4f26205ba8
More then one anonymous aggregates on one line creates chaos when MDNode uniquness is combined with RAUW operation. Right solution is to avoid using RAUW.
...
This fixes PR 6554.
llvm-svn: 98083
2010-03-09 21:32:27 +00:00
f2aa707895
Use getLast() instead of getBasename().
...
llvm-svn: 98072
2010-03-09 19:14:07 +00:00
408dcf60aa
Start using DIFile. Corresponding llvm patch is r98020.
...
llvm-svn: 98021
2010-03-09 00:44:50 +00:00
e18aaf2c4b
add a codegen hack to work around an AST bug, allowing us to compile the
...
code in PR6537. This should be reverted when the ast bug is fixed.
llvm-svn: 97981
2010-03-08 21:08:07 +00:00
d3cbaa1ddf
Avoid using DIDescriptor.isNull().
...
llvm-svn: 97976
2010-03-08 20:53:17 +00:00
9cf2705a5d
Revert r97949.
...
llvm-svn: 97964
2010-03-08 19:20:58 +00:00
4ce8ebb087
Avoid DIDescriptor.isNull() checks.
...
llvm-svn: 97949
2010-03-08 18:26:57 +00:00
de4827dd34
Extend ObjCMessageExpr for class method sends with the source location
...
of the class name.
llvm-svn: 97943
2010-03-08 16:40:19 +00:00
c934bc840c
Perform overload resolution when static_cast'ing from a
...
pointer-to-member-to-derived to a pointer-to-member-to-base. Fixes
PR6072.
llvm-svn: 97923
2010-03-07 23:24:59 +00:00
577cf97cf3
Don't turn off mangling in implicitly extern "C" system headers. GCC
...
doesn't do this on any of the major platforms, and we don't really
support any of the platforms that do (nor will we actually handle
those headers well). Fixes PR6217; see PR6530 for details on what we
would need to do to support these platforms.
llvm-svn: 97899
2010-03-07 05:10:40 +00:00
f0a0f68c79
Add a message to these asserts.
...
llvm-svn: 97873
2010-03-06 09:07:19 +00:00
d51c9b9a9f
Use static method in GlobalValue
...
llvm-svn: 97872
2010-03-06 07:35:18 +00:00
99d20f83ba
PR6515: Implement __builtin_signbit and friends.
...
I'm reasonably sure my implementation is correct, but it would be nice if
someone could double-check.
llvm-svn: 97864
2010-03-06 02:17:52 +00:00
beec5a080f
Implement __builtin_dwarf_sp_column for i386 (Darwin and not), x86-64 (all),
...
and ARM. Implement __builtin_init_dwarf_reg_size_table for i386 (both) and
x86-64 (all).
llvm-svn: 97859
2010-03-06 00:35:14 +00:00
f27ab71d9f
add support for a 1<<29 bit in the block flags field to mark blocks using alternate struct return ABI
...
llvm-svn: 97775
2010-03-05 01:29:59 +00:00
4c489c7030
Don't produce an alias for a destructor if the target is weak.
...
This fixes bootstrap on ELF systems :-)
llvm-svn: 97773
2010-03-05 01:21:10 +00:00
bef98689ad
really fix 6473 by handling weakref in constant expressions.
...
llvm-svn: 97750
2010-03-04 21:26:03 +00:00
2e42fec3a0
Fix PR6473.
...
Clang's support for weakref is now better than llvm-gcc's :-)
We don't introduce a new symbol and we correctly mark undefined references weak only if there is no
definition or regular undefined references in the same file.
llvm-svn: 97733
2010-03-04 18:17:24 +00:00
731be6620c
Revert changes r97693, r97700, and r97718.
...
Our testing framework can't deal with disabled targets yet.
llvm-svn: 97719
2010-03-04 04:29:44 +00:00
9eba9dfb0c
Refactor local class name mangling and make it
...
ABI conforming.
llvm-svn: 97702
2010-03-04 01:02:03 +00:00
81d4d12504
Implement __builtin_dwarf_sp_column().
...
llvm-svn: 97700
2010-03-04 00:44:01 +00:00
6abe39c61e
Create a TargetMachine whenever we create a CodeGenAction. The codegen of
...
some builtins will rely on target knowledge.
llvm-svn: 97693
2010-03-04 00:23:29 +00:00
281aae6328
Fix code gen bug generating code for
...
((id)cat)->isa. Fixes radar 7709015.
llvm-svn: 97672
2010-03-03 22:09:47 +00:00
576def7fbe
fix PR6475, we were doing side-effecting stuff in an assert.
