aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* put the restriction that 'luaC_barrierback' works only on tablesRoberto Ierusalimschy2014-07-193-29/+33
| | | | in its prototype
* removed useless assertion (gcstate != GCSpause already implied byRoberto Ierusalimschy2014-07-191-4/+4
| | | | | other assertions) and wrong assertion (setmetatable uses this barrier for tables too)
* in 'luaL_checkversion_' check numeric types first. (Other testsRoberto Ierusalimschy2014-07-191-4/+3
| | | | depend on correct numeric type.)
* no more "-2" for limits. (Limits should be precise; safety is inRoberto Ierusalimschy2014-07-181-6/+6
| | | | the code that handles these limits.)
* detail (no need to define LUAI_USER_ALIGNMENT_T when it is notRoberto Ierusalimschy2014-07-181-4/+5
| | | | defined; simpler to define 'L_Umaxalign' directly)
* type 'Udata' refers directly to structure inside the union (unionRoberto Ierusalimschy2014-07-187-35/+48
| | | | used only for aligning purposes now)
* type 'TString' refers directly to the structure inside the unionRoberto Ierusalimschy2014-07-1813-104/+102
| | | | (union used only for size purposes)
* ensure size for LUAI_USER_ALIGNMENT_T in tests is larger than theRoberto Ierusalimschy2014-07-181-2/+2
| | | | structures it is aligning (in most architectures)
* added check for conversion 'obj2gco' (and corrections for smallRoberto Ierusalimschy2014-07-1810-39/+51
| | | | problems detected by this check)
* no need for field 'gch' anymoreRoberto Ierusalimschy2014-07-175-81/+71
|
* change in GCObject: instead of being a union, it is now a structureRoberto Ierusalimschy2014-07-172-19/+30
| | | | | | | | 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.
* new type 'lua_Ctx' for continuation-function contexts (to allow typeRoberto Ierusalimschy2014-07-177-42/+63
| | | | to be configurable)
* new conversion float->integer: conversion is valid only whenRoberto Ierusalimschy2014-07-174-35/+46
| | | | float has an exact representation as an integer
* 'requiref' checks 'package.loaded' before loading a moduleRoberto Ierusalimschy2014-07-162-13/+22
|
* function 'type' keeps type names as upvalues to avoid creating stringsRoberto Ierusalimschy2014-07-161-8/+21
| | | | everytime it is called
* detail (added placeholders for non-function fields to preallocateRoberto Ierusalimschy2014-07-163-3/+16
| | | | space for them)
* Table library now respects '__index'/'__newindex' metamethodsRoberto Ierusalimschy2014-07-161-43/+96
|
* removed unused parameter Ä'L' in macro 'api_check' and companyRoberto Ierusalimschy2014-07-156-55/+52
|
* added api check in 'lua_typename' (tag is valid)Roberto Ierusalimschy2014-07-151-2/+3
|
* detail (typos in comments)Roberto Ierusalimschy2014-06-309-28/+28
|
* unsigned-manipulation functions (lua_puhsunsigned, lua_tounsigned, etc.)Roberto Ierusalimschy2014-06-266-32/+46
| | | | deprecated
* emergency collection can happen even when collector is stoppedRoberto Ierusalimschy2014-06-261-5/+3
|
* detail (comment)Roberto Ierusalimschy2014-06-261-2/+2
|
* unsigned-manipulation functions (lua_puhsunsigned, lua_tounsigned, etc.)Roberto Ierusalimschy2014-06-261-48/+1
| | | | deprecated
* var-arguments to the script come from 'arg' table (not from originalRoberto Ierusalimschy2014-06-261-10/+16
| | | | 'argv' array)
* 'IntPoint' -> 'point2int' + ensure that casted value fits inRoberto Ierusalimschy2014-06-262-5/+5
| | | | destination type
* LUA_MAXUNSIGNED was only used to define 2.0^intnumbits (and thatRoberto Ierusalimschy2014-06-242-9/+8
| | | | | | 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)
* use proper macros to convert 'GCObject' to other objects + betterRoberto Ierusalimschy2014-06-191-24/+23
| | | | type cheking in 'set*value' macros
* detail (avoid too large macro)Roberto Ierusalimschy2014-06-191-3/+5
|
* more precision between closure types ('LClosure' x 'CClosure')Roberto Ierusalimschy2014-06-199-49/+49
|
* use appropriate macros to convert GCObject to specific typesRoberto Ierusalimschy2014-06-183-9/+17
|
* allows different 'source' for each prototype, but inherits it fromRoberto Ierusalimschy2014-06-182-21/+34
| | | | | parent when they are equal (only possible case for chunks created by the parser)
* added assertion about sources being the same for all protos in a chunkRoberto Ierusalimschy2014-06-181-2/+4
|
* changed macro for compatibility options + detailsRoberto Ierusalimschy2014-06-181-5/+5
|
* source for all prototypes must be equal in a chunk; no need to storeRoberto Ierusalimschy2014-06-182-4/+6
| | | | each one separated
* 'math.ifloor' is backRoberto Ierusalimschy2014-06-181-1/+15
|
* macro 'checkobjref' accepts NULL (as all its uses checked for NULLRoberto Ierusalimschy2014-06-171-21/+15
| | | | before) + user value from a userdata may not be a GC object
* cleaner way to handle bit CIST_OAH (with auxiliar macros)Roberto Ierusalimschy2014-06-123-16/+16
|
* janitor work on 'lua_resume' and related codeRoberto Ierusalimschy2014-06-111-28/+32
|
* new items in default path for WindowsRoberto Ierusalimschy2014-06-101-6/+10
|
* bit-field CIST_YIELDED removed (it was never consulted)Roberto Ierusalimschy2014-06-102-10/+8
|
* must also reset CIST_OAH if necessary (CallInfo is not a fresh one)Roberto Ierusalimschy2014-06-101-5/+6
|
* more relaxed rules for __eq metamethod (more similar to otherRoberto Ierusalimschy2014-06-103-20/+11
| | | | operators)
* no need for field 'status' in structure 'CallInfo' (after removalRoberto Ierusalimschy2014-06-103-29/+29
| | | | | of 'lua_getctx') + field 'old_allowhook' can be packed into a single bit
* new type lua_KFunction + no more 'lua_getctx'Roberto Ierusalimschy2014-06-106-61/+44
|
* added commentsRoberto Ierusalimschy2014-06-091-6/+19
|
* lots of janitor work (including comments) + 'arg' table createdRoberto Ierusalimschy2014-06-051-153/+223
| | | | | before running any script + changes in the parameters and return of 'collectargs'
* 'math.mof' works with integers, tooRoberto Ierusalimschy2014-06-021-12/+24
|
* 'assert' does not assume that the error object is a stringRoberto Ierusalimschy2014-06-021-4/+9
|
* if numeral overflows the reading buffer, signal it as invalid inputRoberto Ierusalimschy2014-06-021-3/+5
| | | | (resulting in nil)
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-21 03:53:20 +0000