Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | details (comments) | Roberto Ierusalimschy | 2014-10-03 | 1 | -4/+4 |
| | |||||
* | size for array part of a table ('sizearray') changed from 'int' to | Roberto Ierusalimschy | 2014-09-04 | 1 | -4/+4 |
| | | | | 'unsigned int', which allows twice as many elements in the array part | ||||
* | weak tables that must be retraversed are kept in 'grayagain' list | Roberto Ierusalimschy | 2014-09-03 | 1 | -42/+38 |
| | | | | | until atomic phase (instead of going to their special lists) + more comments | ||||
* | 'linktable' -> 'linkgclist' (and used for all links with 'gclist') | Roberto Ierusalimschy | 2014-09-01 | 1 | -22/+16 |
| | |||||
* | simpler definition for 'setobj' (trust the compiler for the assignment) | Roberto Ierusalimschy | 2014-07-29 | 1 | -2/+2 |
| | |||||
* | 'iswhite' and related macros now can work directly on any object | Roberto Ierusalimschy | 2014-07-19 | 1 | -8/+8 |
| | | | | (no need to convert to 'GCObject') | ||||
* | put the restriction that 'luaC_barrierback' works only on tables | Roberto Ierusalimschy | 2014-07-19 | 1 | -9/+7 |
| | | | | in its prototype | ||||
* | removed useless assertion (gcstate != GCSpause already implied by | Roberto Ierusalimschy | 2014-07-19 | 1 | -4/+4 |
| | | | | | other assertions) and wrong assertion (setmetatable uses this barrier for tables too) | ||||
* | type 'Udata' refers directly to structure inside the union (union | Roberto Ierusalimschy | 2014-07-18 | 1 | -2/+2 |
| | | | | used only for aligning purposes now) | ||||
* | type 'TString' refers directly to the structure inside the union | Roberto Ierusalimschy | 2014-07-18 | 1 | -8/+5 |
| | | | | (union used only for size purposes) | ||||
* | added check for conversion 'obj2gco' (and corrections for small | Roberto Ierusalimschy | 2014-07-18 | 1 | -5/+8 |
| | | | | problems detected by this check) | ||||
* | no need for field 'gch' anymore | Roberto Ierusalimschy | 2014-07-17 | 1 | -32/+32 |
| | |||||
* | detail (typos in comments) | Roberto Ierusalimschy | 2014-06-30 | 1 | -2/+2 |
| | |||||
* | simpler handling of 'GCScallfin' state + more comments | Roberto Ierusalimschy | 2014-05-25 | 1 | -57/+47 |
| | |||||
* | 'GCmemtrav' does not need to track the entire collection, only each | Roberto Ierusalimschy | 2014-04-04 | 1 | -8/+8 |
| | | | | single step individually | ||||
* | LUAI_FUNC is being used only in header files | Roberto Ierusalimschy | 2014-04-02 | 1 | -2/+2 |
| | |||||
* | more precise estimation (GCestimate) for total bytes in use after | Roberto Ierusalimschy | 2014-04-01 | 1 | -14/+18 |
| | | | | a GC cycle | ||||
* | details (typos in comments) | Roberto Ierusalimschy | 2014-03-21 | 1 | -2/+2 |
| | |||||
* | userdata can have any Lua value as uservalue | Roberto Ierusalimschy | 2014-02-19 | 1 | -4/+7 |
| | |||||
* | no need to keep threads in a different GC list, now that there is the | Roberto Ierusalimschy | 2014-02-18 | 1 | -5/+1 |
| | | | | 'twups' list | ||||
* | new list 'twups' to allow traversal of upvalues from dead threads | Roberto Ierusalimschy | 2014-02-18 | 1 | -10/+35 |
| | | | | (+ fixed some problems with cycles involving those upvalues) | ||||
* | field 'op' renamed to 'open' | Roberto Ierusalimschy | 2014-02-15 | 1 | -5/+5 |
| | |||||
* | change in the way 'collectgarbage("step", size)' interprets 'size' | Roberto Ierusalimschy | 2014-02-14 | 1 | -15/+7 |
| | | | | | (mimicking the way the GC itself behaves when Lua allocates 'size' Kbytes) | ||||
* | limit to 'gcstepmul' imposed by 'lua_gc' (+ some details in 'lgc.c') | Roberto Ierusalimschy | 2014-02-13 | 1 | -11/+18 |
| | |||||
