D/ldc
1
0
Fork 0
mirror of https://github.com/xomboverlord/ldc.git synced 2026-01-20 23:03:14 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity
1,412 Commits 10 Branches 8 Tags
adfc5b3ee9ceea28af4d655133f42bdf7fcdc11f
Commit Graph

913 Commits

Author SHA1 Message Date
Frits van Bommel
adfc5b3ee9 Fix scope(exit) foreach (ref v; arr) foo(v); 2009-05-22 21:38:01 +02:00
Frits van Bommel
59f9b35cea Count the sret register as well when keeping track of how many integer registers 
are available for extern(C) functions on x86-64.
Interestingly, llvm-g++ seems to have a very similar bug: http://llvm.org/pr4242
(So this breaks ABI-compatibility with llvm-gcc for this corner case, but gains
it with gcc...)

To clarify, this is about code like this:
{{{
struct S { void*[3] data; }
struct T { void*[2] data; }

// The T should be passed in memory, and p in the last int register.
extern(C) S fail(int, int, int, int, T t, void* p) {
    S s;
    s.data[0] = t.data[0];
    s.data[1] = t.data[1];
    s.data[2] = p;
    return s;
}
}}}
which should generate code functionally equivalent to this:
{{{
extern(C) S* succeed(S* s, int, int, int, int, T t, void* p) {
    s.data[0] = t.data[0];
    s.data[1] = t.data[1];
    s.data[2] = p;
    return s;
}
}}}
(with the same definitions for S and T)
 2009-05-22 13:17:06 +02:00
Christian Kamm
86e1457dde Automated merge with http://hg.dsource.org/projects/ldc 2009-05-22 10:18:04 +02:00
Christian Kamm
c433604243 Fix bug #307 by doing what DMD does: add the arrayop to 
sc->module->importedFrom instead of sc->module. That way array ops in template
instantiations will be semantic3'ed.
 2009-05-22 10:17:47 +02:00
Frits van Bommel
2739cd108c Let TargetABI::passByValue() know when a parameter is byref (either explicitly 
or because it's a static array) by giving it the pointer type instead of the
type itself.
This fixes Derelict compilation on x86-64, where the ABI wasn't expecting a
static array to be passed in at all.
 2009-05-22 00:19:54 +02:00
Frits van Bommel
2f150ddd62 Oops, didn't mean to change debug info linkage. 2009-05-21 15:26:46 +02:00
Frits van Bommel
a520661667 Now that templates instantiations are no longer emitted for all modules that 
even blink at them they seem to break due to being linkonce (if compiled with
any optimization level > 0), so let's give them weak linkage instead.
The difference is that unreferenced linkonce symbols can be deleted, while
weak symbols need to be preserved.
 2009-05-21 15:23:28 +02:00
Christian Kamm
f675d36cfb Tentative fix for semantic3 on imported modules and unnecessary template 
function definition issue. Please test!
Also change linkage of __interfaceInfos to external (same as __Class, __vtbl,
__init). The other change might make this superfluous.
 2009-05-21 10:56:04 +02:00
Christian Kamm
8e56fe69a4 Merged xfBuild patch for dependency tree generation. See #286. 2009-05-20 21:13:41 +02:00
Frits van Bommel
95c35225bb Don't print the entire declaration of the alliassee when ->toChars() is 
called on an `AliasDeclaration`; just printing the name will do. This fixes
#305, which otherwise tries to generate
{{{
  class E {
      void A() {
          alias /* recurse into E->toCBuffer() */ m;
      }
  }
}}}
by way of an infinite recursion (causing a segfault when the stack runs out).
 2009-05-20 16:20:59 +02:00
Frits van Bommel
3164ceba17 Remove some overly verbose logging. 2009-05-19 20:46:33 +02:00
Kelly Wilson
16a2512808 remove extra 'test' instructions and fix 'test' properly 2009-05-19 12:28:44 -06:00
Kelly Wilson
8941b6960c fix missing 'test' mnemonic and add several variations 2009-05-19 11:58:54 -06:00
Tomas Lindquist Olsen
f733680251 Try to cut down reallocations when building string literals. 2009-05-18 16:01:22 +02:00
Tomas Lindquist Olsen
be25445461 Added simple check for ClassInfo change introduced in DMD 1.045, it must have 12 fields. 2009-05-17 23:15:04 +02:00
Frits van Bommel
827880e467 Don't try to strip metadata if we don't generate any 2009-05-17 22:40:02 +02:00
Tomas Lindquist Olsen
0b8413b4bc Fixed class -> integer casts. 2009-05-17 22:02:03 +02:00
Tomas Lindquist Olsen
2690f3cfd8 Add debug info for all class fields, including those from super classes... 2009-05-17 17:55:53 +02:00
Tomas Lindquist Olsen
e2ba08388f Added new FreeBSD and Solaris predefined version identifiers. 2009-05-17 17:10:26 +02:00
Tomas Lindquist Olsen
b64ad3299e Use RTTIBuilder to build !ModuleInfo, still needs some work. 2009-05-17 17:04:47 +02:00
Tomas Lindquist Olsen
3e4ade93ba Updated comment in RTTIBuilder.h 2009-05-17 16:31:23 +02:00
Tomas Lindquist Olsen
b1fa43ce24 Increased RTTIBuilder buffer to 14 constants, to be big enough for !ClassInfo as well as !TypeInfo 2009-05-17 16:29:49 +02:00
Tomas Lindquist Olsen
dc4b7e8118 Update !ClassInfo generation to use !RTTIBuilder, slight update of !RTTIBuilder . 2009-05-17 16:27:01 +02:00
