aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* some cleaning on signed opcode parametersRoberto Ierusalimschy2017-12-153-27/+50
|
* (1 << 31) with signed integer has undefined behavior in CRoberto Ierusalimschy2017-12-151-2/+2
|
* 'Proto->numparams' does not include vararg parameterRoberto Ierusalimschy2017-12-154-12/+11
| | | | (one less subtraction when calling functions...)
* 'VRELOCABLE' -> 'VRELOC'Roberto Ierusalimschy2017-12-143-22/+22
|
* bug: memory-allocation error when resizing a table can leave itRoberto Ierusalimschy2017-12-131-2/+27
| | | | in an inconsistent state.
* new opcodes BANDK/BORK/BXORK. (They do not use immediate operandsRoberto Ierusalimschy2017-12-137-23/+107
| | | | | because, too often, masks in bitwise operations are integers larger than one byte.)
* Code should not change the stack level after the initialization of aRoberto Ierusalimschy2017-12-131-3/+3
| | | | string buffer.
* when shrinking stack, always shrinks the CI list.Roberto Ierusalimschy2017-12-121-5/+2
| | | | (Stack overflow is not corelated to CI overflow anymore.)
* back to reallocation when resizing the string table.Roberto Ierusalimschy2017-12-121-30/+60
| | | | | (Not a good idea to explicitly allocate new memory when shrinking something.)
* allows memory-allocation errors when shrinking blocksRoberto Ierusalimschy2017-12-111-2/+3
|
* more freedom in handling memory-allocation errors (not all allocationsRoberto Ierusalimschy2017-12-111-2/+2
| | | | | automatically raise an error), which allows fixing a bug when resizing a table.
* 'luaD_growstack' cannot raise any errors when 'raiseerror' isRoberto Ierusalimschy2017-12-112-17/+30
| | | | false (+ some comments)
* 'luaM_shrinkvector' raises an error if it cannot shrink the blockRoberto Ierusalimschy2017-12-111-2/+2
| | | | | (several parts of Lua use array size in protos as proxies for number of valid elements)
* more freedom in handling memory-allocation errors (not all allocationsRoberto Ierusalimschy2017-12-088-95/+136
| | | | | automatically raise an error), which allows fixing a bug when resizing a table.
* new C instruction "rawcheckstack" (to test failing in 'lua_checkstack')Roberto Ierusalimschy2017-12-081-3/+16
|
* using explicit tests for allocation overflow whenever possibleRoberto Ierusalimschy2017-12-075-46/+80
|
* detail (comment)Roberto Ierusalimschy2017-12-071-1/+2
|
* new test function 'T.allocount' to restrict number of allocationsRoberto Ierusalimschy2017-12-072-4/+20
| | | | before a memory-allocation error
* opcodes for order and shift can use several metamethods,Roberto Ierusalimschy2017-12-071-4/+8
| | | | | so it is better to use a generic description + metamethod names in some error messages shown without '__' prefix
* avoid using one function for different tasks (malloc, free, etc.)Roberto Ierusalimschy2017-12-063-47/+108
|
* bug: 'lua_pushcclosure' should not call the garbage collector whenRoberto Ierusalimschy2017-12-061-2/+35
| | | | 'n' is zero.
* bug: 'lua_pushcclosure' should not call the GC when 'n' is zeroRoberto Ierusalimschy2017-12-061-3/+4
|
* new opcodes OP_SHLI/OP_SHRIRoberto Ierusalimschy2017-12-044-18/+94
|
* 'luaS_resize' can raise memory errorsRoberto Ierusalimschy2017-12-011-4/+9
|
* rehashes string table always allocating a new array instead ofRoberto Ierusalimschy2017-12-011-17/+11
| | | | | reallocating old one. (Avoids problems if reallocation to a small size fails.)
* detailRoberto Ierusalimschy2017-12-011-3/+3
|
* another value for LUAI_MAXCCALLS (must think more about that)Roberto Ierusalimschy2017-12-011-2/+2
|
* warnings from Visual Studio /W3Roberto Ierusalimschy2017-11-303-6/+6
|
* small peephole optimizationsRoberto Ierusalimschy2017-11-305-48/+63
|
* details (comments)Roberto Ierusalimschy2017-11-301-5/+10
|
* detail (spacing)Roberto Ierusalimschy2017-11-301-3/+3
|
* new opcodes OP_RETURN0/OP_RETURN1Roberto Ierusalimschy2017-11-294-17/+77
|
* 'luaV_execute' gets call info as extra argument (it is alwaysRoberto Ierusalimschy2017-11-293-11/+11
| | | | available on call sites)
* order opcodes cannot use 'K' operandsRoberto Ierusalimschy2017-11-281-2/+2
|
* conditional jumps unified in label "condjump' + new variable 'vra'Roberto Ierusalimschy2017-11-281-128/+106
| | | | to avoid excessive use of macro 's2v'
* using register 'k' for conditions in tests (we only need one bit there)Roberto Ierusalimschy2017-11-283-92/+120
|
* detail (identation of switch)Roberto Ierusalimschy2017-11-282-37/+35
|
* new opcodes 'OP_LTI' and 'OP_LEI'Roberto Ierusalimschy2017-11-275-12/+71
|
* detail (typo in comments)Roberto Ierusalimschy2017-11-239-19/+19
|
* small simplifications around 'luaT_callorderTM'Roberto Ierusalimschy2017-11-232-21/+19
|
* small simplifications in 'luaD_poscall'Roberto Ierusalimschy2017-11-232-20/+20
|
* no more 'stackless' implementation; 'luaV_execute' calls itselfRoberto Ierusalimschy2017-11-239-112/+90
| | | | | recursively to execute function calls. 'unroll' continues all executions suspended by an yield (through a long jump)
* more information from 'T.stacklevel'Roberto Ierusalimschy2017-11-231-2/+4
|
* using 'A' for register instead of 'B' in relational opcodesRoberto Ierusalimschy2017-11-223-36/+31
| | | | ('R(A)' is already created by default for all instructions.)
* new opcode 'OP_EQI' for equality with immediate numbersRoberto Ierusalimschy2017-11-224-25/+68
|
* more direct implementation for tail calls.Roberto Ierusalimschy2017-11-213-36/+57
|
* in order comparison opcodes, fast track for floats tooRoberto Ierusalimschy2017-11-201-18/+42
|
* 'io.read' accepts multiple formats in a single string argumentRoberto Ierusalimschy2017-11-161-27/+31
|
* detail ('signal' -> 'sign' in comments)Roberto Ierusalimschy2017-11-165-13/+13
|
* new instruction 'OP_EQK' (for equality with constants)Roberto Ierusalimschy2017-11-164-14/+61
|
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-18 23:05:47 +0000