Commit Graph

  • 2275030b86 details (and missing 'break's) Roberto Ierusalimschy 2011-09-13 15:05:59 -03:00
  • 05de314701 upvalue names always can be NULL (if debug info was removed), so always check for that case Roberto Ierusalimschy 2011-09-13 14:40:20 -03:00
  • d281d23f8d detail (make 'lua_assert' a valid expression even when turned off) Roberto Ierusalimschy 2011-09-13 14:39:23 -03:00
  • ec431ce5db comment Roberto Ierusalimschy 2011-09-12 17:33:03 -03:00
  • c8f47c2934 optimization for |if cond then goto| Roberto Ierusalimschy 2011-08-30 13:38:58 -03:00
  • ca458f1f52 'luaK_goiffalse' exported Roberto Ierusalimschy 2011-08-30 13:26:41 -03:00
  • 121dc8af66 cleaner code for 'if' construct Roberto Ierusalimschy 2011-08-25 10:45:24 -03:00
  • a8675966ec detail: centralize all tests for Windows on variable 'LUA_WIN' Roberto Ierusalimschy 2011-08-23 14:25:42 -03:00
  • 3dc5475e23 'nCcalls' should be local to each thread, as each thread may have its own C stack (with LuaThreads or something similar) Roberto Ierusalimschy 2011-08-23 14:24:34 -03:00
  • 8a008a2057 bug: __newindex metamethod may not work if metatable is its own metatable. Roberto Ierusalimschy 2011-08-17 17:38:51 -03:00
  • 89b59eee73 bug: __newindex metamethod may not work if metatable is its own metatable + luaV_settable does not create entry when there is a metamethod (and therefore entry is useless) Roberto Ierusalimschy 2011-08-17 17:26:47 -03:00
  • 166dd0261a new option '-E' to avoid environment variables Roberto Ierusalimschy 2011-08-17 17:19:52 -03:00
  • 51471ba748 comments about use of 'luaH_set' Roberto Ierusalimschy 2011-08-15 16:41:58 -03:00
  • 0b65cb74cd new implementation for 'getobjname': first search for relevant instruction (new function 'findsetreg') and then try to build a meaningful name Roberto Ierusalimschy 2011-08-12 17:01:44 -03:00
  • ea46b750dc comment Roberto Ierusalimschy 2011-08-11 14:06:12 -03:00
  • 92afcf2823 no more 'luaH_setstr (used only once) + 'luaH_setint' receives value to be set. Roberto Ierusalimschy 2011-08-09 17:58:29 -03:00
  • b5bf7d9ef4 'string.format' checks whether values for integer formats are actually integers Roberto Ierusalimschy 2011-08-09 17:58:29 -03:00
  • 6a5d89b39f detail (using array instead of several variables to keep track of enabled command-line options) Roberto Ierusalimschy 2011-08-04 15:16:16 -03:00
  • ae14adc272 better error message when trying to open files Roberto Ierusalimschy 2011-08-02 15:00:01 -03:00
  • ff98f17d27 detail: factoring in common code for opening files and checking for errors Roberto Ierusalimschy 2011-07-28 15:41:15 -03:00
  • 59bcd137ae reducing even more use of C stack by the parser: struct 'FuncState' does not need field 'L' + number of labels/gotos in a chunk may be limited to SHRT_MAX. (Also removed some non-needed 'unsigned's.) Roberto Ierusalimschy 2011-07-27 15:09:01 -03:00
  • 5ab6d36d99 no more field 'L' in 'FuncState' Roberto Ierusalimschy 2011-07-27 15:07:37 -03:00
  • 43c873895f tonumber: base 10 is not special, no base is Roberto Ierusalimschy 2011-07-27 09:14:06 -03:00