...
llvm-svn: 97669
2010-03-03 21:52:23 +00:00
a529c25555
Implements mangling of local class names to
...
fix a code gen crash. This is WIP as not
all ABI cases are covered (there is a FIXME to
this effect). Fixes radar 7696748.
llvm-svn: 97658
2010-03-03 19:41:08 +00:00
5cc15e058b
add framework for ARM builtins, Patch by Edmund Grimley Evans!
...
llvm-svn: 97656
2010-03-03 19:03:45 +00:00
515c3c548c
Sketch out an implementation for __builtin_dwarf_cfa. I have no idea
...
why the front-end is calculating the argument to llvm.eh.dwarf.cfa().
llvm-svn: 97653
2010-03-03 10:30:05 +00:00
66769f8544
Implement __builtin_eh_return.
...
llvm-svn: 97643
2010-03-03 05:38:58 +00:00
5b0057c05b
Fix a bug with base offset merging that Devang noticed.
...
llvm-svn: 97641
2010-03-03 04:58:02 +00:00
d4f4b7f5ee
Add proper target hooks for __builtin_extract_return_address and
...
__builtin_frob_return_address. The implementations for both are
still trivial in the default case.
llvm-svn: 97638
2010-03-03 04:15:11 +00:00
1950a11939
Don't emit derived-to-base destructor aliases if we don't have a definition
...
for the base destructor, because aliases to declarations aren't legal.
Fixes PR 6471.
llvm-svn: 97637
2010-03-03 03:40:11 +00:00
70e040d552
During codegen assert that any copy assignment, destructor or constructor that
...
we need to synthesize has been marked as used by Sema.
Change Sema to avoid these asserts.
llvm-svn: 97589
2010-03-02 21:28:26 +00:00
93b1b49082
Simplify code.
...
llvm-svn: 97551
2010-03-02 05:40:45 +00:00
b6cc2c0439
Inspired by seeing "MIPS" go by in the commits, I've gone ahead and
...
implemented a (codegen) target hook for __builtin_extend_pointer.
I'm also making it return a uint64_t instead of an unsigned word; this
comports with typical usage (i.e. the one use I know of).
I don't know if any of the existing targets requires this hook to be
set (other than x86 and x86_64, which I know do not).
llvm-svn: 97547
2010-03-02 03:50:12 +00:00
6f1f002651
Rename BaseOffset to Offset and make it signed in preparation of more construction vtable work.
...
llvm-svn: 97546
2010-03-02 03:44:06 +00:00
4b613fae35
After much consultation aimed at figuring out what this builtin actually
...
does, document the results and then implement __builtin_extend_pointer for
platforms where it's a no-op.
llvm-svn: 97540
2010-03-02 02:31:24 +00:00
bd8d9bd39c
Split out types that are non-canonical unless dependent as their own
...
category. Use this in a few places to eliminate unnecessary TST cases and
do some future-proofing. Provide terrible manglings for typeof. Mangle
decltype with some hope of accuracy.
Our manglings for some of the cases covered in the testcase are different
from gcc's, which I've raised as an issue with the ABI list.
llvm-svn: 97523
2010-03-01 23:49:17 +00:00
de9607bc6f
The latest draft uses 'dt' to mangle member expressions, and now so do we.
...
llvm-svn: 97479
2010-03-01 19:12:25 +00:00
a51c67424e
Improve name mangling for dependently-scoped declaration references.
...
llvm-svn: 97422
2010-02-28 22:05:49 +00:00
5610db6923
Add name mangling for DeclRefExprs that refer to external names
...
llvm-svn: 97418
2010-02-28 21:40:32 +00:00
5b79c31471
Remove debug output.
...
llvm-svn: 97407
2010-02-28 18:39:30 +00:00
7162ab947b
Handle unused functions in construction vtables correctly.
...
llvm-svn: 97406
2010-02-28 18:37:33 +00:00
8b37bb7c48
When laying out vtables for virtual bases in construction vtables, we need to check if the vtable is a primary base in the layout class.
...
llvm-svn: 97402
2010-02-28 18:08:38 +00:00
e3385f566b
More improvements to construction vtables; we know handle vbase offsets correctly (I hope).
...
llvm-svn: 97361
2010-02-28 01:43:58 +00:00
62c6c72bab
Pass information about whether a base is virtual or not down to getCtorVtable, we need this information in the vtable builder.
...
llvm-svn: 97356
2010-02-28 00:36:23 +00:00
31f595f5ef
Add new function.
