Commit Graph

301 Commits

Author SHA1 Message Date
Roberto Ierusalimschy
a2a2abcba4 new function 'luaC_runtilstate' to advance GC until a "valid" state 2009-12-11 17:14:59 -02:00
Roberto Ierusalimschy
3c4d970a7b comment typos 2009-11-26 09:39:20 -02:00
Roberto Ierusalimschy
ce444bff33 (huge) simplification of GC management 2009-11-18 11:13:47 -02:00
Roberto Ierusalimschy
88eb901f81 registry and global table may be changed through the API without a
write barrier, so GC should visit them in the atomic phase.
2009-11-09 16:29:21 -02:00
Roberto Ierusalimschy
a921d81033 avoid using 'ttype' when there is an explicit test + macro 'checkdeadkey'
to avoid repetitions
2009-11-06 15:06:19 -02:00
Roberto Ierusalimschy
9756f56354 better control over accesses to TValue fields 2009-11-05 15:43:54 -02:00
Roberto Ierusalimschy
5bc91c6405 no more one environment per thread: all threads share a single global
environment
2009-10-23 17:12:19 -02:00
Roberto Ierusalimschy
5938212748 information about upvalues (where they come from) kept in Proto structure,
instead of sequence of pseudo-opcodes after OP_CLOSURE
2009-09-28 13:32:50 -03:00
Roberto Ierusalimschy
1829911d7c some operations may shrink g->totalbytes so g->estimate must be
more flexible
2009-09-28 10:50:34 -03:00
Roberto Ierusalimschy
55dc7fb240 bug: stack must be cleared until its end (including extra size) +
control of stack size moved to 'ldo.c'
2009-07-16 13:26:09 -03:00
Roberto Ierusalimschy
4a67e48611 new macro 'condmovestack' instead of 'condhardstacktests' 2009-06-08 16:35:59 -03:00
Roberto Ierusalimschy
019ebcb85f errors in finalizers are propagated with code LUA_ERRGCMM (ERRor in
__gc MetaMethod)
2009-05-21 17:06:11 -03:00
Roberto Ierusalimschy
673c456cba resize string hash table only when new size is smaller than current one 2009-04-29 14:09:41 -03:00
Roberto Ierusalimschy
e091a254df new way to GC stacks: the entire stack must be correct all the times;
the 'dead' part of a stack (after the top) must have only nil's, so
that 'top' may go up without cleaning the stack.
2009-04-28 16:04:36 -03:00
Roberto Ierusalimschy
4f88418170 'CallInfo' stack implemented as double-linked list instead of an array 2009-04-17 11:28:06 -03:00
Roberto Ierusalimschy
ba484b9eb1 yielding across lua_call (first version) 2009-03-10 14:14:37 -03:00
Roberto Ierusalimschy
d07abcc6c7 finalizers always called protected 2009-02-17 16:47:58 -03:00
Roberto Ierusalimschy
0c8f5fc2fd simplification in the handling of finalizers: no more 'tmudata' list +
no more GCSsweeptmu collector's state
2008-06-26 16:42:45 -03:00
Roberto Ierusalimschy
c3525610fe bug: when closing the state, 'luaC_separateudata' might mark
userdata in the wrong phase of collection, therefore avoiding
their traversal
2008-06-23 19:07:44 -03:00
Roberto Ierusalimschy
7ba62e2985 code reorganization (only changed comments and order of functions) 2008-06-23 13:51:28 -03:00
Roberto Ierusalimschy
e2b366c760 userdata with finalizers are kept in a separated list 2008-02-19 15:55:09 -03:00
Roberto Ierusalimschy
5f0a2f6248 more options for 'kinds' of Garbage Collection + more agressive
shrinking of stacks and string hash
2008-02-11 13:46:03 -02:00
Roberto Ierusalimschy
5e8dd55574 first implementation of ephemerons 2007-10-31 13:41:19 -02:00
Roberto Ierusalimschy
0e961ad47a some changes toward ephemerons 2007-10-29 14:51:20 -02:00
Roberto Ierusalimschy
a7c9e45c64 avoid trailing white spaces 2006-09-11 11:07:24 -03:00
Roberto Ierusalimschy
3ca9af51a4 emergency garbage collector (core forces a GC when allocation fails) 2006-07-11 12:53:29 -03:00
Roberto Ierusalimschy
0079efc479 bug: collector did not compensate enough for userdata with finalizers 2006-05-24 11:34:06 -03:00
Roberto Ierusalimschy
c505f341d6 small changes in casts 2005-12-22 14:19:56 -02:00
Roberto Ierusalimschy
4ff1e5b4e1 details 2005-08-24 14:06:36 -03:00
Roberto Ierusalimschy
680fc88cc2 a small bug 2005-08-04 10:37:38 -03:00
Roberto Ierusalimschy
746a1d612b small bug (type error) 2005-06-07 15:53:45 -03:00
Roberto Ierusalimschy
90de38bf1f warnings in VS .Net 2005-05-31 11:25:18 -03:00
Roberto Ierusalimschy
6cf85dcc90 metatables for all types 2005-05-05 12:34:03 -03:00
Roberto Ierusalimschy
390256edf7 better name for GC pause 2005-03-22 13:04:29 -03:00
Roberto Ierusalimschy
4494094031 detail 2005-03-16 17:02:48 -03:00
Roberto Ierusalimschy
370d31a559 luac´ -> luai' (to avoid confusion with other luac stuff) 2005-03-09 13:28:07 -03:00
Roberto Ierusalimschy
f8df21bd20 `luaconf.h´ exports all its definitions always (so all of them
must have a lua/LUA prefix).
