 florian
|
bb8b8f2eae
* do not set IsMultithread on windows, if no second thread is started, resolves #30535
|
8 years ago |
 svenbarth
|
f8b458106e
* fix for Mantis #30747: for anonymous events CreateEvent() needs to be called with a Nil parameter for the name, not an empty string
|
8 years ago |
|
svenbarth
|
535621faaa
Rename FPC_HAS_INDIRECT_MAIN_INFORMATION to FPC_HAS_INDIRECT_ENTRY_INFORMATION which is more logical considering the record's name
|
9 years ago |
|
svenbarth
|
6afda909d4
Rework TlsKey handling on Windows so that it works as intended with indirect main information
|
9 years ago |
 sergei
|
83deae45b8
* First part of 64-bit Windows fixes from Mantis #23074.
|
11 years ago |
|
sergei
|
179586f589
+ SEH support for Win32. Enable by cycling with OPT=-dTEST_WIN32_SEH.
|
11 years ago |
|
sergei
|
6788fd6433
+ Win64 SEH: install a top-level exception handler around thread functions. This was a missing puzzle piece of SEH support. Now behavior of unhandled exceptions in non-main threads is consistent with that of main thread, and, in general, behavior of SEH-enabled RTL is consistent with one without SEH.
|
12 years ago |
|
sergei
|
7969fb4ee1
- Removed TlsSize variable, it duplicates threadvarblocksize.
|
12 years ago |
 pierre
|
f80bc70707
+ Add _FPC_TlsSize public symbol for use by heaptrc unit
|
13 years ago |
|
pierre
|
4a9c8e330b
+ Move EXEC_callback into separate file win/systlsdir.inc
|
13 years ago |
|
sergei
|
ef48f93d85
* win/systhrd.inc: modified ThreadMain so it initializes RTL if not already done, allowing TLS callbacks to be optional.
|
14 years ago |
|
sergei
|
5c3aca5148
* thread.inc, DoneThread: allow CurrentTM.ReleaseThreadVars to be unassigned and reset ThreadID (in preparation to support Windows native threadvars)
|
14 years ago |
|
sergei
|
02a5efc619
* win/systhrd.inc, clean up:
|
14 years ago |
|
sergei
|
138c2b6b0a
* Hacking TLS callbacks into proper shape:
|
14 years ago |
|
pierre
|
1339c66ec9
* SysAllocateThreadVars: Make sure the data slot is only allocated once
|
14 years ago |
|
pierre
|
0fabda01a8
* Added code to support Windows 95 again
|
14 years ago |
 michael
|
cc0be1be4d
* Completely initialize threadvars when an external thread is detected. Needs still cleanup code (or a better thread initialization/finalization).
|
14 years ago |
|
florian
|
27ff441975
o patch from Sven Barth:
|
14 years ago |
 marco
|
8cce3d3f91
* closethread thread manager field/function. Mantis 13160
|
15 years ago |
|
marco
|
c477df5046
* TCriticalSection.Tryenter support (Mantis 15928) + short test/demo
|
15 years ago |
|
florian
|
e6648da567
* make heaptrc aware of the fact that multi threading could be initialized before heaptrc itself is initialized, this is currently the case for windows dlls
|
15 years ago |
|
florian
|
fb07fe5856
* always assume a multithreaded application when using threadvars in dlls, resolves #14992
|
15 years ago |
|
michael
|
12fcb19bf4
* Added comment in SysRelocateThreadVar, not working in Windows Vista
|
16 years ago |
|
florian
|
2f8e263560
* allow only one thread at a time to attach on a dll so the tls slot is allocated properly, resolves #12987
|
16 years ago |
|
florian
|
a8c6d9ec3a
* partially reverted 13571, I didn't get it work on W7 x64 in win32 mode
|
16 years ago |
|
florian
|
d3d3596cd6
* accelerated/fixed TLS access on win32 by Pierre Muller
|
16 years ago |
|
florian
|
380f3ac78f
* fix from Nikolay Samofatov for #12987: initialize thread vars in libraries on windows correctly
|
16 years ago |
 yury
|
3b32571557
* Fixed compilation for win64. ThreadID is DWORD on both win32 and win64.
|
16 years ago |
|
florian
|
88f6b8854e
* proper error handling in case of thread creation failure on windows, resolves #13768, no test case due because such a test is too sensitive regarding race conditions etc. and might cause false failures
|
16 years ago |
|
florian
|
6c5471f324
* don't destroy LastError when resolving a threadvar on windows, resolves #10205 (?)
|
17 years ago |
