| Wed, 27 May 2009 19:20:18 +0200 |
Christian Kamm |
Error instead of assert on delegate literals as constant expressions.
0.9.1
|
| Sun, 17 May 2009 04:41:10 +0200 |
Tomas Lindquist Olsen |
Cleaned up TypeInfo_Struct.
|
| Sat, 16 May 2009 18:19:52 +0200 |
Tomas Lindquist Olsen |
Added IrTypeFunction and IrTypeDelegate and eliminated IrDType. This means the Type::ir field can be removed. It's the final part needed for the move to a slightly more sane type system. Now the whole thing just needs to be cleaned up :P
|
| Thu, 14 May 2009 20:36:55 +0200 |
Frits van Bommel |
Remove struct padding when passing or returning in registers on x86-64 (extern(D) only)
|
| Tue, 12 May 2009 16:02:51 +0200 |
Frits van Bommel |
Remove some other code that tested for LLVM before r67588.
|
| Sat, 09 May 2009 18:57:00 +0200 |
Christian Kamm |
Update for LLVM rev > 71348: getTypePaddedSize -> getTypeAllocSize
|
| Wed, 29 Apr 2009 18:59:39 +0200 |
Tomas Lindquist Olsen |
Fixed array operation codegen when used from imported templates. All array operations are now emitted with template linkage. Fixed #270 .
|
| Fri, 17 Apr 2009 01:19:03 +0200 |
Tomas Lindquist Olsen |
Merge with rev [1221]
|
| Thu, 16 Apr 2009 20:31:21 +0200 |
Tomas Lindquist Olsen |
Merge.
|
| Wed, 15 Apr 2009 20:06:25 +0200 |
Tomas Lindquist Olsen |
Added Doxygen file.
|
| Fri, 17 Apr 2009 01:14:35 +0200 |
Frits van Bommel |
Unbreak nested naked functions.
|
| Wed, 15 Apr 2009 21:37:01 +0200 |
Frits van Bommel |
Name some types.
|
| Mon, 13 Apr 2009 16:16:03 +0200 |
Frits van Bommel |
Give all symbols nested in functions internal linkage, unless it's one of the
|
| Sun, 12 Apr 2009 21:56:43 +0200 |
Frits van Bommel |
Add some alignment info where LLVM might otherwise be more pessimistic.
|
| Fri, 03 Apr 2009 16:34:11 +0200 |
Tomas Lindquist Olsen |
Started seperating type resolution from the rest of codegen again, the merge had too many regressions.
|
| Mon, 30 Mar 2009 16:34:36 +0200 |
Frits van Bommel |
Give function & delegate literals, foreach bodies and lazy parameters internal
|
| Tue, 24 Mar 2009 15:24:59 +0100 |
Frits van Bommel |
Update DtoConstFP() to be correct after LLVM r67562, which changed the way the
|
| Sun, 08 Mar 2009 16:13:10 +0100 |
Frits van Bommel |
Make LDC work with LLVM trunk (s/LinkOnceLinkage/LinkOnceOdrLinkage/)
|
| Fri, 06 Mar 2009 21:15:13 +0100 |
Frits van Bommel |
Fix a bug in the X86 ABI. The size of a struct is different from the size of a
|
| Tue, 03 Mar 2009 02:51:21 +0100 |
Tomas Lindquist Olsen |
Abstracted more (most) ABI details out of the normal codegen.
|
| Thu, 26 Feb 2009 14:51:02 +0100 |
Frits van Bommel |
Use LLVM OStream wrapper instead of <iostream> in the logger.
|
| Thu, 26 Feb 2009 14:11:49 +0100 |
Tomas Lindquist Olsen |
- Fixed LLVM style CL args for D2.
|
| Sun, 08 Feb 2009 05:26:54 +0100 |
Tomas Lindquist Olsen |
SWITCHED TO LLVM 2.5 !
|
| Wed, 04 Feb 2009 18:48:03 +0100 |
Tomas Lindquist Olsen |
Changed the way LDC determines if a template instantiation needs to get a definition, seems to speed up compile times quite a bit in some cases.
|
| Tue, 03 Feb 2009 08:54:57 +0100 |
Tomas Lindquist Olsen |
Implemented proper support for naked asm using llvm module level asm. Still not 100% complete, but already 1000 times better that what we had before. Don's BignumX86 implementation from Tango (when turned into a standalone unittest) seems to fully work with no changes, and great performance :)
|
| Mon, 02 Feb 2009 01:44:51 +0100 |
Tomas Lindquist Olsen |
Changed templates and typeinfo to use linkonce linkage instead of weak linkage, this should fix inlining problems, fixing bug #197 . If problems show up, it's easy to change it back by changing the define in mars.h . I'm 95% sure this is safe, given how we handle templates.
|
| Thu, 22 Jan 2009 17:26:50 +0100 |
Tomas Lindquist Olsen |
Makes sure only function definitions inside templates get weak linkage, llvm doesn't allow weak linkage on just declarations, so we use external linkage for those. Fixes #162
|
| Sun, 14 Dec 2008 04:11:32 +0100 |
Tomas Lindquist Olsen |
Runtime now compiles with D2 again. This does NOT mean it works flawlessly !!!
|
| Mon, 01 Dec 2008 02:10:16 +0100 |
Tomas Lindquist Olsen |
Fixed ModuleInfo generation to no longer use the ModuleInfo class' default initializer for types/defaults, it's unsafe as initializers don't necesarily match the "formal" type. There might be explicit padding.
|
| Sat, 29 Nov 2008 21:25:43 +0100 |
Tomas Lindquist Olsen |
Removed KDevelop3 project files, CMake can generate them just fine!
|
