c510d94549
includes counts from 'sweeptolive' in cost of atomic step
Roberto Ierusalimschy
2012-05-31 18:28:59 -03:00
4dd0622d0c
bug in luaL_getsubtable (calling lua_absindex not with original stack)
Roberto Ierusalimschy
2012-05-31 17:28:45 -03:00
ab2c3d5cde
small bug in error handling of finalizers (cannot call lua_tostring inside the core) + small bug in luaC_checkfinalizer (when avoiding remove object being sweeped from sweep list) + small changes in GC pace control (for the atomic part)
Roberto Ierusalimschy
2012-05-31 17:26:14 -03:00
01e5f358bd
small improvement in lua_checkmemory (only allow dead objects in part of the sweep list not yet sweeped)
Roberto Ierusalimschy
2012-05-31 17:25:42 -03:00
69c775de21
update of 'lua_checkmemory'
Roberto Ierusalimschy
2012-05-30 13:40:29 -03:00
6fcdfc6d4f
bug: object being moved to 'finobj' list might not be sweeped by the collector
Roberto Ierusalimschy
2012-05-30 13:01:10 -03:00
b77a90681e
typos in comments
Roberto Ierusalimschy
2012-05-30 09:33:44 -03:00
6efffb9acd
only count in 'atomic' objects marked for the first time
Roberto Ierusalimschy
2012-05-29 14:52:17 -03:00
51e8f08e60
more efficient way to apply 'stepmul' + some changes in GC parameters
Roberto Ierusalimschy
2012-05-28 17:41:00 -03:00
5adb5a4812
definition for 'MAX_LMEM'
Roberto Ierusalimschy
2012-05-28 17:32:28 -03:00
e29f3a5751
definition of 'GCSTEPSIZE' moved to header file + small changes
Roberto Ierusalimschy
2012-05-23 12:43:14 -03:00
21ed264a38
compensate for 'GCSTEPSIZE' in GC steps
Roberto Ierusalimschy
2012-05-23 12:42:27 -03:00
8e7149f496
detail ('char' should be 'const char')
Roberto Ierusalimschy
2012-05-22 15:38:56 -03:00
086da10dac
merge of fields 'lastmajormem' (used in gen. mode) and 'estimate' (used in inc. mode)
Roberto Ierusalimschy
2012-05-22 14:50:39 -03:00
b36b4b521f
try to avoid sweeping new objects created with new white (and therefore not collectable in the current cycle)
Roberto Ierusalimschy
2012-05-22 14:32:25 -03:00
398811a313
simpler macro 'luaC_condGC' + better 'step' in 'lua_gc' + micro bug in 'luaC_checkfinalizer' (current sweep object could be removed from 'allgc' list)
Roberto Ierusalimschy
2012-05-21 10:18:10 -03:00
2a66b34f72
revamp of the GC pace control; more like 5.1: any X Kbytes allocated makes the GC handle f(X) Kbytes of objects
Roberto Ierusalimschy
2012-05-20 17:36:44 -03:00
8d0e1ed52f
extend optimization of 'if a then break end' for the case 'if a then break; end'
Roberto Ierusalimschy
2012-05-20 11:51:23 -03:00
3ac595da8a
allow "long double" precision for PI constant if needed
Roberto Ierusalimschy
2012-05-18 14:47:53 -03:00
3cadc37f47
no more 'Proto' objects on the stack. Protos are anchored on outer Protos or on a Closure, which must be created before the Proto.
Roberto Ierusalimschy
2012-05-08 10:53:33 -03:00
bb1146dc39
wrong error message when 'coroutine.create' or 'coroutine.wrap' are called with no arguments (new thread is assumed to be the missing argument).
Roberto Ierusalimschy
2012-04-27 15:59:04 -03:00
fc24a3a6c4
small bug: a reader function should not modify the Lua stack
Roberto Ierusalimschy
2012-04-27 11:13:19 -03:00
a94ce68e0a
bug: wrong handling of 'nCcalls' in coroutines
Roberto Ierusalimschy
2012-04-26 17:41:18 -03:00
ec22fc963a
detail (separate code to create 'searchers' table into a new function)
Roberto Ierusalimschy
2012-04-26 16:38:52 -03:00
c54f5f64c9
primaryexp -> suffixedexp; prefixexp -> primaryexp + more 'syntactical' way to distinguish between function calls and assignments
Roberto Ierusalimschy
2012-04-20 16:20:05 -03:00
4cca1a436d
details (using lua_setglobal/lua_getglobal instead of explicit use of the global table)
Roberto Ierusalimschy
2012-04-20 14:05:17 -03:00
048be9cf27
different scheme to check arguments to '%d', '%x', etc. Old scheme did not work well when lua_Number is an integer.