2005-03-08 17:10:05 -03:00
Roberto Ierusalimschy
d55bb795fa details 2005-02-23 14:30:22 -03:00
Roberto Ierusalimschy
7d45a5f48f C functions and userdata also have environments 2005-02-18 10:40:02 -02:00
Roberto Ierusalimschy
73d7640244 more secure way to stop GC cycles during finalizers 2005-02-14 11:19:50 -02:00
Roberto Ierusalimschy
678d2fb2ac all collected userdata must go to the end of the list of finalizers 2005-02-11 18:03:35 -02:00
Roberto Ierusalimschy
4df8800a01 cleaner way to free all objects 2005-02-10 11:25:02 -02:00
Roberto Ierusalimschy
334ba8132b cleaner way to remark open upvalues 2005-01-18 15:18:09 -02:00
Roberto Ierusalimschy
5be517602e no more generational collector (and no more `noinc' mode) 2005-01-14 12:19:42 -02:00
Roberto Ierusalimschy
e2498e079e change in hash algorithm so that it does not need empty slot
(tables can be 100% full)
2005-01-05 16:20:51 -02:00
Roberto Ierusalimschy
c6254dceff a different option for the GC 2004-12-13 10:15:11 -02:00
Roberto Ierusalimschy
39a8082f50 more options for controling the GC 2004-12-06 15:53:42 -02:00
Roberto Ierusalimschy
0e002005b1 better names for `luaM_free...' macros 2004-11-24 17:20:21 -02:00
Roberto Ierusalimschy
2f82bf6fe9 better support for 64-bit machines 2004-11-24 16:55:56 -02:00
Roberto Ierusalimschy
6f1ea817f5 better control over memory-size overflows 2004-11-19 13:52:40 -02:00
Roberto Ierusalimschy
8050e75f9d bug: gc metamethod must disable GC steps (to avoid nested calls) 2004-10-08 13:00:34 -03:00
Roberto Ierusalimschy
bd38017ddf small optimization for table size in machines with double allignment 2004-10-06 15:34:16 -03:00
Roberto Ierusalimschy
0e54d2be36 bug: barrier was wrong for generational phase 2004-09-15 17:38:15 -03:00
Roberto Ierusalimschy
0de2065f4e tighter tests for stack overflow 2004-09-08 11:23:09 -03:00
Roberto Ierusalimschy
0b06241483 better control for GC cycles 2004-08-30 10:44:44 -03:00
Roberto Ierusalimschy
32d4f304db first implementation of generational GC 2004-08-24 17:12:06 -03:00
Roberto Ierusalimschy
26ae992129 less conservative write barrier for tables 2004-08-10 16:17:23 -03:00
Roberto Ierusalimschy
b4cd38ba6c new scheme for configuration through `luaconf.h' 2004-04-30 17:13:38 -03:00
Roberto Ierusalimschy
178246062c reuse `sweeplist' for all lists 2004-03-23 09:57:12 -03:00
Roberto Ierusalimschy
a4e1230f95 better way to control open upvalues 2004-03-15 18:04:33 -03:00
Roberto Ierusalimschy
b876ec61c0 new (temporary?) API for garbage collector 2004-03-09 14:34:35 -03:00
Roberto Ierusalimschy
2aaf7394ad more and better tools (assertions & inspectors) to check incremental GC 2004-02-16 16:09:52 -03:00
Roberto Ierusalimschy
beb2aa5a46 atomic operations are not GC "states" 2003-12-12 16:29:34 -02:00
Roberto Ierusalimschy
47fc57a252 TObject' renamed to TValue' + other name changes and better assertions
for incremental garbage collection
2003-12-10 10:13:36 -02:00
Roberto Ierusalimschy
df429f163a First version of incremental GC 2003-12-09 14:56:11 -02:00
Roberto Ierusalimschy
fe595a45c2 `grayagain' list 2003-12-04 16:52:23 -02:00
Roberto Ierusalimschy
9db1942bac sweep of strings also incremental 2003-12-04 15:22:42 -02:00
Roberto Ierusalimschy
c6eac44a94 two different white flags (to distinguish dead elements from new ones) 2003-12-03 18:03:07 -02:00
Roberto Ierusalimschy
8878554b85 single list for all collectible objects, with udata separated at the
end of the list
2003-12-03 10:30:41 -02:00
Roberto Ierusalimschy
