ea1086b7f2
PR2659 was fixed by r61847. Add the testcase as a regression test.
...
llvm-svn: 61986
2009-01-09 08:16:12 +00:00
f50aa6ae5c
Implement rdar://6480391, extending of equality icmp's to avoid a truncation.
...
I noticed this in the code compiled for a routine using std::map, which produced
this code:
%25 = tail call i32 @memcmp(i8* %24, i8* %23, i32 6) nounwind readonly
%.lobit.i = lshr i32 %25, 31 ; <i32> [#uses=1]
%tmp.i = trunc i32 %.lobit.i to i8 ; <i8> [#uses=1]
%toBool = icmp eq i8 %tmp.i, 0 ; <i1> [#uses=1]
br i1 %toBool, label %bb3, label %bb4
which compiled to:
call L_memcmp$stub
shrl $31, %eax
testb %al, %al
jne LBB1_11 ##
with this change, we compile it to:
call L_memcmp$stub
testl %eax, %eax
js LBB1_11
This triggers all the time in common code, with patters like this:
%169 = and i32 %ply, 1 ; <i32> [#uses=1]
%170 = trunc i32 %169 to i8 ; <i8> [#uses=1]
%toBool = icmp ne i8 %170, 0 ; <i1> [#uses=1]
%7 = lshr i32 %6, 24 ; <i32> [#uses=1]
%9 = trunc i32 %7 to i8 ; <i8> [#uses=1]
%10 = icmp ne i8 %9, 0 ; <i1> [#uses=1]
etc
llvm-svn: 61985
2009-01-09 07:47:06 +00:00
0f7cf1d7e1
Remove some old code that looks like a remanant from signed-types days.
...
llvm-svn: 61984
2009-01-09 07:10:58 +00:00
482eb70a10
Fix PR3298, a crash in Jump Threading. Apparently even
...
jump threading can have bugs, who knew? ;-)
llvm-svn: 61983
2009-01-09 06:08:12 +00:00
d48d1ec320
this doesn't depend on the gcc early inliner anymore.
...
llvm-svn: 61982
2009-01-09 05:49:27 +00:00
7f88a1b512
PR3290 is now fixed.
...
llvm-svn: 61981
2009-01-09 05:46:19 +00:00
fef138b140
Fix part 3/2 of PR3290, making instcombine zap (gep(bitcast)) when possible.
...
llvm-svn: 61980
2009-01-09 05:44:56 +00:00
9170731cb7
this test should not run opt -std-compile-opts, it should run
...
just llc.
llvm-svn: 61979
2009-01-09 05:32:00 +00:00
8b80ed7bf7
add a helper method.
...
llvm-svn: 61978
2009-01-09 05:27:40 +00:00
729a21f947
fit in 80 cols
...
llvm-svn: 61977
2009-01-09 04:58:01 +00:00
a784a2ce01
move some code, check to see if the input to the GEP is a bitcast
...
(which is constant time and cheap) before checking hasAllZeroIndices.
llvm-svn: 61976
2009-01-09 04:53:57 +00:00
c7c9ab7947
Fix rdar://6480479 - [parser] infinite loop on invalid input
...
llvm-svn: 61975
2009-01-09 04:34:13 +00:00
cc1b6e8fce
Convert block types in IRgen. This is not the correct type, but
...
matches llvm-gcc (?).
llvm-svn: 61974
2009-01-09 02:48:46 +00:00
9b21e63251
Block pointer types are not aggregate types.
...
llvm-svn: 61973
2009-01-09 02:44:18 +00:00
bdc0f8b627
Add load-folding table entries for MOVDQA.
...
llvm-svn: 61972
2009-01-09 02:40:34 +00:00
e907a0a527
Whitespace and other minor adjustments to make SSE instructions have
...
the same formatting as their corresponding SSE2 instructions, for
consistency.
llvm-svn: 61971
2009-01-09 02:27:34 +00:00
1e21c19be6
Don't crash when our FunctionDecl has a non-identifier name
...
llvm-svn: 61970
2009-01-09 01:47:02 +00:00
4755d9df78
Adjustments to last patch based on review.