  • 03a078493e refuse things like 'inf' or 'Nan' as numerals Roberto Ierusalimschy 2011-07-27 09:13:08 -03:00
  • bae57ea088 'M_IX86' -> '_M_IX86' Roberto Ierusalimschy 2011-07-25 14:20:47 -03:00
  • 2557107e91 removed unused variable 'found' (in 'findfield') + removed legacy test for 'what' being "tail" Roberto Ierusalimschy 2011-07-25 14:18:49 -03:00
  • 32a12e2f3f detail (cleaned whitespaces at end of lines) Roberto Ierusalimschy 2011-07-15 09:50:29 -03:00
  • 559bb554c9 no more 'zungetc' (better not to read next char) Roberto Ierusalimschy 2011-07-15 09:48:03 -03:00
  • fd5e810e08 macro 'char2int' replaced by 'cast_uchar' (used by other files) Roberto Ierusalimschy 2011-07-15 09:35:32 -03:00
  • 067f761739 better code for numerical escapes Roberto Ierusalimschy 2011-07-15 09:30:41 -03:00
  • 7978a8d8b2 added '_M_IX86' in NaN trick v5.2-beta Roberto Ierusalimschy 2011-07-08 17:07:11 -03:00
  • 32e09c2a16 detail Roberto Ierusalimschy 2011-07-08 17:01:38 -03:00
  • df19931ddc error on invalid escape sequences Roberto Ierusalimschy 2011-07-08 16:17:30 -03:00
  • b5084fdafe detail (comment) Roberto Ierusalimschy 2011-07-06 13:45:14 -03:00
  • d8200d5f3c '\*' -> '\z' + '\?' is not an official escape sequence Roberto Ierusalimschy 2011-07-05 16:13:45 -03:00
  • 79cbc3468c removed and deprecated functions really removed from the code base Roberto Ierusalimschy 2011-07-05 09:49:35 -03:00
  • 9e89a55a4f NILCONSTANT also depends no endianess (with NaN trick) Roberto Ierusalimschy 2011-07-04 17:29:02 -03:00
  • b707e81c3d comments about some compiler options that should be tested once in a while Roberto Ierusalimschy 2011-07-04 13:10:15 -03:00
  • ee37ee50d6 'table.pack' also returns 'n' + 'deprecated' changed to 'removed' Roberto Ierusalimschy 2011-07-02 13:01:44 -03:00
  • ad1a54b5c0 reordering of some defines (more logical grouping) Roberto Ierusalimschy 2011-07-02 13:00:15 -03:00
  • 88c29c0cd5 compilation warning about integer size x pointer size Roberto Ierusalimschy 2011-07-02 12:59:17 -03:00
  • d93df22bf6 small simplifications (and assert was wrongly written) Roberto Ierusalimschy 2011-07-02 12:58:14 -03:00
  • 5017cc6ba2 simpler 'luai_apicheck' (and avoids compilation error...) Roberto Ierusalimschy 2011-07-02 12:57:25 -03:00
  • 2f125ebc33 'deprecated' x 'removed' (they are not the same thing) Roberto Ierusalimschy 2011-07-02 12:56:43 -03:00
  • 49a01c0991 'package.loaders' -> 'package.searchers' Roberto Ierusalimschy 2011-06-28 14:14:12 -03:00
  • 17526ed2fe no more errors on accesses to 'string' and strings Roberto Ierusalimschy 2011-06-28 14:13:52 -03:00
  • f180822fa6 back to internal table for 'ctype', no more locale manipulation Roberto Ierusalimschy 2011-06-28 12:42:04 -03:00
  • 594220c624 use standard "underscore + uppercases" for reserved keys in the registry Roberto Ierusalimschy 2011-06-27 16:42:31 -03:00
  • 3dd68624dc comments Roberto Ierusalimschy 2011-06-27 15:27:18 -03:00
  • f9eea453b7 avoid "magic constant" 32 (the difference from 'A' to 'a') Roberto Ierusalimschy 2011-06-27 15:22:46 -03:00