...
llvm-svn: 97353
2010-02-28 00:10:58 +00:00
f98b47d3b9
Fix to dumpLayout; we want to be able to dump address points even if the vtable doesn't have any methods.
...
llvm-svn: 97350
2010-02-27 23:58:01 +00:00
c3f92a16f2
Add a simple construction vtable test.
...
llvm-svn: 97344
2010-02-27 21:09:00 +00:00
9391375d95
Start fleshing out construction vtable support.
...
llvm-svn: 97342
2010-02-27 20:39:05 +00:00
603d7d1910
Enable the new vtable layout code for vtables that aren't construction vtables. (This doesn't mean that we emit LLVM IR using it yet, it just means that it's running and hopefully not crashing or asserting).
...
llvm-svn: 97341
2010-02-27 20:02:53 +00:00
9f19aaaa3a
Move ComputeThisAdjustmentBaseOffset to VtableBuilder.
...
llvm-svn: 97340
2010-02-27 19:57:44 +00:00
b26c2abac7
Make sure to insert the primary base in the set :)
...
llvm-svn: 97339
2010-02-27 19:51:04 +00:00
e02a926a1c
Use the real base offset when calculating vbase offsets.
...
llvm-svn: 97338
2010-02-27 19:21:58 +00:00
4d0729a7cc
Figured out why the test was failing, this will hopefully fix it.
...
llvm-svn: 97336
2010-02-27 19:00:53 +00:00
2b381bbb5c
Don't add this adjustments for pure virtual member functions.
...
llvm-svn: 97334
2010-02-27 18:16:50 +00:00
01d3c7da60
We want to store method info for unused functions.
...
llvm-svn: 97333
2010-02-27 18:09:40 +00:00
4068798cbc
Finish up the changes to this adjustments.
...
llvm-svn: 97328
2010-02-27 16:52:49 +00:00
d46ed89d79
Stub out more of the 'this' pointer adjustment fixes I've been planning. Start using a set vector for primary bases so they will be ordered.
...
llvm-svn: 97326
2010-02-27 16:18:19 +00:00
d2025417ef
Fix another vtable layout bug; we weren't looking hard enough for overriden functions when determining if an overrider will ever be used.
...
llvm-svn: 97306
2010-02-27 06:38:03 +00:00
d69b2f93ea
Handle vcall offset sharing between destructors.
...
llvm-svn: 97304
2010-02-27 04:12:52 +00:00
b08aaa3e10
Fix a bug where we were generating an unnecessary vtable for a virtual base that's already a primary virtual base.
...
llvm-svn: 97303
2010-02-27 04:05:52 +00:00
43c6ffbeec
Sundry fixes to the new vtable builder.
...
llvm-svn: 97258
2010-02-26 20:22:44 +00:00
36c6320d78
Don't generate method metadata for @dynamic properties. Fixes PR6354.
...
llvm-svn: 97199
2010-02-26 01:11:38 +00:00
2da83a3a38
Use the power of types to track down another canonicalization bug in
...
the ABI-computation interface. Fixes <rdar://problem/7691046>.
llvm-svn: 97197
2010-02-26 00:48:12 +00:00
9e3b3a3bbe
Improve vcall offset handling.
...
llvm-svn: 97174
2010-02-25 22:23:13 +00:00
ed7d0e8be8
Fux a bug where we were trying to add overriders for non-virtual bases of virtual bases more than once.
...
llvm-svn: 97173
2010-02-25 22:18:35 +00:00
9ac3092d34
Targets (like pic16) may have mangled the name of global variables,
...
so get the name from Var rather than the original decl.
llvm-svn: 97114
2010-02-25 05:20:44 +00:00
01bbef516e
Remove dead code.
...
llvm-svn: 97109
2010-02-25 03:57:50 +00:00
422f778ba1
Move the vcall and vbase offset layout code out into its own class.
...
llvm-svn: 97107
2010-02-25 03:45:56 +00:00
3c1ec97a28
Get rid of 'this' adjustments from the FinalOverriders class since they can be different for the same overrider in different parts of the vtable.
...
llvm-svn: 97068
2010-02-24 22:32:18 +00:00
d8d698da60
Improve this adjustment pointer calculation.
...
llvm-svn: 97067
2010-02-24 22:27:12 +00:00
8025e7b1b9
Make ComputeThisAdjustmentBaseOffset public for now.
