mirror/lua
1
0
Fork 0
mirror of https://github.com/lua/lua.git synced 2026-01-31 11:55:07 +00:00
Code Issues Packages Projects Releases Wiki Activity
5,809 Commits 4 Branches 58 Tags
Commit Graph

545 Commits

Author SHA1 Message Date
Roberto I
5b7d998764 External strings are as good as internal ones 
A '__mode' metafield and an "n" key both can be external strings.
 2025-11-11 14:40:30 -03:00
Roberto Ierusalimschy
c33bb08ffe Added some casts for 32-bit machines 
When both 'int' and 'l_obj' have 32 bits, an unsigned int needs a
cast to be assigned to 'l_obj'. (As long as 'l_obj' can count the
total memory used by the system, these casts should be safe.)
 2025-07-29 11:50:20 -03:00
Roberto Ierusalimschy
9386e49a31 New metatable in an all-weak table can fool the GC 
All-weak tables are not being revisited after being visited during
propagation; if it gets a new metatable after that, the new metatable
may not be marked.
 2025-06-16 16:29:32 -03:00
Roberto Ierusalimschy
8cd7ae7da0 Simpler code for 'traversetable' 
Check the mode in a separate function (getmode), instead of using
comma expressions inside the 'if' condition.
 2025-06-16 15:50:12 -03:00
Roberto Ierusalimschy
9b014d4bcd Details (typos in comments) 2025-04-23 11:36:09 -03:00
Roberto Ierusalimschy
3dbb1a4b89 In gen. GC, some gray objects stay in gray lists 
In generational collection, objects marked as touched1 stay in gray
lists between collections. This commit fixes a bug introduced in
commit 808976bb59.
 2025-04-15 17:00:30 -03:00
Roberto Ierusalimschy
808976bb59 Small correction in 'traverseweakvalue' 
After a weak table is traversed in the atomic phase, if it does not
have white values ('hasclears') it does not need to be retraversed
again. (Comments were correct, but code did not agree with them.)
 2025-03-12 12:35:36 -03:00
Roberto Ierusalimschy
fa1382b5cd Main thread is a regular field of global_State 
They were already allocated as a single block, so there is no need
for the global_State to point to its main thread.
 2025-01-31 13:51:38 -03:00
Roberto Ierusalimschy
d1e677c52b New type 'TStatus' for thread status/error codes 2025-01-30 11:41:39 -03:00
Roberto Ierusalimschy
2d8d5c74b5 Details 
New year (2024->2025), typos in comments
 2025-01-16 11:51:16 -03:00
Roberto Ierusalimschy
002beeebe7 New way to keep hints for table length 
Instead of using 'alimit' for keeping the size of the array and at
the same time being a hint for '#t', a table now keeps these two
values separate. The Table structure has a field 'asize' with the
size of the array, while the length hint is kept in the array itself.
That way, tables with no array part waste no space with that field.
Moreover, the space for the hint may have zero cost for small arrays,
if the array of tags plus the hint still fits in a single word.
 2024-11-29 17:26:20 -03:00
Roberto Ierusalimschy
a4762b6ffe 'objsize' returns 'l_mem' 
Sums of size_t may not fit in a size_t.
 2024-11-15 12:04:53 -03:00
Roberto Ierusalimschy
5ffcd458f0 Some changes in default GC parameters 2024-10-23 17:15:06 -03:00
Roberto Ierusalimschy
9b01da97e3 Small bug in 'luaE_luaE_statesize' 
Plus, function was renamed to 'luaE_threadsize'.
 2024-10-21 15:30:35 -03:00
Roberto Ierusalimschy
258355734d Better support in 'ltests' for tracing the GC 2024-10-21 15:18:20 -03:00
Roberto Ierusalimschy
d0815046d0 Some adjustments in transition minor->major 
Plus extra comments and other details.
 2024-10-18 17:10:20 -03:00
Roberto Ierusalimschy
3d54b42d59 'objsize' broke in smaller pieces 2024-09-30 14:01:42 -03:00
Roberto Ierusalimschy
ddfa1fbccf GC back to controling pace counting bytes 
Memory is the resource we want to save. Still to be reviewed again.
 2024-09-19 19:02:14 -03:00
Roberto Ierusalimschy
a04e0ffdb9 Rename of fields in global state that control GC 
All fields in the global state that control the pace of the garbage
collector prefixed with 'GC'.
 2024-09-06 14:38:39 -03:00
Roberto Ierusalimschy
0acd55898d Added gcc option '-Wconversion' 
No warnings for standard numerical types. Still pending alternative
numerical types.
 2024-07-27 13:32:59 -03:00
Roberto Ierusalimschy
262dc5729a Details 
Corrections in comments and manual. Added note in the manual about
local variables in the REPL.
 2024-05-08 17:50:10 -03:00
Roberto Ierusalimschy
86a8e74824 Details 2024-03-28 17:11:33 -03:00
Roberto Ierusalimschy
ba71060381 Removed "bulk operations" 
Negligible performance gains don't justify extra complexity.