...
llvm-svn: 61969
2009-01-09 01:30:11 +00:00
f4c6616d9d
Fix crash on null deference when searching for readwrite properties in
...
categories.
- Also, simplify nesting via early return.
llvm-svn: 61968
2009-01-09 01:04:21 +00:00
a3491665a6
ccc: Start defining host information.
...
- For use by the driver in places where the host alters driver
behavior (for example, running as a driver driver on darwin).
- Allow user override for testing purposes.
llvm-svn: 61967
2009-01-09 01:00:40 +00:00
c25d7a7fe0
Addressed the issue in <rdar://problem/6479085>, where we failed to
...
rewrite @class declarations that showed up within linkage
specifications because those @class declarations never made it any
place where the rewriter could find them.
Moved all of the ObjC*Decl nodes over to ScopedDecls, so that they can
live in the appropriate top-level or transparent DeclContext near the
top level, e.g., TranslationUnitDecl or LinkageSpecDecl. Objective-C
declarations now show up in a traversal of the declarations in a
DeclContext (they didn't before!). This way, the rewriter finds all
Objective-C declarations within linkage specifications.
llvm-svn: 61966
2009-01-09 00:49:46 +00:00
339153d3be
Re-enable PTH testing for Cocoa.h and Carbon.h (and include testing for Objective-C++).
...
llvm-svn: 61965
2009-01-09 00:41:48 +00:00
e84858ccf9
Adding support for ObjC methods which have c-style
...
parameter list. This is work in progress.
llvm-svn: 61964
2009-01-09 00:38:19 +00:00
b54c19a754
Enable support for '-x objective-c++-header'.
...
llvm-svn: 61963
2009-01-09 00:38:08 +00:00
6697945069
Simpler solution to LiteralSupport compatibility: just add one whitespace character after each cached string.
...
llvm-svn: 61962
2009-01-09 00:37:37 +00:00
8ae06625b5
Invert assertion condition.
...
llvm-svn: 61961
2009-01-09 00:36:11 +00:00
cb5a95f3a4
Temporarily revert r61956 and r61957 (PTH tests failing).
...
llvm-svn: 61960
2009-01-09 00:27:29 +00:00
68b135cc9c
61949 accidentally introduced an escaped newline. Fix this by making
...
the comment a little more verbose.
llvm-svn: 61959
2009-01-08 23:51:48 +00:00
510b4da49d
Enhance -fsyntax-only test of Carbon.h to also include testing for PTH.
...
llvm-svn: 61958
2009-01-08 23:42:56 +00:00
8598addcd7
Enhance -fsyntax-only test of Cocoa.h to also include testing for PTH.
...
llvm-svn: 61957
2009-01-08 23:41:35 +00:00
d4a4fd99de
PTH: For the cached spellings of literals, store one whitespace character after the spelling to accomodate sanity checking in LiteralSuppoert.cpp.
...
llvm-svn: 61956
2009-01-08 23:40:50 +00:00
f646668799
Convert DwarfWriter into a pass.
...
Now Users request DwarfWriter through getAnalysisUsage() instead of creating an instance of DwarfWriter object directly.
llvm-svn: 61955
2009-01-08 23:40:34 +00:00
bfd47a3c00
Add mm_malloc.h, patch by Sam Weinig.
...
llvm-svn: 61954
2009-01-08 23:30:09 +00:00
d11dc0b1d9
Place warning about 'readonly' property attributes which
...
are related to setter syntax under -Wreadonly-setter-attrs
to prevent warnings in projects built with gcc.
llvm-svn: 61953
2009-01-08 23:23:10 +00:00
e4b414c781
Revert my previous, failed attempt to pretty-print anonymous struct/union accesses well. Added a FIXME so we know to revisit this later
...
llvm-svn: 61951
2009-01-08 22:45:41 +00:00
f87dc9264a
Delete unnecessary parens around return values.
...
llvm-svn: 61950
2009-01-08 22:19:34 +00:00
138436f158
Fix the comment for lltok::backslash.