...
llvm-svn: 97066
2010-02-24 22:18:01 +00:00
aea181de04
Fix an iterator-invalidation bug that was causing selfhost errors
...
on non-darwin platforms. Fixes PR6411. Test case doesn't reduce,
unfortunately.
llvm-svn: 97055
2010-02-24 20:32:01 +00:00
a9f633b060
Generate correct vcall offsets when we have a primary virtual base that is not a primary base in the complete class hierarchy.
...
llvm-svn: 97039
2010-02-24 16:43:12 +00:00
8ee376f08a
Canonicalize parameter and return types before computing ABI info. Eliminates
...
a common source of oddities and, in theory, removes some redundant ABI
computations. Also fixes a miscompile I introduced yesterday by refactoring
some code and causing a slightly different code path to be taken that
didn't perform *parameter* type canonicalization, just normal type
canonicalization; this in turn caused a bit of ABI code to misfire because
it was looking for 'double' or 'float' but received 'const float'.
llvm-svn: 97030
2010-02-24 07:14:12 +00:00
b407338fe2
Emit debug info for VectorType.
...
llvm-svn: 96999
2010-02-23 22:59:39 +00:00
c18086ae17
Add support for the weakref attribute. We still produce "alias weak" as llvm-gcc does, but are more strict on what uses of weakref we accept.
...
llvm-svn: 96992
2010-02-23 22:00:30 +00:00
fc83aa04db
Unconditionally support block introspection data in a new field at the end
...
of the block descriptor field. This field is the ObjC style @encode
signature of the implementation function, and was to this point
conditionally provided in the block literal data structure. That
provisional support is removed.
Additionally, eliminate unused enumerations for the block literal flags field.
The first shipping ABI unconditionally set (1<<29) but this bit is unused
by the runtime, so the second ABI will unconditionally have (1<<30) set so
that the runtime can in fact distinguish whether the additional data is
present or not.
llvm-svn: 96989
2010-02-23 21:51:17 +00:00
1e6bc133ce
Make previous fix handle a few more edge cases.
...
llvm-svn: 96962
2010-02-23 18:25:09 +00:00
ad5f0789f6
PR6400: Handle an extreme edge case in mangling correctly.
...
llvm-svn: 96961
2010-02-23 18:20:18 +00:00
d2c0de6b44
PR6386: Fix a recent regression in IRGen of cast-to-union constructs.
...
llvm-svn: 96958
2010-02-23 17:58:35 +00:00
8939e400b6
Implement IsOverriderUsed. This can't be tested yet due to some other bugs :)
...
llvm-svn: 96897
2010-02-23 06:34:44 +00:00
a8796c692c
Stub out IsOverriderUsed.
...
llvm-svn: 96883
2010-02-23 04:26:39 +00:00
0570f71255
More fixes. Don't try to emit a virtual base vtable if the virtual base in question is a primary virtual base of some other base.
...
llvm-svn: 96881
2010-02-23 03:48:14 +00:00
bae46277fe
Always emit vcall offset for the primary base, not only if it's virtual. Remove a debug printf, and add the test case that now passes.
...
llvm-svn: 96880
2010-02-23 03:26:17 +00:00
510d22d097
Simplify the vcall offset calculation and make it give the correct answers :) My test case now has the right values but in the wrong order.
...
llvm-svn: 96877
2010-02-23 03:14:49 +00:00
30f5301e54
More work on vcall offsets. We now emit the right number of vcall offsets in my local test case, but not the right values.
...
llvm-svn: 96874
2010-02-23 02:47:31 +00:00
075056a8f9
Move BaseOffset out of the FinalOverriders class.
...
llvm-svn: 96853
2010-02-23 01:34:28 +00:00
f8ff7b9fd1
Perform two more constructor/destructor code-size optimizations:
...
1) emit base destructors as aliases to their unique base class destructors
under some careful conditions. This is enabled for the same targets that can
support complete-to-base aliases, i.e. not darwin.
2) Emit non-variadic complete constructors for classes with no virtual bases
as calls to the base constructor. This is enabled on all targets and in
theory can trigger in situations that the alias optimization can't (mostly
involving virtual bases, mostly not yet supported).
These are bundled together because I didn't think it worthwhile to split them,
not because they really need to be.
llvm-svn: 96842
2010-02-23 00:48:20 +00:00
f4f2e0247f
Change the name of the vtable-debugging environment variable to
...
CLANG_VTABLE_DEBUG.
llvm-svn: 96785
2010-02-22 16:48:26 +00:00
76eb74a1cd
Don't use NamedDecl::getNameAsCString() when dealing with C++ methods,
...
since they may not have normal identifiers for names. Fixes PR6369.
llvm-svn: 96784
2010-02-22 16:44:27 +00:00
36c569fb33
Eliminate the default arguments to ASTContext::getFunctionType(),
...
fixing up a few callers that thought they were propagating NoReturn
information but were in fact saying something about exception
specifications.
llvm-svn: 96766
2010-02-21 22:15:06 +00:00
350b56eed5
Add the CK_UnusedFunctionPointer component kind for unused function pointers.
