lua, branch v5.4.9A mirror of https://github.com/lua/lua
https://git.lua4.win/lua/atom?h=v5.4.92026-08-07T18:59:21+00:00Small corrections in the manual2026-08-07T18:59:21+00:00Roberto Iroberto@inf.puc-rio.br2026-08-07T18:59:21+00:00urn:sha1:312b9efaa1061c2c4cad08554dbc1351c3270eefNew copyright year (2026)2026-08-05T18:04:44+00:00Roberto Iroberto@inf.puc-rio.br2026-08-05T18:04:44+00:00urn:sha1:03aabbc636eacf27a16694e5d86e4551ade8a954New release number (5.4.9)2026-07-24T14:39:24+00:00Roberto Iroberto@inf.puc-rio.br2026-07-24T14:39:24+00:00urn:sha1:2f8fe031c75d925e01aedc49507e247657d751fbBug: 'luaL_newmetatable' used in a wrong way2026-07-24T14:38:45+00:00Roberto Iroberto@inf.puc-rio.br2026-07-24T14:38:45+00:00urn:sha1:f4f4968fd3947519fdca385b6f9757e55ae60441
The call to 'luaL_newmetatable' in 'newbox' can leave an incomplete
metatable in the registry, if 'luaL_setfuncs' raises a memory error.
Bug: wrong initialization in result from 'gmatch'2026-07-24T14:21:24+00:00Roberto Iroberto@inf.puc-rio.br2026-07-24T14:21:24+00:00urn:sha1:e282da498cbf0bd5ea474b38bd61f6a1453d035d
Function returned by 'string.gmatch' can be left in an inconsistent
state after an error.
Bug: Loading a binary chunk does not run the GC2026-07-24T14:16:38+00:00Roberto Iroberto@inf.puc-rio.br2026-07-24T14:16:38+00:00urn:sha1:b18086ea3422ac89d1f7709dd7a37589040e233aBug: shift overflow in utf-8 decode2026-07-22T18:07:52+00:00Roberto Iroberto@inf.puc-rio.br2026-07-22T18:07:52+00:00urn:sha1:af19891bce064a625deeea0fa3d0005f6e22842a
An initial byte \xFF will ask for 7 continuation bytes, and then the
shift by (count * 5) will try to shift 35 bits.
Bug: GC checks stack space before running finalizer2026-07-22T17:42:55+00:00Roberto Iroberto@inf.puc-rio.br2026-07-22T17:42:55+00:00urn:sha1:8511e90b7df5daf139aba290b9c9c1595927e7b3
If some stack does not have a minimum available space, the GC defers
calling a finalizer until the next cycle. That avoids errors while
running a finalizer that the programmer cannot control.
Bug: Issues with write barrier for __newindex2026-07-13T18:55:36+00:00Roberto Iroberto@inf.puc-rio.br2026-07-13T18:55:36+00:00urn:sha1:0f781f836ab348716eb9232d3831a871b9821a3c
In 'luaV_finishset', there is an update on a table that is a field on
another table. If the first table is the same as the one with the field
(e.g., after 't.__newindex = t'), the update can change the value on
that field (if the field being updated is '__newindex' itself). After
that, the barrier is called with the table stored in that field, which
is not the correct table anymore.
Bug: Constructors with nils can overflow counters2025-08-27T17:58:02+00:00Roberto Ierusalimschyroberto@inf.puc-rio.br2025-08-27T17:58:02+00:00urn:sha1:934fdd481ced3a9d4a7aaace4479ce889ab23582