* | better control for number of finalizers called at each GC cycle | Roberto Ierusalimschy | 2014-02-13 | 1 | -22/+40 |
| | | | | (increases progressively) | ||||
* | no more local collection | Roberto Ierusalimschy | 2014-02-13 | 1 | -157/+27 |
| | |||||
* | detail (better presentation for 'luaC_step') | Roberto Ierusalimschy | 2014-02-11 | 1 | -12/+11 |
| | |||||
* | keep a single list of objects to be finalized (with local and non-local | Roberto Ierusalimschy | 2014-02-11 | 1 | -31/+16 |
| | | | | objects), to ensure finalization order | ||||
* | details | Roberto Ierusalimschy | 2013-12-13 | 1 | -17/+13 |
| | |||||
* | use goto to implement a tail call in 'reallymarkobject' | Roberto Ierusalimschy | 2013-12-13 | 1 | -4/+9 |
| | |||||
* | CallInfo lists shrinks together with their associated stacks | Roberto Ierusalimschy | 2013-09-17 | 1 | -2/+1 |
| | |||||
* | GC local pause configurable | Roberto Ierusalimschy | 2013-09-13 | 1 | -27/+19 |
| | |||||
* | check for shrinking string table done only at the end of a GC cycle | Roberto Ierusalimschy | 2013-09-11 | 1 | -5/+10 |
| | |||||
* | new names and better order for GC states (sweep first lists that | Roberto Ierusalimschy | 2013-09-11 | 1 | -15/+15 |
| | | | | can have dead objects) | ||||
* | threads are kept in a separated GC list, linked after the main thread | Roberto Ierusalimschy | 2013-09-11 | 1 | -9/+12 |
| | |||||
* | detail: 'sweepstep' checks end of phase after calling 'sweeplist', so | Roberto Ierusalimschy | 2013-09-11 | 1 | -6/+7 |
| | | | | that phases with small lists return 0 at the first call to 'sweepstep' | ||||
* | objects in list 'tobefnz' have a GC life-cycle like all others | Roberto Ierusalimschy | 2013-09-11 | 1 | -9/+12 |
| | | | | (specifically they are cleaned during sweep phase) | ||||
* | 'luaC_newobj' does not handle special cases; only special case | Roberto Ierusalimschy | 2013-09-11 | 1 | -13/+6 |
| | | | | now is threads, which do not use 'luaC_newobj' anymore. | ||||
* | local collection now calls finalizers | Roberto Ierusalimschy | 2013-09-03 | 1 | -33/+87 |
| | |||||
* | new GC state to sweep 'localgc' list + small changes in sweep control | Roberto Ierusalimschy | 2013-08-30 | 1 | -34/+34 |
| | |||||
* | bug: local collection must clear rest of stack | Roberto Ierusalimschy | 2013-08-29 | 1 | -2/+5 |
| | |||||
* | tables and userdata all go to local list, too | Roberto Ierusalimschy | 2013-08-28 | 1 | -2/+4 |
| | |||||
* | LOCALBLACK changed to LOCALMARK and used also to control whether object | Roberto Ierusalimschy | 2013-08-27 | 1 | -6/+10 |
| | | | | | is in 'localgc' list + luaC_newobj by default puts object in 'localgc' list | ||||
* | upvalues collected by reference count | Roberto Ierusalimschy | 2013-08-27 | 1 | -84/+58 |
| | |||||
* | Lua closures go to local, too | Roberto Ierusalimschy | 2013-08-26 | 1 | -6/+35 |
| | |||||
* | C functions and strings now go to the local list; first version | Roberto Ierusalimschy | 2013-08-23 | 1 | -8/+94 |
| | | | | of the local collector | ||||
* | "fixed" objects kept in a separated list (instead of being kept in | Roberto Ierusalimschy | 2013-08-21 | 1 | -1/+12 |
| | | | | 'allgc' list with a bit marking them) | ||||
* | change in string table: string table is now independent of GC lists; all | Roberto Ierusalimschy | 2013-08-21 | 1 | -32/+14 |
| | | | | strings live in 'normal' GC lists | ||||
* | GC bits SEPARATED and FINALIZEDBIT mixed in FINALIZEDBIT (with simpler | Roberto Ierusalimschy | 2013-08-20 | 1 | -12/+7 |
| | | | | control) | ||||
* | "barrier" for link prototype->cache changed to be consistent with | Roberto Ierusalimschy | 2013-08-19 | 1 | -23/+1 |
| | | | | GC behavior (link is cleared to preserve invariant) |