...
llvm-svn: 96695
2010-02-19 20:08:13 +00:00
b81884d347
More refactoring around constructor/destructor code generation.
...
Fix some bugs with function-try-blocks and simplify normal try-block
code generation.
This implementation excludes a deleting destructor's call to
operator delete() from the function-try-block, which I believe
is correct but which I can't find straightforward support for at
a moment's glance.
llvm-svn: 96670
2010-02-19 09:25:03 +00:00
2d3d903872
Dump this-adjustments for destructors as well.
...
llvm-svn: 96660
2010-02-19 06:03:53 +00:00
d4324148d7
Re-introduce the ctor/dtor alias optimization, this time hidden behind a
...
command-line option which defaults off.
llvm-svn: 96649
2010-02-19 01:32:20 +00:00
aef662206b
Patch removes IVars list from ObjCInterfaceDecl and
...
instead relies on their DeclContext for iteration, etc.
llvm-svn: 96638
2010-02-19 00:31:17 +00:00
5421a83e4a
remove unused function & enumerations
...
llvm-svn: 96635
2010-02-19 00:24:37 +00:00
334ce7c1c4
Revert the ctor/dtor alias optimization for now; the buildbots can detect
...
some failure here that I can't.
llvm-svn: 96612
2010-02-18 21:31:48 +00:00
5c60a6f597
Make deleting and complete dtor variants defer to other dtor variants by
...
calling them as subroutines. This triggers whenever the alias optimization
doesn't, i.e. when the dtor has linkonce linkage or there are virtual bases
or it's the deleting dtor.
llvm-svn: 96605
2010-02-18 19:59:28 +00:00
f8b53536ed
Attempt to fix the 32-bit test failures.
...
llvm-svn: 96595
2010-02-18 18:20:49 +00:00
e5faca7cda
Fix another bug and add another class.
...
llvm-svn: 96590
2010-02-18 17:32:33 +00:00
8847d9e40a
More work on vcall offsets.
...
llvm-svn: 96587
2010-02-18 17:26:40 +00:00
78277c7705
Start stubbing out vcall offset handling.
...
llvm-svn: 96585
2010-02-18 17:00:09 +00:00
5a7654a9d8
Store the base offset of the final overrider in the OverriderInfo struct, to be used for vcall offsets.
...
llvm-svn: 96582
2010-02-18 16:29:24 +00:00
c8316a2e08
Remove some dead code.
...
llvm-svn: 96581
2010-02-18 16:24:20 +00:00
89b12b34f6
Extract out function-body code generation into its own method. No functionality
...
change.
llvm-svn: 96564
2010-02-18 03:17:58 +00:00
d2408f78a5
silence warning in a cleaner way
...
llvm-svn: 96520
2010-02-17 21:42:34 +00:00
999ddbcf8e
Patch to remove arbitrary imporation of 'self' into
...
a block without it being used. This causes over release
of objects in certain runtime-senitive apps.
(fixes radar 7581175).
llvm-svn: 96501
2010-02-17 19:54:08 +00:00
586ca163b4
Did not intend to check this in.
...
llvm-svn: 96486
2010-02-17 18:11:55 +00:00
c0309cd3d1
Use proper lexcial context for newly added ivars.
...
llvm-svn: 96484
2010-02-17 18:10:54 +00:00
231461f88f
Make FinalOverriders handle virtual bases correctly. Unfortunately this can't be tested just yet.
...
llvm-svn: 96481
2010-02-17 17:48:25 +00:00
d2fd302726
Handle primary bases in AddVCallOffsets.
...
llvm-svn: 96461
2010-02-17 06:30:02 +00:00
c20a9586c4
Silence GCC warning by marking an assert-only variable as unused.
...
llvm-svn: 96459
2010-02-17 06:24:18 +00:00
b0a588f038
First cut at emitting vcall offsets.
...
llvm-svn: 96455
2010-02-17 06:07:19 +00:00
67cea74745
Emit complete constructors and destructors as aliases to base constructors
...
and destructors when the two entities are semantically identical, i.e. when
the class has no virtual base classes. We only do this for linkage types
for which aliases are supported, i.e. internal and external, i.e. not linkonce.
llvm-svn: 96451
2010-02-17 03:52:49 +00:00
db042200c0
Pass through whether a base is virtual or not.