  • 374773748b back to "one-label-per-block" Roberto Ierusalimschy 2011-06-27 15:18:59 -03:00
  • 29644366fc useless assignments removed (macros already do the assignment) Roberto Ierusalimschy 2011-06-24 11:36:21 -03:00
  • 6eadedbfa1 resort to standard C ctype for non-ASCII systems + 'ltoupper' replaced by 'ltolower' Roberto Ierusalimschy 2011-06-24 09:25:33 -03:00
  • 1978094b3a resort to standard C ctype for non-ASCII systems Roberto Ierusalimschy 2011-06-24 09:25:33 -03:00
  • 5acc5b0c1e using 'tolower' instead of 'toupper' (slightly simpler to implement) Roberto Ierusalimschy 2011-06-24 09:25:02 -03:00
  • 59547c6cda macro 'ltoupper' instead of 'lisupper' (more usefull for Lua) Roberto Ierusalimschy 2011-06-23 13:01:06 -03:00
  • 3d3355b4f3 label syntax changed to '::label::' Roberto Ierusalimschy 2011-06-23 11:54:48 -03:00
  • 801f43f09a change in the representation of file handles Roberto Ierusalimschy 2011-06-21 10:43:48 -03:00
  • 719c01359f label syntax changed to '::label::' Roberto Ierusalimschy 2011-06-20 13:52:48 -03:00
  • 43f13729a2 comment Roberto Ierusalimschy 2011-06-20 13:50:59 -03:00
  • 136f8e148d added cast to void to calls to 'setlocale' (to easy redefining it as a constant) Roberto Ierusalimschy 2011-06-20 13:36:03 -03:00
  • f90a130b07 typo in message Roberto Ierusalimschy 2011-06-20 13:35:23 -03:00
  • cb4950a6d8 parser must work with C locale, but external locale must be preserved Roberto Ierusalimschy 2011-06-18 14:08:58 -03:00
  • fdede85419 label names must be unique inside a function Roberto Ierusalimschy 2011-06-16 13:36:39 -03:00
  • beab626061 correct length for LUA_COPYRIGHT (without trailing '\0') Roberto Ierusalimschy 2011-06-16 11:30:58 -03:00
  • 817f8674af avoid warning about -unsigned value Roberto Ierusalimschy 2011-06-16 11:14:31 -03:00
  • cf0562e1e7 detail (missing cast to unsigned char when converting char to int) Roberto Ierusalimschy 2011-06-16 11:13:22 -03:00
  • 22e1c5ebad avoid conversion from 'const char*' to 'void *' (warning about 'const') Roberto Ierusalimschy 2011-06-16 11:12:24 -03:00
  • 470dd56a89 'os.execute' (and similars) should return 'exit' and code in case of success, too. Roberto Ierusalimschy 2011-06-16 11:11:04 -03:00
  • 20d30bcd33 declarations should be present only when implementations are Roberto Ierusalimschy 2011-06-16 11:10:12 -03:00
  • 44b6f496b5 just in case, avoid side effects in 'ctype' macros Roberto Ierusalimschy 2011-06-15 11:35:55 -03:00
  • fd80e63468 configuration for NaN trick big-endian + macro 'luai_checknum' to ensure numbers comming from C are not "improper" (some kinds of signaling NaNs) Roberto Ierusalimschy 2011-06-13 11:13:06 -03:00
  • f62565abea avoid warnings with -Wstrict-overflow Roberto Ierusalimschy 2011-06-09 15:24:22 -03:00
  • c9d1d9f9c6 better way to test overflows with string indices Roberto Ierusalimschy 2011-06-09 15:22:47 -03:00
  • 35391d9b1a macro 'NNMARK' (mark for non-number values using the NaN trick) Roberto Ierusalimschy 2011-06-09 15:21:25 -03:00
