIR-/tmp}/mawk-out$$
 STDERR=${TMPDIR-/tmp}/mawk-err$$
 
-ulimit -v 25000
+# The ulimit command interfers with valgrind (uncomment for ad hoc testing).
+#ulimit -v 25000
+
 trap 'echo mawk_test failed ; rm -f $STDOUT $STDERR; exit 1'  0
 
 PATH=.:$PATH
@@ -36,60 +78,65 @@ $PROG -W version
 
 NULLS=`$PROG -W version 2>&1 |fgrep 'internal regex' 2>/dev/null`
 #################################
-echo
-echo testing input and field splitting
+Begin "testing input and field splitting"
 
-LC_ALL=C $PROG -f wc.awk $dat | cmp -s - wc-awk.out || exit
-LC_ALL=C $PROG -f null-rs.awk null-rs.dat | cmp -s - null-rs.out || exit
+LC_ALL=C $PROG -f wc.awk $dat | cmp -s - wc-awk.out || Fail "wc.awk"
+LC_ALL=C $PROG -f null-rs.awk null-rs.dat | cmp -s - null-rs.out || Fail "null-rs.awk"
 
 LC_ALL=C $PROG -F '(a?)*b' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F 'a*b' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F 'a*b' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 2"
 
 LC_ALL=C $PROG -F '(a?)+b' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F 'a*b' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F 'a*b' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 3"
 
 LC_ALL=C $PROG -F '[^^]' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F '(.)' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F '(.)' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 4"
 
 LC_ALL=C $PROG -F '[^]]' -f wc.awk $dat > $STDOUT
 LC_ALL=C $PROG -F '[[#a-zA-Z0-9/*!=<>+,;.&_%(){}" -]' -f wc.awk $dat |
-	cmp -s - $STDOUT || exit
+	cmp -s - $STDOUT || Fail "case 5"
 
 LC_ALL=C $PROG -F '[a[]' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F '[[a]' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F '[[a]' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 6"
 
 LC_ALL=C $PROG -F '(])' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F '[]]' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F '[]]' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 7"
 
 # check that the regexp [\ does not make mawk segfault
-LC_ALL=C $PROG -F '[\' 2> $STDERR || exit
+LC_ALL=C $PROG -F '[\' 2> $STDERR || Fail "case 8"
 
 LC_ALL=C $PROG -F '(^)?)' -f wc.awk $dat > $STDOUT
-LC_ALL=C $PROG -F ')' -f wc.awk $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG -F ')' -f wc.awk $dat | cmp -s - $STDOUT || Fail "case 9"
 
 echo baaab | LC_ALL=C $PROG -F 'a*+' '{print NF}' > $STDOUT
-echo baaab | LC_ALL=C $PROG -F 'a*' '{print NF}' | cmp -s - $STDOUT || exit
+echo baaab | LC_ALL=C $PROG -F 'a*' '{print NF}' | cmp -s - $STDOUT || Fail "case 10"
 
 if test -n "$NULLS" ; then
 	LC_ALL=C $PROG -F '\000'    -f nulls0.awk $nulldat > $STDOUT
 	LC_ALL=C $PROG -F '[\000 ]' -f nulls0.awk $nulldat >> $STDOUT
-	cmp -s  nulls.out $STDOUT || exit
-	echo "... $PROG supports matches with NUL bytes"
+	if ( cmp -s  nulls.out $STDOUT )
+	then
+		echo "... $PROG supports matches with NUL bytes"
+	else
+		echo "... $PROG does NOT supports matches with NUL bytes"
+	fi
 fi
 
-echo input and field splitting OK
+Finish "input and field splitting"
+
 #####################################
 
-echo
-echo testing regular expression matching
+Begin "testing regular expression matching"
+
 LC_ALL=C $PROG -f reg0.awk $dat >  $STDOUT
 LC_ALL=C $PROG -f reg1.awk $dat >> $STDOUT
 LC_ALL=C $PROG -f reg2.awk $dat >> $STDOUT
+LC_ALL=C $PROG -f reg3.awk $dat >> $STDOUT
 LC_ALL=C $PROG -f reg4.awk $dat >> $STDOUT
 LC_ALL=C $PROG -f reg5.awk $dat >> $STDOUT
 LC_ALL=C $PROG -f reg6.awk $dat >> $STDOUT
 
-cmp -s  reg-awk.out $STDOUT || exit
+cmp -s  reg-awk.out $STDOUT || Fail "reg0-reg6 case"
 
 # 640 backslashes
 backslashes='\\\\\\\\\\'
@@ -97,68 +144,66 @@ backslashes="$backslashes$backslashes$backslashes$backslashes"
 backslashes="$backslashes$backslashes$backslashes$backslashes"
 backslashes="$backslashes$backslashes$backslashes$backslashes"
 backslashes="$backslashes$backslashes$backslashes$backslashes"
-( set +e; LC_ALL=C $PROG "/a$backslashes/" $dat 2> $STDERR ; test $? = 2 ) || exit
+( set +e; LC_ALL=C $PROG "/a$backslashes/" $dat 2> $STDERR ; test $? = 2 ) || Fail "buffer-overflow"
 