...
llvm-svn: 96449
2010-02-17 03:11:55 +00:00
ba1e267a4c
implement EmitVAArg. pretty much the same way other targets do.
...
llvm-svn: 96446
2010-02-17 02:25:52 +00:00
03db9e9cd3
Correctly calculate base offsets for 'this' pointer adjustments involving virtual bases. This can't be tested yet due to lack of vcall offsets :)
...
llvm-svn: 96441
2010-02-17 02:05:23 +00:00
347132b32f
IRgen optimization: cache the value of 'this' and 'vtt' instead of
...
repeatedly reloading from an alloca. We still need to create the alloca
for debug info purposes (although we currently create it in all cases
because of some abstraction boundaries that're hard to break down).
llvm-svn: 96403
2010-02-16 22:04:33 +00:00
7585580ccc
Distinguish two lexical blocks at the same level.
...
llvm-svn: 96397
2010-02-16 21:41:20 +00:00
fd09df7839
IRgen: Switch 'retval' to use CreateIRTemp.
...
llvm-svn: 96376
2010-02-16 19:45:20 +00:00
d004918ccb
IRgen: Add CreateIRTemp, which creates a temporary alloca but with type converted "not-for-memory". Dunno a better name.
...
llvm-svn: 96374
2010-02-16 19:44:13 +00:00
27bacafb71
IRgen: Switch EmitCompoundLiteralLValue to use CreateMemTemp.
...
llvm-svn: 96373
2010-02-16 19:43:39 +00:00
13f0f66223
Handle layout of vtables for virtual bases.
...
llvm-svn: 96355
2010-02-16 16:49:35 +00:00
b716a543c9
Move some code around in preparation for virtual base vtables.
...
llvm-svn: 96354
2010-02-16 16:26:28 +00:00
14157939ab
Fix a bug where we would not emit secondary vtables for bases of a primary base.
...
llvm-svn: 96351
2010-02-16 16:02:57 +00:00
77904f1d5b
Emit vbase offsets.
...
llvm-svn: 96329
2010-02-16 04:59:55 +00:00
dc2dba306c
More work on the new layout code.
...
llvm-svn: 96328
2010-02-16 04:49:44 +00:00
6ce747220a
When emitting complete destructors for classes with virtual bases, compute
...
the offset to the virtual bases statically inside of relying on the virtual
base offsets in the object's vtable(s). This is both more efficient and
sound against the destructor's manipulation of the vtables.
Also extract a few helper routines.
Oh and we seem to pass all tests with an optimized clang now.
llvm-svn: 96327
2010-02-16 04:15:37 +00:00
e8814ce0a6
Use getLocStart(), instead of getLocEnd(), to record starting location of objc method. :)
...
llvm-svn: 96245
2010-02-15 18:08:38 +00:00
998f9d975b
Uniformize the names of type predicates: rather than having isFloatTy and
...
isInteger, we now have isFloatTy and isIntegerTy. Requested by Chris!
llvm-svn: 96224
2010-02-15 16:14:01 +00:00
7538eec67b
When emitting an aggregate into a temporary, make sure we set the alignment
...
on the alloca.
The fact that codegen makes this class of bug so wonderfully easy to make is
embarrassing.
llvm-svn: 96204
2010-02-15 01:23:36 +00:00
033d48697f
Don't compute final overriders or build vtables for bases that don't need a vtable.
...
llvm-svn: 96171
2010-02-14 17:05:59 +00:00
073a0c88cc
Build fix.
...
llvm-svn: 96140
2010-02-14 00:44:19 +00:00
6a7e6a4c3c
Baby steps towards teaching FinalOverriders about virtual bases.
...
llvm-svn: 96139
2010-02-14 00:37:35 +00:00
1888b44988
Don't try to layout construction vtables for now.
...
llvm-svn: 96138
2010-02-14 00:22:59 +00:00
258a1e35e8
Improve support for non-virtual 'this' pointer adjustments. With this, it should be possible to use the new vtable layout code for all class hierarchies that do not involve virtual bases.
...
llvm-svn: 96137
2010-02-14 00:16:19 +00:00
d3adb0ced9
Add basic support for simple non-virtual 'this' pointer adjustments.
...
llvm-svn: 96136
2010-02-13 23:40:17 +00:00
cce9ee8970
Add support for very simple non-virtual this adjustments in the FinalOverriders class.
...
llvm-svn: 96133
2010-02-13 23:17:31 +00:00
c809cbdc3f
Cleanup; remove some duplicated code.