Tomas Lindquist Olsen
b6daf3d545 Renamed !TypeInfoBuilder to !RTTIBuilder. 2009-05-17 15:20:58 +02:00
Tomas Lindquist Olsen
0414a5acbb Fixed filename in array bounds check for mixed in imported template function. Fixes ticket #295 . 2009-05-17 14:56:29 +02:00
Tomas Lindquist Olsen
9be5694168 Added error messages when failed to open files for .bc and .ll output, instead of just trying to write to the stream, pretty similar to the #281 problem. 2009-05-17 14:40:09 +02:00
Tomas Lindquist Olsen
0a77bd9fa6 Add error message when failed to open output stream for native asm output. Fixes ticket #281 . 2009-05-17 14:20:27 +02:00
Frits van Bommel
b52ee2de13 Use %La to hex-format a real instead of pretending it's an integer. 2009-05-17 11:06:14 +02:00
Frits van Bommel
a75ddbf803 Specify a large integer constant more portably. 2009-05-17 10:33:21 +02:00
Tomas Lindquist Olsen
5078019c54 Cleaned up TypeInfo_Tuple generation. 2009-05-17 05:07:51 +02:00
Tomas Lindquist Olsen
d80de9deb3 Cleaned up TypeInfo_Struct. 
Fixed problem with DtoConstSize_t taking a size_t argument, this is not enough for cross compiling from 32bit host to a 64bit target. It now takes uint64_t. There's probably a lot of similar case around to code ...
 2009-05-17 04:41:10 +02:00
Tomas Lindquist Olsen
f3c7278e88 Cleaned up TypeInfo generation, still need to do TypeInfo_Struct/Tuple. Eventually do ClassInfo and ModuleInfo as well using same interface. 2009-05-17 03:10:55 +02:00
Frits van Bommel
371cdb71ba Remove an incorrect attribute: noalias doesn't apply to struct types, even if 
we prefer to think of that type as an array :(.
 2009-05-17 02:22:21 +02:00
Frits van Bommel
76ae0b0ab6 Fix format-string bugs by adding __attribute__((__format__)) in all applicable 
places and fixing all warnings my gcc produced.
Among other things, this should fix several segfaults (including one I just
ran into).
 2009-05-17 00:15:25 +02:00
Tomas Lindquist Olsen
6633d696c9 Forgot to set the (ClassInfo.flags & 32) bit added in 1.045. 2009-05-16 22:24:49 +02:00
Tomas Lindquist Olsen
e8780d50e8 Merged DMD 1.045 !!! 2009-05-16 22:21:31 +02:00
Tomas Lindquist Olsen
c0f2af5afd 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 
Added -v-cg switch, which right now just prints "codegen: module.name (module/name.d)" to stdout, this can really help figuring out where, in some complex build command, things go wrong.
 2009-05-16 18:19:52 +02:00
Benjamin Kramer
46b200a4ed silence a gcc warning 2009-05-16 13:50:44 +02:00
Frits van Bommel
c05183123d Automated merge with http://hg.dsource.org/projects/ldc 2009-05-16 13:10:35 +02:00
Frits van Bommel
541f3723cd Intrinsics shouldn't see struct padding, so use a special TargetABI for them 
that removes it.
This unbreaks the `llvm_*_with_overflow` intrinsics.
 2009-05-16 13:06:49 +02:00
Christian Kamm
53c3ac0313 Fix overload resolution issue in dmd bug 313/314 fix. 2009-05-16 12:17:33 +02:00
Tomas Lindquist Olsen
62e8384da0 Set explicit alignment for normal global variables, ensures alignment is correct for globals with union type. 2009-05-15 17:17:20 +02:00
Tomas Lindquist Olsen
0680e6a3ca Enabled fix for ticket #294 . It will try it's best to let LLVM handle the alignment, and fall back to manual padding when that's not possible. If this causes problems again, we'll have to consistently pad manually using packed structs. 2009-05-15 17:05:35 +02:00
Tomas Lindquist Olsen
ad142b5fda Erm... Revert the rev [1352] change, the original code was correct ... 2009-05-15 15:30:59 +02:00
Frits van Bommel
41c738f4da Re-apply [1346], which was reverted in [1348], now that padding should no longer 
be a problem.
As noted before, this will break ABI-compatibility on x86-64, so recompile old
code before linking with newly-compiled code if you're on x86-64.
As an added bonus, I actually tested it this time :).
 2009-05-14 20:40:56 +02:00
Frits van Bommel
287b1278d3 Remove struct padding when passing or returning in registers on x86-64 (extern(D) only) 2009-05-14 20:36:55 +02:00
Tomas Lindquist Olsen
a0678dc9ec Oups, fix assertion condition in disabled code from rev [1351]. 2009-05-14 17:23:55 +02:00
Tomas Lindquist Olsen
b147ecc66d Initial (but disabled) fix for ticket #294 , the actual part that fixes the bug is in a #if 0 block as I'm afraid it will cause regressions. I'm most likely not going to be around tonight, and maybe not tomorrow as well, so I'm pushing it in case someone wants to run some serious testing/investigate the problem noted in llvmhelpers.cpp : realignOffset . 2009-05-14 17:20:17 +02:00
Tomas Lindquist Olsen
f5d635dfc7 Adds explicit alignment information for alloca instructions in general, there's a few cases that still needs to be looked at but this should catch the majority. Fixes ticket #293 . 2009-05-14 13:26:40 +02:00
Christian Kamm
5694ec18ff Use getenv on Windows, closes #248. 2009-05-13 18:08:40 +02:00