b6e9fcbd70
use of macro ('vmcase') to format all cases in main switch of 'luaV_execute' (to facilitate experiences with direct threading)
Roberto Ierusalimschy2010-05-12 17:40:35 -03:00
8554baeacf
small changes in 'luaV_execute' to make cases more regularly formatted
Roberto Ierusalimschy2010-05-12 17:31:33 -03:00
de0f51a02c
wrong type being assigned (not detected because of obj2gco macro)
Roberto Ierusalimschy2010-05-11 17:48:36 -03:00
0567221848
udata in 'tobefnz' list be have old bit on (it will be cleared when udata is moved to 'allgc' list)
Roberto Ierusalimschy2010-05-10 15:24:36 -03:00
1c1a98e872
corrected some places where an old object could end up in front of a new one + minimal documentation about this problem
Roberto Ierusalimschy2010-05-10 15:23:45 -03:00
a48e330f37
macros 'lua_number2int' and 'lua_number2uint' for Visual Studio corrected and tested
Roberto Ierusalimschy2010-05-10 13:38:58 -03:00
dcc070683c
avoid reserving LUA_RIDX_LAST slots in any table used by reference system. Store free list in a field indexed by a unique name, instead of using a numerical index.
Roberto Ierusalimschy2010-05-10 12:25:02 -03:00
2d81cfa4e1
corrected definition of 'lua_register' (there is no LUA_ENVIRONINDEX anymore)
Roberto Ierusalimschy2010-05-10 10:50:20 -03:00
0fac33da9e
bug: incremental sweep was not cleaning old bits (as it stopped in the first old object) + bug: moving udata to 'udgc' list might put old object in front a new one + some new macros + generational mode may be in 'pause' state (it just skips 'markroot')
Roberto Ierusalimschy2010-05-07 15:19:36 -03:00
9c1347e647
more tests in 'lua_checkmemory' + more information in function 'gccolor'
Roberto Ierusalimschy2010-05-07 15:09:23 -03:00
b373a40133
new macro 'isgenerational' + new macro 'isold' + better deffinition for 'isdead', compatible with the code used by 'sweeplist'
Roberto Ierusalimschy2010-05-07 15:08:05 -03:00
25951e0ea5
some cleaning in 'sweeplist' (threads do not need to be traversed if they are going to be collected + use of bit masks to unify differences in generational and incremental modes)
Roberto Ierusalimschy2010-05-06 15:18:07 -03:00
d2ea5b00b7
new function 'luaC_changemode' + bug: objects entering the 'allgc' list must have their OLDBIT cleared (upvalues being cleared) + bug: in 'checkSizes', KGC_EMERGENCY is stored in 'gckind' field, not in 'gcstate' + current white changes when entering sweep phase (so there are dead objects only in that phase)
Roberto Ierusalimschy2010-05-05 15:58:36 -03:00
3e662cec89
full GC must run finalizers only after finishing everything else (including a return to generational mode if needed)
Roberto Ierusalimschy2010-05-04 15:09:06 -03:00
658ea8752b
no need to handle '\0' differently from other control chars in format '%q'
Roberto Ierusalimschy2010-05-04 14:20:33 -03:00
aa73403ba4
no need of a KGC_FORCED collection kind; it has the same behavior of a normal collection
Roberto Ierusalimschy2010-05-03 14:39:48 -03:00
0c27de2e7b
no more 'finalize' phase in GC; finalizers are called along the entire cycle
Roberto Ierusalimschy2010-05-03 14:33:39 -03:00
d25f7f9d78
items in 'tobefnz' are kept black (as before recent change) and changed to white only when needed (being moved to 'allgc' when not keeping invariant).
Roberto Ierusalimschy2010-05-03 08:55:40 -03:00
0062db1e06
default PAUSE should be 200 (differences may be corrected in internal multiplers) + check for unbilt stack equal the one used by lgc.c
Roberto Ierusalimschy2010-04-30 11:22:23 -03:00
23001d8607
nasty GC bug: upvalue must be turned white when not keeping invariant, but barrier was not being called when uv->v were already white.
Roberto Ierusalimschy2010-04-29 18:43:36 -03:00
49ca1f7083
allows thread switches (when/if there are thread switches) in the same places that finalizers can run (so they should be safe...)
Roberto Ierusalimschy2010-04-19 14:58:46 -03:00
4db6f20770
ensure that 'luai_userstatethread' is always called (even if 'stack_init' throws a memory error)
Roberto Ierusalimschy2010-04-19 14:40:13 -03:00
8003e9544d
'string.format' may get buffer as an argument when there are missing arguments and format string is too long
Roberto Ierusalimschy2010-04-19 13:39:25 -03:00
1bb407e494
in lua_gc/step, 'luaC_step' changes GCthreshold, so there was little control over the real step size.
Roberto Ierusalimschy2010-04-19 13:33:19 -03:00
e7ffb1b501
macro LUA_USE_LONGJMP allows choosing longjmp when compiling Lua as C++ code
Roberto Ierusalimschy2010-04-18 09:41:35 -03:00
c066ac10a9
better line numbers for function calls and unary/binary operators + null statement + no more "ambiguous syntax" restriction
Roberto Ierusalimschy2010-04-17 09:46:32 -03:00
1e6940f291
'gcstate' now also runs collector until given state + small changes in 'testC' to test 'lua_topointer' and 'lua_tocfunction'
Roberto Ierusalimschy2010-04-16 14:42:49 -03:00
d8d81ba891
new escape sequence '\*' + several comments + moving options from switch default into cases (as now locale is fixed)
Roberto Ierusalimschy2010-04-16 09:31:07 -03:00
cf22133b69
no need to avoid calling ctype functions as now they are implemented by us (no inefficiencies due to accessing locale information)
Roberto Ierusalimschy2010-04-15 16:44:43 -03:00
9100f7479a
new implementation for Generic Buffer manipulation (using userdata as temporary buffer space)
Roberto Ierusalimschy2010-04-09 13:14:46 -03:00
055104f5b6
keep memory-error message in the global state, so that its use does not depend on Lua internalizing strings to avoid a string creation on memory errors
Roberto Ierusalimschy2010-04-08 14:16:46 -03:00
Roberto Ierusalimschy