 svenbarth
|
345d83c3e3
Merge RTTI changes from packages branch (including adjustments that were required due to changes in trunk since then). These changes favor source backwards compatibility in contrast to Delphi compatibility. Binary compatiblity is however drastically broken due to the indirect references that are platform independant!
|
9 years ago |
 Jonas Maebe
|
5af40f22e8
* handle enums with a packenum setting that is larger than necessary to store
|
9 years ago |
|
Jonas Maebe
|
f55a403270
* specify the packenum setting to use when creating a tenumdef
|
9 years ago |
|
Jonas Maebe
|
d5e5ca0531
* fixed conversion of LOC_JUMP in location_force_reg() for C-style booleans
|
9 years ago |
|
Jonas Maebe
|
b76539a049
* fixed handling of uninitialised complex function results
|
9 years ago |
|
Jonas Maebe
|
a59fc7c165
- disable attempts to generate debug info when using llvm for now
|
9 years ago |
|
Jonas Maebe
|
ea54b6b2c5
* factored comp/currency-in-register handling out of a_loadfpu_reg_reg()
|
9 years ago |
 joost
|
561c7b0862
* Added ability to clear global-variable containing the compiler-version info
|
9 years ago |
 florian
|
7f44774852
* i386 uses OptPass1And from aoptx86
|
9 years ago |
|
florian
|
ff506eb219
* enabled And2Test optimization
|
9 years ago |
|
florian
|
a5fc8b629a
* factored out OptPass1AND from x86-64 and moved it to aoptx86
|
9 years ago |
|
florian
|
5e8e21c1be
* factored out OpPass2MOV code, x86-64 uses it as well now
|
9 years ago |
|
florian
|
4758271c91
* x86-64 uses OptPass1MOV
|
9 years ago |
|
florian
|
6609e8def2
* correctly set the size of the operation when carrying out Mov Op Mov optimizations
|
9 years ago |
|
florian
|
b274599a47
* made some assembler optimizer methods class methods
|
9 years ago |
|
florian
|
e56147ac6e
* integrated mov op mov -> op optimization in aoptx86
|
9 years ago |
|
florian
|
9ed6638baa
* moved MovAnd2Mov optimization to aoptx86
|
9 years ago |
|
Jonas Maebe
|
7f2d970a1e
* fixed type encoding of C-style blocks for LLVM
|
9 years ago |
|
Jonas Maebe
|
951ca4e62b
* fixed procdef used to call C-style blocks (only relevant for LLVM)
|
9 years ago |
|
Jonas Maebe
|
ccf011dbc5
* C-style blocks are implicit pointers -> don't parse such typed constants
|
9 years ago |
|
florian
|
8c5fdc9dc2
* softfpu patch by Max Nazhalov: Small fixes to softfpu.pp complementing r27215
|
9 years ago |
|
Jonas Maebe
|
0ed6c3d80e
* also use the last floating point parameter register on AArch64 in case of
|
9 years ago |
 Tomas Hajny
|
cd236d1811
* fcl-pdf package added
|
9 years ago |
|
Tomas Hajny
|
942360289f
* added short name for fcl-pdf
|
9 years ago |
|
florian
|
dfe41db13f
* call GetNextInstruction only once when carrying out mov optimizations
|
9 years ago |
 michael
|
022e099e41
* Corrected implementation of ServiceRunning and WaitForServicecompletion
|
9 years ago |
|
michael
|
ef7c327c7a
* Revert rev r33905, started from too old version
|
9 years ago |
|
michael
|
d37dfecbd5
* Fix unit name
|
9 years ago |
 Károly Balogh
|
2eeb078f3e
m68k: removed obsolete and wrong comment. the '881 extended type is actually 12 bytes during load/store, not 10, which means longword alignment is fine. no functional changes.
|
9 years ago |
|
florian
|
d05222990e
* while updating used register sets, all labels can/must be skipped
|
9 years ago |