...
llvm-svn: 96131
2010-02-13 22:39:18 +00:00
a678aa1e51
Assert if we encounter this adjustments.
...
llvm-svn: 96125
2010-02-13 22:23:31 +00:00
e75aaa97b3
Start laying out secondary vtables.
...
llvm-svn: 96123
2010-02-13 22:05:23 +00:00
3f9917de6e
Merge base offsets and dump them.
...
llvm-svn: 96121
2010-02-13 21:33:22 +00:00
c1290adab7
Don't make return adjustments for pure virtual member functions.
...
llvm-svn: 96120
2010-02-13 21:16:54 +00:00
cf5a882da1
Handle virtual bases in return adjustment types.
...
llvm-svn: 96119
2010-02-13 21:07:32 +00:00
f3f67846c8
Handle virtual bases in ComputeBaseOffset.
...
llvm-svn: 96117
2010-02-13 20:41:15 +00:00
6a6cbfc55f
More work on covariant return types. We now handle non-virtual adjustments fine.
...
llvm-svn: 96114
2010-02-13 20:11:51 +00:00
4ea31ab369
Emit the 'alignstack' LLVM function attribute when we encounter a function
...
marked 'force_align_arg_pointer'. Almost there; now all I need to do is finish
up the backend.
llvm-svn: 96100
2010-02-13 15:54:06 +00:00
a932bbca75
Fix think-o, attributes can't come *within* the type of the variable.
...
llvm-svn: 96099
2010-02-13 10:42:55 +00:00
94eab4ad04
Silence unused variable warning in a build without assertions.
...
llvm-svn: 96098
2010-02-13 10:38:52 +00:00
e047e11189
Remove dead {include, semicolon, variable}.
...
llvm-svn: 96087
2010-02-13 09:15:07 +00:00
02c673781d
Use a different name for this iterator. MSVC and clang++ didn't like "I" in the same scope twice.
...
llvm-svn: 96086
2010-02-13 09:11:28 +00:00
eee53d9c06
More work on return type adjustments in the new vtable builder.
...
llvm-svn: 96070
2010-02-13 02:02:03 +00:00
ef3057c419
Switch the standard DeclarationName comparator to be a tri-valued comparator.
...
Use that while fixing a nasty misuse of qsort in vtable codegen which, somehow,
has not actually caused a crash.
llvm-svn: 96062
2010-02-13 01:04:05 +00:00
18c28f6cdc
if-0 out printf.
...
llvm-svn: 96052
2010-02-13 00:03:21 +00:00
0760481f26
Fix a refacto that broke the clang-on-clang build.
...
llvm-svn: 95994
2010-02-12 18:14:46 +00:00
5272c2574a
Start stubbing out more of the covariant thunk support.
...
llvm-svn: 95990
2010-02-12 17:37:14 +00:00
8e661e15f2
Keep track of whether a final overrider needs a return type adjustment.
...
llvm-svn: 95985
2010-02-12 17:13:23 +00:00
136bd19224
Move overrider out into a separate struct.
...
llvm-svn: 95984
2010-02-12 16:55:34 +00:00
e8a9756e63
Keep track of the address points for all primary bases, and add the ability to dump multiple address points for a single offset.
...
llvm-svn: 95970
2010-02-12 07:43:48 +00:00
84c416b9f6
Fix a bug causing an assertion when a covariant return type differed from
...
an overriden type only by reduced qualification.
llvm-svn: 95968
2010-02-12 06:15:07 +00:00
65b4978f7f
More work on vtable layout. We can now layout vtables with primary bases.
...
llvm-svn: 95965
2010-02-12 05:25:12 +00:00
09da3372b6
When dumping vtables, dump whether a virtual member function is pure or not.
...
llvm-svn: 95963
2010-02-12 02:38:13 +00:00
9fbf6d95f6
More work on the final overriders.
...
llvm-svn: 95954
2010-02-12 01:40:03 +00:00
f4df65cf60
Do not ignore anonymous records.
...
llvm-svn: 95953
2010-02-12 01:31:06 +00:00
e003dd1721
Cache new compile unit.
...
llvm-svn: 95952
2010-02-12 01:30:31 +00:00
e8ac42ad5c
Stub out the final overriders class.
...
llvm-svn: 95922
2010-02-11 21:24:32 +00:00
dd27b5d7bb
Use getAs instead of cast/dyn_cast on types. (I'm sure Doug will appreciate this).
...
llvm-svn: 95911
2010-02-11 19:45:15 +00:00
727ffb14a1
Move two functions to the top. No functionality change.