 2024-03-15 11:23:35 -03:00
Roberto Ierusalimschy
3823fc6c81 Added "bulk operations" to arrays 
A few operations on arrays can be performed "in bulk", treating all
tags of a cell as a simple (or a few) word(s).
 2024-03-15 11:01:34 -03:00
Roberto Ierusalimschy
8eb0abc9db Removed uses of LUA_NUMTAGS 
That constant was already deprecated (see commit 6aabf4b15e7).
 2024-01-13 18:10:50 -03:00
Roberto Ierusalimschy
7827c40c49 A few more tweaks in the garbage collector 2024-01-10 14:45:58 -03:00
Roberto Ierusalimschy
e2cc179454 New option "setparms" for 'collectgarbage' 
The generational mode also uses the parameters for the incremental
mode in its major collections, so it should be easy to change those
parameters without having to change the GC mode.
 2023-12-22 14:48:07 -03:00
Roberto Ierusalimschy
ad0ea7813b GC parameters encoded as floating-point bytes 
This encoding brings more precision and a larger range for these
parameters.
 2023-12-20 16:25:20 -03:00
Roberto Ierusalimschy
666e95a66d Option 0 for step multiplier makes GC non-incremental 2023-12-20 11:06:27 -03:00
Roberto Ierusalimschy
ad73b33224 Check minor->major made at the end of a minor cycle 
It does not make sense to wait for another cycle to decide when much of
the information about creation of old objects is already available.
 2023-12-07 15:45:41 -03:00
Roberto Ierusalimschy
925fe8a0f2 First criteria for shifts minor<->major 2023-12-07 15:45:11 -03:00
Roberto Ierusalimschy
789e7acdea Major collections done incrementally 
Major collections do not need to "stop the world". Still pending:
criteria for shifts minor-major, shifts generational-incremental.
 2023-12-06 10:49:56 -03:00
Roberto Ierusalimschy
74b4013538 Removed macro 'changeage' 
It is simpler to use always 'setage'. The saving from 'changeage'
is too irrelevant.
 2023-12-01 16:42:01 -03:00
Roberto Ierusalimschy
63d68bd657 Comments detailing the ages for generational GC 
Plus other comments and small details.
 2023-11-29 16:22:09 -03:00
Roberto Ierusalimschy
024f9064f1 External strings 
Strings can use external buffers to store their contents.
 2023-11-09 17:05:42 -03:00
Roberto Ierusalimschy
7f4906f565 Towards external strings 
Long strings have a pointer to string contents.
 2023-11-08 13:24:38 -03:00
Roberto Ierusalimschy
37c215b43f Merge branch 'newarray' into nextversion 2023-11-07 17:26:15 -03:00
Roberto Ierusalimschy
fa075b7953 Merge branch 'master' into newarray 2023-11-03 15:39:14 -03:00
Roberto Ierusalimschy
08a077d673 Full implementation of new representation for arrays 2023-11-03 15:26:13 -03:00
Roberto Ierusalimschy
43c8e5bded Full abstraction for representation of array values 2023-10-30 14:25:59 -03:00
Roberto Ierusalimschy
c815c2f0eb Merge branch 'master' into nextversion 2023-08-23 15:14:03 -03:00
Roberto Ierusalimschy
9b4f39ab14 More disciplined use of 'getstr' and 'tsslen' 
We may want to add other string variants in the future; this change
documents better where the code may need to handle those variants.
 2023-08-17 15:59:28 -03:00
Roberto Ierusalimschy
6b51133a98 Thread stacks resized in the atomic phase 
Although stack resize can be a little expensive, it seems unusual to
have too many threads needing resize during one GC cycle. On the other
hand, the change allows full collections to skip the propagate phase,
going straight from a pause to the atomic phase.
 2023-07-13 14:55:46 -03:00
Roberto Ierusalimschy
9be74ccc21 Several functions turned 'static' 
Several functions that were already being used only inside their
own file have been declared as 'static'.
 2023-05-22 14:47:54 -03:00
Roberto Ierusalimschy
8dea54877a Do not avoid major collections when GCdebt is zero 
'collectgarbage("step")' (without an argument) does not have any
special meaning, it means "do a step with some default size".
 2022-12-29 15:41:07 -03:00
Roberto Ierusalimschy
5d8b5b9290 Changed signal of GC debt 
Positive debts seems more natural then negative ones.
 2022-12-13 15:45:57 -03:00
Roberto Ierusalimschy
40565b4a08 Revamp of GC parameters 
More uniformity when handling GC parameters + avoid divisions by 100
when applying them.
 2022-12-13 11:55:14 -03:00
Roberto Ierusalimschy
ff106c028c Merge branch 'master' into nextversion 2022-12-12 14:08:55 -03:00
Roberto Ierusalimschy
fa2f294dd1 Reduce calls to 'luaC_step' when GC is stopped 2022-12-09 16:35:19 -03:00
Roberto Ierusalimschy
0270c204c2 Simplification in handling of GC debt 
Each incremental step has always the same size (stepsize), and the
debt for next step also is always the same.
 2022-12-06 12:02:34 -03:00