  • 99b1b8e918 first version with 'NANTRICK' (packing all Lua values inside a 'double') Roberto Ierusalimschy 2011-06-07 16:02:33 -03:00
  • 3b44821334 stricter control (using tag variants) over closure kinds (Lua x C) Roberto Ierusalimschy 2011-06-02 16:31:40 -03:00
  • 0214dab989 yet another macro ('num_') Roberto Ierusalimschy 2011-05-31 16:15:01 -03:00
  • 821bd7025e more uniform names for 'equalobj'-related functions Roberto Ierusalimschy 2011-05-31 15:27:56 -03:00
  • 9b7dddad7d no need for two different implementations for equality (one raw and one with metamethods) Roberto Ierusalimschy 2011-05-31 15:24:36 -03:00
  • 3f04a9f2c0 code is simpler without 'goto' here Roberto Ierusalimschy 2011-05-30 13:36:38 -03:00
  • 2a59a999d4 yet another macro abstracting TValue representation Roberto Ierusalimschy 2011-05-26 14:12:31 -03:00
  • ee7d0c2649 new macro 'luai_writeline' to print newlines (and flush 'stdout') Roberto Ierusalimschy 2011-05-26 13:09:40 -03:00
  • aa13c591f5 avoid use of 'ifdef/ifndef'; use 'defined' instead (simpler and more powerful) Roberto Ierusalimschy 2011-05-25 11:12:28 -03:00
  • 6c8a32217a changes by lhf (better control of chars x bytes) Roberto Ierusalimschy 2011-05-17 09:42:43 -03:00
  • cb09f4fef4 details Roberto Ierusalimschy 2011-05-06 13:00:23 -03:00
  • 3cf1729a02 new macro 'lua_longassert' that is equivalent to an assertion without a stringfication of the condition, to avoid too long string literals (limited by C90 to ~510 characters) Roberto Ierusalimschy 2011-05-05 16:43:14 -03:00
  • 74b0bb3a45 macro 'checkconsistency' moved to this file, as it is used only here Roberto Ierusalimschy 2011-05-05 16:42:25 -03:00
  • e4076fe6d9 break long assertion in two shorter ones Roberto Ierusalimschy 2011-05-05 13:18:53 -03:00
  • 81952e6884 avoid using expression as argument to unsafe macro 'l_isfalse' Roberto Ierusalimschy 2011-05-05 13:16:33 -03:00
  • 25b4e35ee8 new macros abstracting TValue representation (to easy change to other representations) Roberto Ierusalimschy 2011-05-05 12:53:23 -03:00
  • 7e1facaa4e detail in assertion Roberto Ierusalimschy 2011-05-04 14:04:06 -03:00
  • ad2531a0ee more complete (and hopefuly more correct) handling of 'sizeof(char)' Roberto Ierusalimschy 2011-05-03 13:01:57 -03:00
  • bc1c718cc0 removed TOKEN_LEN (useless) Roberto Ierusalimschy 2011-05-03 12:51:16 -03:00
  • c9ce754e38 comment Roberto Ierusalimschy 2011-05-02 14:33:01 -03:00
  • d34271fe42 MacOS 'strtod' handles hexa formats too Roberto Ierusalimschy 2011-04-29 10:56:28 -03:00
  • d120ec29ca bug in OP_SELF when method name goes to a register Roberto Ierusalimschy 2011-04-28 11:00:11 -03:00
  • 2aff901c93 Mac OS defines LUA_USE_AFORMAT and LUA_USE_LONGLONG by default Roberto Ierusalimschy 2011-04-20 15:25:54 -03:00
  • bad231c481 'LUA_USELONGLONG' -> 'LUA_USE_LONGLONG' Roberto Ierusalimschy 2011-04-20 13:36:28 -03:00
  • a2b1c279f1 'LUA_USELONGLONG' on by default on Linux + LUA_USE_HEXAFLOAT renamed to LUA_USE_STRTODHEX (as Lua 5.2 always *uses* hexa floats) Roberto Ierusalimschy 2011-04-19 15:31:30 -03:00