Roberto Ierusalimschy
2012-04-20 10:16:48 -03:00
1485ea2ee7
Finalizers may call functions from a dynamic library after the library has been unloaded
Roberto Ierusalimschy
2012-04-12 13:25:25 -03:00
b10dbe5c72
bug: C libraries must be unloaded after all other finalizers have run, because a finalizer may use a C function from a C library
Roberto Ierusalimschy
2012-04-11 13:35:32 -03:00
462375ba47
'if' to avoid empty 'memcpy' (may be expensive)
Roberto Ierusalimschy
2012-04-03 16:06:19 -03:00
cfbe2333a4
string hash may not use all bytes (but this is configurable now) + small other changes
Roberto Ierusalimschy
2012-03-28 15:28:25 -03:00
afc3fcf952
definition for 'LUA_MAXSHORTLEN' moved to 'lstring.c' (used only there)
Roberto Ierusalimschy
2012-03-28 15:27:25 -03:00
99bd1a000f
error function can be 'l_noret'
Roberto Ierusalimschy
2012-03-19 19:58:09 -03:00
d19f1da6ef
new macro 'isreserved' + 'eqstr' -> 'luaS_eqstr' (may be a function)
Roberto Ierusalimschy
2012-01-23 21:04:07 -02:00
f88806a1fb
"default: lua_assert(0)" in switches helps debugging + uses non-variant types in binary files
Roberto Ierusalimschy
2012-01-23 21:02:10 -02:00
fb6eee2805
Lua never uses Ã'assert' (it is always 'lua_assert')
Roberto Ierusalimschy
2012-01-23 20:59:02 -02:00
0010ec68b1
documentation comment (small correction about strings being gray)
Roberto Ierusalimschy
2012-01-23 18:29:12 -02:00
28f8ed9113
new macro 'checktype'
Roberto Ierusalimschy
2012-01-20 20:47:11 -02:00
fd22ccd6d0
object tag keeps variant bits too -> no need for 'isC' field in Closures + more strick typing for closure variants
Roberto Ierusalimschy
2012-01-20 20:05:50 -02:00
76eab106df
bug: Lexical gets confused with some combination of arithmetic operators and hexadecimal numbers
Roberto Ierusalimschy
2012-01-20 16:35:36 -02:00
81ed85ecfb
BUG: memory hoarding when creating Lua hooks for coroutines BUG: Lexical gets confused with some combination of arithmetic operators and hexadecimal numbers
Roberto Ierusalimschy
2012-01-20 16:32:13 -02:00
7dcc02b165
BUG: coroutine hooks were not collected together with coroutine
Roberto Ierusalimschy
2012-01-19 18:14:44 -02:00
e2fc2ce8df
new way to handle -E option (write a mark in the registry to avoid reading environment variables)
Roberto Ierusalimschy
2011-12-12 14:34:03 -02:00
51650eac78
small change to avoid warnings of non-specified order of access to volatile fields
Roberto Ierusalimschy
2011-11-29 13:39:48 -02:00
aea68e5c54
explicit option RTLD_LOCAL instead of 0 (Posix does not fix a default for dlopen)
Roberto Ierusalimschy
2011-11-28 15:27:51 -02:00
ece042fef0
'table.pack' does not return 'n' (may be confusing when using table.pack as last argument in a call)
Roberto Ierusalimschy
2011-11-28 15:26:30 -02:00
217b65e6d9
avoid some warnings about converting 32-bit shifts into 64-bit results
Roberto Ierusalimschy
2011-11-28 15:25:48 -02:00
ef900f224c
strutcture for file handles exported in 'lauxlib.h'
Roberto Ierusalimschy
2011-11-11 17:59:17 -02:00
3105febf9a
using Lua buffers to build error messages
Roberto Ierusalimschy
2011-11-10 09:42:58 -02:00
7133e20c94
make treatment of 'pcall' and 'xpcall' more similar
Roberto Ierusalimschy
2011-11-09 17:38:00 -02:00
888d39ea75
refactoring of common parts of 'pcall' and 'xpcall' + avoids stack errors on these functions (return error code instead)
Roberto Ierusalimschy
2011-11-09 17:28:27 -02:00
eef22fa4ce
create error messages with a single concat, to avoid problems with creating strings piecemeal
Roberto Ierusalimschy
2011-11-09 17:11:20 -02:00
eda87f1156
in 'luaL_checkstack', include extra stack space in test to allow correct execution of error routines, if necessary
Roberto Ierusalimschy
2011-11-09 17:08:55 -02:00