961760c9b6
message handler always adds a traceback to messages (even if they are not strings), unless they have a __tostring method.
Roberto Ierusalimschy
2014-10-20 16:19:26 -02:00
57559547a3
in 'pack'/'unpack', endianness and alignment treated like options + small changes in names and handling of internal options ('KOption')
Roberto Ierusalimschy
2014-10-20 14:44:54 -02:00
cfa84e1232
does not use 'long long' when LUA_ANSI is set
Roberto Ierusalimschy
2014-10-20 14:32:05 -02:00
c172a4f7c2
'dumpint' and related functions replaced by 'string.pack'/'string.unpack'
Roberto Ierusalimschy
2014-10-17 07:55:28 -03:00
f8e0d33b25
added "_CRT_SECURE_NO_WARNINGS" for Windows (to avoid warnings about several standard C functions) + small changes in '#include's
Roberto Ierusalimschy
2014-10-15 11:53:20 -03:00
34b6664dcb
better to use 'long' to represent UTF-8 code points
Roberto Ierusalimschy
2014-10-01 08:52:33 -03:00
d35fff16d5
'ipairs' always stops at first nil element
Roberto Ierusalimschy
2014-09-30 10:53:26 -03:00
56699cd603
small change to avoid bug in some versions of the clang compiler
Roberto Ierusalimschy
2014-09-26 11:09:21 -03:00
8ede2c353c
detail in 'report' + message handler always is called with an argument + 'report' already handles non-string error messages
Roberto Ierusalimschy
2014-09-25 11:20:37 -03:00
6384475ec4
'luaL_getmetafield' returns type of metafield (instead of a boolean)
Roberto Ierusalimschy
2014-09-22 03:42:15 -03:00
3a15c7ce43
size for array part of a table ('sizearray') changed from 'int' to 'unsigned int', which allows twice as many elements in the array part
Roberto Ierusalimschy
2014-09-04 15:15:29 -03:00
0a6b58c3aa
weak tables that must be retraversed are kept in 'grayagain' list until atomic phase (instead of going to their special lists) + more comments
Roberto Ierusalimschy
2014-09-03 13:54:41 -03:00
a1ab5ab396
'table.copy' -> 'table.move' + optional parameter moved to the end + several functions operate on "virtual" tables too
Roberto Ierusalimschy
2014-08-21 16:13:55 -03:00
57d8424146
extra space for new threads is initialized with a copy of the main thread, not of the creator thread.
Roberto Ierusalimschy
2014-07-24 13:17:56 -03:00
baa0e23456
better support for extra user space associated with a Lua state
Roberto Ierusalimschy
2014-07-24 11:00:16 -03:00
55a710545c
detail (extra byte in LUAI_EXTRASPACE to test proper alignment of a lua_State)
Roberto Ierusalimschy
2014-07-23 14:16:50 -03:00
753f2c1da4
better(?) alignment for some structures (pointers first)
Roberto Ierusalimschy
2014-07-23 14:15:43 -03:00
b3b8580606
comments + proper undef of compatibility macros + no need to undef internal macros (this is the first definition)
Roberto Ierusalimschy
2014-07-23 13:47:47 -03:00
73ad16b5c6
no need for type 'pCallInfo'
Roberto Ierusalimschy
2014-07-23 13:44:30 -03:00
97e299c395
'lua_replace' implemented as a macro using 'lua_copy'
Roberto Ierusalimschy
2014-07-22 15:07:47 -03:00
c697aa30bc
more precise type for argument to 'aux_upvalue'
Roberto Ierusalimschy
2014-07-21 13:02:57 -03:00
17ee57f8e0
'iswhite' and related macros now can work directly on any object (no need to convert to 'GCObject')
Roberto Ierusalimschy
2014-07-21 13:02:10 -03:00
3c6d0aaa7d
'iswhite' and related macros now can work directly on any object (no need to convert to 'GCObject')
Roberto Ierusalimschy
2014-07-19 12:14:46 -03:00
e43612aaf6
put the restriction that 'luaC_barrierback' works only on tables in its prototype
Roberto Ierusalimschy
2014-07-19 12:09:37 -03:00
2898e2fd12
removed useless assertion (gcstate != GCSpause already implied by other assertions) and wrong assertion (setmetatable uses this barrier for tables too)
Roberto Ierusalimschy
2014-07-19 11:44:19 -03:00
8704fca0c9
in 'luaL_checkversion_' check numeric types first. (Other tests depend on correct numeric type.)
