 michael
|
b09139a7c4
* Some fixes after first test round
|
13 tahun lalu |
 marco
|
f89e32c034
* expanded VK_ codes. Patch by Nikolay N. Mantis #20801
|
13 tahun lalu |
 Tomas Hajny
|
c0c526efbc
* DosQuerySysState and DosReplaceModule added
|
13 tahun lalu |
 svenbarth
|
aca2bd83c7
pexpr.pas, post_comp_expr_gendef:
|
13 tahun lalu |
|
michael
|
a051684e61
* Added lots of commands to fpdoc file manager, moved to separate class and unit for reuse
|
13 tahun lalu |
 Mattias Gaertner
|
9d30bf6a9f
avl_tree: fixed FindPointer if on last node
|
13 tahun lalu |
 sergei
|
8dc6ebb5ed
rtl/generic.inc, minor optimizations:
|
13 tahun lalu |
|
sergei
|
32094761ef
* win32/system.pp, win64/system.pp: Moved duplicated declarations of WinAPI functions to sysos.inc (Florian's comment about include ordering is no longer true, wstrings.inc comes after sysos.inc these days).
|
13 tahun lalu |
|
michael
|
c17d15cf72
* Start of project file manipulation program
|
13 tahun lalu |
|
michael
|
e07646a653
* Implemented project file writing functionality, based on patch from Hans-Peter Diettrich
|
13 tahun lalu |
|
svenbarth
|
06ebfcd360
Added a few more tests. All except tgeneric65.pp (object inside generic record) and tgeneric68.pp (object inside generic object) are successfully compiled.
|
13 tahun lalu |
|
svenbarth
|
35377bae67
Fix a remaining artefact from the overloaded symbols approach (just a comment, but nevertheless a change...)
|
13 tahun lalu |
|
svenbarth
|
6aeceb92e8
ptype.pas, read_named_type, expr_type:
|
13 tahun lalu |
|
svenbarth
|
d94a55604e
Fix the test. It's mode Delphi, but does not compile in Delphi, because "TSomeRecord" and "TSomeRecord<T>" are different identifiers.
|
13 tahun lalu |
|
svenbarth
|
6a83ebc4a5
Fix introduced regressions.
|
13 tahun lalu |
|
svenbarth
|
d55684937f
generate_specialization:
|
13 tahun lalu |
|
sergei
|
37abdd52df
* Reverted r19693, calling SysInitMultithreading from PROCESS_ATTACH is plain wrong because it depends on initialized thread manager. Calling InitSystemThreads is more correct, but in case of Win32 it further depends on EnrtyInformation, which is not available at this point. Properly solving the issue requires moving the entire TLS stuff into sysinit units or something like that.
|
14 tahun lalu |
|
sergei
|
e62cd71a2f
* Fixed cycling with 2.4.4 after r19707 (2.4.4 does not define FPC_HAS_FEATURE_RESOURCES, so related stuff is not compiled)
|
14 tahun lalu |
|
sergei
|
0093517227
* sysutils: changed RunErrorToExcept to use table instead of case statement for OS exceptions. This is to be able to extract exception class alone for the upcoming SEH support.
|
14 tahun lalu |
|
marco
|
0d562f04ff
* Improve CHM versioning, now that a compiler is added. Main chm version is in chmbase
|
14 tahun lalu |
|
sergei
|
816953990d
* Do not access fields of potentially unassigned resultdef. p1.resultdef is not assigned at this point when dotypecheck=false. Mantis #20765.
|
14 tahun lalu |
|
sergei
|
266acd1bed
* Windows TLS callbacks, call SysInitMultithreading early, it is necessary because if a statically loaded DLL creates a thread, apparently DLL_THREAD_ATTACH callback executes right away before the entry point. In turn, DLL_THREAD_ATTACH requires TLSIndex to be initialized.
|
14 tahun lalu |
|
sergei
|
c8e65c501a
* Don't optimize away implicit initialization/finalization procedures if corresponding clause is present in the source but is empty. Resolves #19701.
|
14 tahun lalu |
 Jonas Maebe
|
b584c71e42
* use {$IFNDEF cpu64bitaddr} instead of {$IFDEF cpu32bit} so that it's
|
14 tahun lalu |
|
Jonas Maebe
|
f7adaef12a
* defined cpu32bit for mips
|
14 tahun lalu |
|
Jonas Maebe
|
5816f0951d
* fix alignment of sc80real on darwin
|
14 tahun lalu |
|
Jonas Maebe
|
a1ff3b24b9
* fixed warning/error when cross-compiling from 64 bit to 32/16 bit target
|
14 tahun lalu |
|
Jonas Maebe
|
de3a7f0cfd
* the size of the FPC_RESSYMBOL pointer must depend on the target
|
14 tahun lalu |
 pierre
|
2e4b7ae05c
Apply fixes from Sven Barth
|
14 tahun lalu |
|
svenbarth
|
68624aab3a
Incorporate the changes from trunk into "postfixoperators" and "handle_factor_typenode". The latter needed to be extended by a parameter "typeonly" which is "false" in almost all calls except the one inside "factor_read_id" where the "typeonly" parameter of "factor" is used.
|
14 tahun lalu |