af850484a9 default metatable can be NULL 2003-12-01 16:22:56 -02:00
Roberto Ierusalimschy
1d10acb355 incremental GC phases 2003-12-01 14:33:30 -02:00
Roberto Ierusalimschy
57b6ed6815 initial implementation of white/gray/black coloring 2003-11-19 17:41:57 -02:00
Roberto Ierusalimschy
366e4af3c9 towards incremental GC 2003-11-18 12:55:11 -02:00
Roberto Ierusalimschy
ab7d9bfd0c cleaner code for manipulation of `marked' field 2003-11-17 17:50:05 -02:00
Roberto Ierusalimschy
8332d5c8a5 parser fully reentrant(!) 2003-08-27 18:01:44 -03:00
Roberto Ierusalimschy
a0a384a6b1 bug: userdata to be collected still counts into new GC threshold,
increasing memory consumption
2003-07-29 16:25:37 -03:00
Roberto Ierusalimschy
fa26d294ae new way to control `pc' of running functions 2003-07-16 17:49:02 -03:00
Roberto Ierusalimschy
9e05503ffa bug: C functions also may have stacks larger than current top 2003-07-07 10:32:19 -03:00
Roberto Ierusalimschy
ecf5730c0c (much) smarter way to clear weak tables 2003-05-16 15:58:39 -03:00
Roberto Ierusalimschy
572a69df78 Lua does not need all those different types... 2003-04-28 16:26:16 -03:00
Roberto Ierusalimschy
0d88545b82 warnings from several compilers (mainly typecasts when lua_Number is float) 2003-04-03 10:35:34 -03:00
Roberto Ierusalimschy
ab9e062928 name changes to avoid name collision between macros and variables 2003-03-18 09:50:04 -03:00
Roberto Ierusalimschy
7a40cdbda0 different variables for number of upvalues and size of upvalue array
(makes code clearer)
2003-02-11 08:46:24 -02:00
Roberto Ierusalimschy
6f207b15fb resist errors in finalizers during lua_close 2003-02-10 15:32:50 -02:00
Roberto Ierusalimschy
dd8edecae1 new functions to manipulate upvales (get/setupvalue) 2002-12-19 09:11:55 -02:00
Roberto Ierusalimschy
27c6b4d422 each .c file defines its own name 2002-12-04 15:38:31 -02:00
Roberto Ierusalimschy
c75c3cfd19 avoid names `str...' (may conflict with string.h) 2002-12-02 10:06:10 -02:00
Roberto Ierusalimschy
e47baca75a warnings from Visual C++ 2002-11-25 10:38:47 -02:00
Roberto Ierusalimschy
dbc5451bea comments 2002-11-22 16:01:46 -02:00
Roberto Ierusalimschy
5142e630bf new macro `condhardstacktests' to control hard stack tests 2002-11-21 15:19:11 -02:00
Roberto Ierusalimschy
2e4e888de0 Lua functions must have a frame or a saved pc. 2002-11-21 13:46:20 -02:00
Roberto Ierusalimschy
aff97fffc4 C functions cannot increase its top without filling the stack 2002-11-21 12:17:15 -02:00
Roberto Ierusalimschy
43013b39cc new representation for hooks (to allow asynchronous calls to sethook) 2002-11-18 09:01:55 -02:00
Roberto Ierusalimschy
5c5d9b2703 back to `__mode' metafield to specify weakness 2002-11-14 10:01:35 -02:00
Roberto Ierusalimschy
3010eb0536 all objects with several children (tables, closures, stacks, prototypes)
go to `gray' queue
2002-11-13 09:49:19 -02:00
Roberto Ierusalimschy
ac27b7a842 avoid traversing extra elements when cleaning weak tables 2002-11-11 09:52:43 -02:00
Roberto Ierusalimschy
dff9be4224 new macros to distinguish different types of object moves (for future GC
evolution).
2002-11-07 13:37:10 -02:00
Roberto Ierusalimschy
96e15b8501 threads now are real Lua objects, subject to garbage collection 2002-10-25 17:05:28 -03:00
Roberto Ierusalimschy
81bc5711a8 only one instance of registry and default metatable per global state 2002-10-22 14:58:14 -03:00
Roberto Ierusalimschy
b3d0682fb9 use of different buffers for scanner and concatenation 2002-10-08 15:46:08 -03:00