 joost
|
addb92e62b
* Replaced global variable for field
|
8 years ago |
 svenbarth
|
11493de99b
+ add test from Mantis #31305 for completeness sake
|
8 years ago |
|
svenbarth
|
c1390b3442
* fix for Mantis #31249: applied (adjusted) patch provided by Maciej Izak
|
8 years ago |
 Károly Balogh
|
a34b6c7c53
m68k: atari also uses syscalls, so update the comment for that too
|
8 years ago |
|
svenbarth
|
d149f260fc
- remove no longer needed gen_fpc_dummy()
|
8 years ago |
|
svenbarth
|
cc42bb8a7a
* generate the package stub for all Windows and NativeNT targets
|
8 years ago |
|
svenbarth
|
48bddbbc43
* use new potype_pkgstub instead of gen_fpc_dummy() as the former works on all platforms that needs it
|
8 years ago |
|
svenbarth
|
1a6a8b7c9f
+ introduce a new potype for the main stub of a package library (which on Windows is simply a DLLMain returning True)
|
8 years ago |
|
svenbarth
|
1c98a48c47
* mark __FPC_specific_handler as an external symbol
|
8 years ago |
|
svenbarth
|
cfa5b9d809
- the external declaration of PascalMain is no longer needed for Win64 on 3.1.1+
|
8 years ago |
|
svenbarth
|
69fa837266
* NASM supports DQ already since version 2.00 (from 10 years ago) and it's already in the ait_const2str array, so use it! (doesn't mean that it works on Win64 however...)
|
8 years ago |
|
svenbarth
|
f517311227
* ensure that the indirect symbols for VMT, RTTI and IID reside in the same sections as the symbols they point to -> reduces the amount of sections for them by half
|
8 years ago |
|
svenbarth
|
c1db48c8c6
* add ability to disable the generation of an indirect symbol if needed (useful if that's done by the constant builder)
|
8 years ago |
|
svenbarth
|
822d96776b
* extend tai_typedconstbuilder with the ability to add an indirect symbol to the same section as the one that's generated for the constant data (at least if it's also rodata)
|
8 years ago |
|
svenbarth
|
b813fc641f
* put data that's used by the VMT into the same section (that also means changing some sections from rodata_norel to rodata); this way the amount of sections per class is reduced (e.g. two for classes that inherited from a class with interfaces)
|
8 years ago |
|
svenbarth
|
deb86a9e70
* Windows and GNU AS both correctly handle read only sections with relocations, so make use of that
|
8 years ago |
 michael
|
f55215f53a
* Case else without semicolon in front (bug ID 31284)
|
8 years ago |
|
michael
|
a51efe50b7
* noreturn procedure modifier (bug ID 31285)
|
8 years ago |
|
michael
|
0dc3d03755
* Packed sets support
|
8 years ago |
|
michael
|
892b616adf
* Fix compilation
|
8 years ago |
|
michael
|
a809705fb8
* New simple program to test if pparser parses some code
|
8 years ago |
|
michael
|
81a00358a1
* Modeswitches now parsed correctly (bug ID 30724)
|
8 years ago |
|
michael
|
8d7fd39282
* Fix isNan, isInfinity, IsPositiveInfinity
|
8 years ago |
|
svenbarth
|
060cb62f2a
* correctly handle alignment for TParameterLocations
|
8 years ago |
|
svenbarth
|
1d40e7e159
* correctly encapsulate the location records
|
8 years ago |
|
michael
|
27b51188b6
* Range typed consts support
|
8 years ago |
|
svenbarth
|
c2d68e35b5
* also assume register calling convention for m68k
|
8 years ago |
|
svenbarth
|
4d41decd5d
* since m68k has a working implementation for the register calling convention and it's in fact used as default calling convention it should also be possible to explicitely declare routines as that calling convention
|
8 years ago |
|
svenbarth
|
d35e9a69fa
* update comment regarding pocall_syscall in supported_calling_conventions
|
8 years ago |
 Jonas Maebe
|
098d64b631
* don't swap and/or boolean operations, they always need to be evaluated in
|
8 years ago |