 # To be uncommented when they stop taking so long to run.
 echo "''Italics with an apostrophe' embedded''" |
-	LC_ALL=C $PROG -f noloop.awk || exit
+	LC_ALL=C $PROG -f noloop.awk || Fail "noloop2 test"
 #echo "''Italics with an apostrophe'' embedded''" |
-#	LC_ALL=C $PROG -f noloop.awk || exit
+#	LC_ALL=C $PROG -f noloop.awk || Fail "noloop1 test"
 
 LC_ALL=C $PROG '/^[^^]*$/' $dat > $STDOUT
-cmp -s $dat $STDOUT || exit
+cmp -s $dat $STDOUT || Fail "case 1"
 
 LC_ALL=C $PROG '!/^[^]]*$/' $dat > $STDOUT
-LC_ALL=C $PROG '/]/' $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG '/]/' $dat | cmp -s - $STDOUT || Fail "case 2"
 
 LC_ALL=C $PROG '/[a[]/' $dat > $STDOUT
-LC_ALL=C $PROG '/[[a]/' $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG '/[[a]/' $dat | cmp -s - $STDOUT || Fail "case 3"
 
 LC_ALL=C $PROG '/]/' $dat > $STDOUT
-LC_ALL=C $PROG '/[]]/' $dat | cmp -s - $STDOUT || exit
+LC_ALL=C $PROG '/[]]/' $dat | cmp -s - $STDOUT || Fail "case 4"
 
 echo aaa | LC_ALL=C $PROG '/a*+/' > $STDOUT
-echo aaa | LC_ALL=C $PROG '/a*/' | cmp -s - $STDOUT || exit
-echo aaa | cmp -s - $STDOUT || exit
+echo aaa | LC_ALL=C $PROG '/a*/' | cmp -s - $STDOUT || Fail "case 5"
+echo aaa | cmp -s - $STDOUT || Fail "case 6"
+
+Finish "regular expression matching"
 
-echo regular expression matching OK
 #######################################
 
-echo
 if [ -c /dev/full ]; then
-    echo testing checking for write errors
+    Begin "testing checking for write errors"
     # Check for write errors noticed when closing the file
-    LC_ALL=C $PROG '{print}' <full-awk.dat >/dev/full 2>/dev/null && exit
+    LC_ALL=C $PROG '{print}' <full-awk.dat >/dev/full 2>/dev/null && Fail "case 1"
     # Check for write errors noticed on writing
     # The file has to be bigger than the buffer size of the libc
-    LC_ALL=C $PROG '{print}' <$SRC/scan.c >/dev/full 2>/dev/null && exit
+    LC_ALL=C $PROG '{print}' <$SRC/scan.c >/dev/full 2>/dev/null && Fail "case 2"
 
-    echo checking for write errors OK
+    Finish "checking for write errors"
 else
+    echo
     echo "No /dev/full - check for write errors skipped"
 fi
 
 #######################################
 
-echo
-echo testing arrays and flow of control
+Begin "testing arrays and flow of control"
 
-LC_ALL=C $PROG -f wfrq0.awk $dat | cmp -s - wfrq-awk.out || exit
+LC_ALL=C $PROG -f wfrq0.awk $dat | cmp -s - wfrq-awk.out || Fail
+
+Finish "array test"
 
-echo array test OK
 #################################
 
-echo
-echo testing function calls and general stress test
+Begin "testing function calls and general stress test"
 
-LC_ALL=C $PROG -f $SRC/examples/decl.awk $dat | cmp -s - decl-awk.out || exit
+LC_ALL=C $PROG -f $SRC/examples/decl.awk $dat | cmp -s - decl-awk.out || Fail
 
-echo general stress test passed
+Finish "general stress test"
 
-
-echo
-echo  tested $PROG seems OK
+Summary "tested $PROG"
 
 trap 0
 rm -f $STDOUT
-exit 0
+exit $ERRS
diff --git a/test/mawktest.bat b/test/mawktest.bat
index d819586..2e1d267 100644
--- a/test/mawktest.bat
+++ b/test/mawktest.bat
@@ -1,5 +1,5 @@
 echo off
-rem $MawkId: mawktest.bat,v 1.4 2009/07/27 18:59:23 tom Exp $
+rem $MawkId: mawktest.bat,v 1.5 2010/02/21 01:04:23 tom Exp $
 rem vile:rs=lf
 rem
 rem  This is a simple test that a new made mawk seems to
@@ -35,6 +35,7 @@ echo testing regular expression matching
 ..\mawk -f reg0.awk %dat% > temp$$
 ..\mawk -f reg1.awk %dat% >> temp$$
 ..\mawk -f reg2.awk %dat% >> temp$$
+..\mawk -f reg3.awk %dat% >> temp$$
 ..\mawk -f reg4.awk %dat% >> temp$$
 ..\mawk -f reg5.awk %dat% >> temp$$
 %CMP% temp$$ reg-awk.out
diff --git a/test/reg-awk.out b/test/reg-awk.out
index 276c64a..927aecc 100644
--- a/test/reg-awk.out
+++ b/test/reg-awk.out
@@ -1,6 +1,7 @@
 3
 4
 1
+0 -1
 reg4.1<<:  >>
 reg4.2<<:  >>
 reg4.3<<:  >>
diff --git a/test/reg3.awk b/test/reg3.awk
new file mode 100644
index 0000000..e8676ae
--- /dev/null
+++ b/test/reg3.awk
@@ -0,0 +1,5 @@
+BEGIN {
+	test = "MFG"
+	match(test, "[^0-9A-Za-z]")
+	print RSTART, RLENGTH
+}