Roberto Ierusalimschy
2014-07-19 11:37:09 -03:00
325e44d87e
no more "-2" for limits. (Limits should be precise; safety is in the code that handles these limits.)
Roberto Ierusalimschy
2014-07-18 15:29:12 -03:00
54ac253343
detail (no need to define LUAI_USER_ALIGNMENT_T when it is not defined; simpler to define 'L_Umaxalign' directly)
Roberto Ierusalimschy
2014-07-18 15:18:45 -03:00
bb12903120
type 'Udata' refers directly to structure inside the union (union used only for aligning purposes now)
Roberto Ierusalimschy
2014-07-18 11:46:47 -03:00
ca41b43f53
type 'TString' refers directly to the structure inside the union (union used only for size purposes)
Roberto Ierusalimschy
2014-07-18 10:36:14 -03:00
3511e186cd
ensure size for LUAI_USER_ALIGNMENT_T in tests is larger than the structures it is aligning (in most architectures)
Roberto Ierusalimschy
2014-07-18 10:27:45 -03:00
56137d58ff
added check for conversion 'obj2gco' (and corrections for small problems detected by this check)
Roberto Ierusalimschy
2014-07-18 09:17:54 -03:00
9aec500a26
no need for field 'gch' anymore
Roberto Ierusalimschy
2014-07-17 14:27:49 -03:00
5a9cc57a5e
change in GCObject: instead of being a union, it is now a structure with the common header of all collectable objects; union is used only for conversions. (Goal is to be able to check that the cast 'obj2gco' can have a check to ensure that object being converted is really a collectable object.). This is the first step in the change.
Roberto Ierusalimschy
2014-07-17 14:09:50 -03:00
1aa4f69b51
new type 'lua_Ctx' for continuation-function contexts (to allow type to be configurable)
Roberto Ierusalimschy
2014-07-17 10:53:37 -03:00
8bb272a3e3
new conversion float->integer: conversion is valid only when float has an exact representation as an integer
Roberto Ierusalimschy
2014-07-17 09:30:53 -03:00
c229ed597f
'requiref' checks 'package.loaded' before loading a module
Roberto Ierusalimschy
2014-07-16 11:51:36 -03:00
16b4110521
function 'type' keeps type names as upvalues to avoid creating strings everytime it is called
Roberto Ierusalimschy
2014-07-16 10:56:59 -03:00
f14662fca6
detail (added placeholders for non-function fields to preallocate space for them)
Roberto Ierusalimschy
2014-07-16 10:56:14 -03:00
10200585a7
var-arguments to the script come from 'arg' table (not from original 'argv' array)
Roberto Ierusalimschy
2014-06-26 14:08:52 -03:00
c94f11d783
'IntPoint' -> 'point2int' + ensure that casted value fits in destination type
Roberto Ierusalimschy
2014-06-26 13:17:35 -03:00
54ac5960a8
LUA_MAXUNSIGNED was only used to define 2.0^intnumbits (and that definition was not strictly portable), so it was removed and replaced by a more correct and direct definition for 2.0^intnumbits (only where it was needed)
Roberto Ierusalimschy
2014-06-24 14:02:00 -03:00
0e6229a953
use proper macros to convert 'GCObject' to other objects + better type cheking in 'set*value' macros
Roberto Ierusalimschy
2014-06-19 15:39:36 -03:00
d2ab1aef31
detail (avoid too large macro)
Roberto Ierusalimschy
2014-06-19 15:29:30 -03:00
89b56e7d84
more precision between closure types ('LClosure' x 'CClosure')
Roberto Ierusalimschy
2014-06-19 15:27:20 -03:00
14929f5764
use appropriate macros to convert GCObject to specific types
Roberto Ierusalimschy
2014-06-18 19:59:29 -03:00
7021cc9bc8
allows different 'source' for each prototype, but inherits it from parent when they are equal (only possible case for chunks created by the parser)
Roberto Ierusalimschy
2014-06-18 15:35:43 -03:00
04da0a40c0
added assertion about sources being the same for all protos in a chunk
Roberto Ierusalimschy
2014-06-18 10:54:31 -03:00
Roberto Ierusalimschy