...
llvm-svn: 61949
2009-01-08 22:18:13 +00:00
6031a4353b
Fix the path to llvm/Assembly/Parser.h in a comment.
...
llvm-svn: 61948
2009-01-08 22:17:05 +00:00
385d246e0d
Correct the form of the atomic opcode names in a comment.
...
llvm-svn: 61947
2009-01-08 22:14:50 +00:00
b48fc71fc6
Do not inline functions with (dynamic) alloca into
...
functions that don't already have a (dynamic) alloca.
Dynamic allocas cause inefficient codegen and we shouldn't
propagate this (behavior follows gcc). Two existing tests
assumed such inlining would be done; they are hacked by
adding an alloca in the caller, preserving the point of
the tests.
llvm-svn: 61946
2009-01-08 21:45:23 +00:00
715b2897d8
Use mayBeOverridden here, in anticipation of the
...
day when more linkage types will be handled.
llvm-svn: 61944
2009-01-08 20:55:49 +00:00
0667123ba1
Fix ObjCInterfaceDecl::Destroy and ObjCProtocolDecl::Destroy to iterate and destroy all contained ObjCMethodDecls in one sweep. This fixes a use-after-free error found by valgrind.
...
llvm-svn: 61943
2009-01-08 20:49:27 +00:00
c8938e09ca
Remove double-insertion of EnumConstantDecls. Thanks to Zhongxing Xu for pointing this out
...
llvm-svn: 61942
2009-01-08 20:48:26 +00:00
48ed48212d
Added iterator mechanism to ObjCContainerDecl to iterate over both class and instance methods at the same time.
...
llvm-svn: 61941
2009-01-08 20:48:19 +00:00
82ac25e4a7
Unify the code for defining tags in C and C++, so that we always
...
introduce a Scope for the body of a tag. This reduces the number of
semantic differences between C and C++ structs and unions, and will
help with other features (e.g., anonymous unions) in C. Some important
points:
- Fields are now in the "member" namespace (IDNS_Member), to keep
them separate from tags and ordinary names in C. See the new test
in Sema/member-reference.c for an example of why this matters. In
C++, ordinary and member name lookup will find members in both the
ordinary and member namespace, so the difference between
IDNS_Member and IDNS_Ordinary is erased by Sema::LookupDecl (but
only in C++!).
- We always introduce a Scope and push a DeclContext when we're
defining a tag, in both C and C++. Previously, we had different
actions and different Scope/CurContext behavior for enums, C
structs/unions, and C++ structs/unions/classes. Now, it's one pair
of actions. (Yay!)
There's still some fuzziness in the handling of struct/union/enum
definitions within other struct/union/enum definitions in C. We'll
need to do some more cleanup to eliminate some reliance on CurContext
before we can solve this issue for real. What we want is for something
like this:
struct X {
struct T { int x; } t;
};
to introduce T into translation unit scope (placing it at the
appropriate point in the IdentifierResolver chain, too), but it should
still have struct X as its lexical declaration
context. PushOnScopeChains isn't smart enough to do that yet, though,
so there's a FIXME test in nested-redef.c
llvm-svn: 61940
2009-01-08 20:45:30 +00:00
40e1fd1f79
Add missing castToDeclContext/castFromDeclContext hooks.
...
llvm-svn: 61939
2009-01-08 20:35:58 +00:00
5ca5f4bc0a
Remove a dead decl.
...
llvm-svn: 61938
2009-01-08 20:22:04 +00:00
6c879a1448
Move FIXME to a better location.
...
llvm-svn: 61937
2009-01-08 20:17:34 +00:00
e538c5f3ae
Removed ObjCContainerDecl::getPropertyMethods()...doesn't belong in the AST.
...
Moved logic to Sema::ProcessPropertyDecl().
llvm-svn: 61936
2009-01-08 20:15:03 +00:00
11b387fee0
Remove redundant method context (now that ObjCMethodDecl isa ScopedDecl).
...
Convert clients to use the standard getDeclContext() API.
Doug, thanks for the review!
llvm-svn: 61935
2009-01-08 19:41:02 +00:00
Dan Gohman