...
llvm-svn: 95908
2010-02-11 19:39:49 +00:00
5bd8d19291
More vtable layout dumper improvements. Handle destructors, dump the complete function type of the member functions (using PredefinedExpr::ComputeName.
...
llvm-svn: 95887
2010-02-11 18:20:28 +00:00
932c2f2a9a
Keep track of, and dump, vtable address points.
...
llvm-svn: 95874
2010-02-11 17:18:51 +00:00
5d40c6f658
Check in the beginnings of my new vtable layout builder idea.
...
Right now, it's off by default but can be tested by passing -fdump-vtable-layouts to clang -cc1. This option will cause all vtables that will normally be emitted as part of codegen to also be dumped using the new layout code.
I've also added a very simple new vtable layout test case.
llvm-svn: 95865
2010-02-11 08:02:13 +00:00
7f4945aa9c
Remove use of 'std::string' from Attr objects, using instead a byte
...
array allocated using the allocator in ASTContext. This addresses
these strings getting leaked when using a BumpPtrAllocator (in
ASTContext).
Fixes: <rdar://problem/7636765>
llvm-svn: 95853
2010-02-11 05:28:37 +00:00
d505812422
Eliminate a bunch of unnecessary ASTContexts from members functions of
...
Decl subclasses. No functionality change.
llvm-svn: 95841
2010-02-11 01:19:42 +00:00
0a5a2216e2
Eliminate the ASTContext parameter from RecordDecl::getDefinition()
...
and CXXRecordDecl::getDefinition(); it's totally unnecessary. No
functionality change.
llvm-svn: 95836
2010-02-11 01:04:33 +00:00
989908fbac
Generate the objc_read_weak API when calling
...
a __weak block. Fixes radar 7628591.
llvm-svn: 95822
2010-02-10 23:34:57 +00:00
c55042c20f
hopefully silence a warning on the clang-i386-darwin9 tester.
...
llvm-svn: 95812
2010-02-10 21:41:43 +00:00
3241d400c8
Switch to using -fsjlj-exceptions instead of hard-coding it. Notably, this fixes
...
calls to the UnwindResumeOrRethrow function for C++/Obj-C exception handling,
for Darwin ARM.
llvm-svn: 95787
2010-02-10 18:49:11 +00:00
535fdaf020
Refactor code that generates debug info for variables that has BlocksAttr.
...
llvm-svn: 95786
2010-02-10 18:49:08 +00:00
1b5330afe9
Use current location as the location of compiler generated arguments, e.g. self, _cmd etc.
...
llvm-svn: 95743
2010-02-10 01:09:50 +00:00
f3a499ad88
StringRefize two random methods, remove a dead variable and a weird constructor call.
...
This maybe fixes some of the errors GCC 4.5 gives on this code.
llvm-svn: 95684
2010-02-09 19:31:24 +00:00
0ae70d16b8
Fix virtual bases' debug info.
...
llvm-svn: 95678
2010-02-09 19:09:28 +00:00
cce7e85165
Mark implicit "this" argument as an artificial argument.
...
llvm-svn: 95673
2010-02-09 17:57:50 +00:00
a7566f163a
IRgen: Add CreateMemTemp, for creating an temporary memory object for a particular type, and flood fill. - CreateMemTemp sets the alignment on the alloca correctly, which fixes a great many places in IRgen where we were doing the wrong thing.
...
- This fixes many many more places than the test case, but my feeling is we need to audit alignment systematically so I'm not inclined to try hard to test the individual fixes in this patch. If this bothers you, patches welcome!
PR6240.
llvm-svn: 95648
2010-02-09 02:48:28 +00:00
819e54b65c
Pass inline keyword to optimizer as the new InlineHint function attribute.
...
At the moment the inlinehint attribute is ignored by the Inliner unless you
pass a -respect-inlinehint option. This will soon be the default.
The inlinehint attribute is set if the inline keyword is explicitly specified
on any declaration.
llvm-svn: 95623
2010-02-09 00:10:00 +00:00
3d33fab7fc
Reapply r95393, without the change to CGExpr. I was wrong in assuming that the
...
element type always matched the converted LLVM type for ExprType.
llvm-svn: 95596
2010-02-08 22:53:07 +00:00
340aafaac6
Emit global references with constant initializers as constants. Fixes PR5585.
...
The standard actually says that such references should have internal linkage,
but gcc doesn't do that, so we probably can't get away with it.
llvm-svn: 95577
2010-02-08 21:46:50 +00:00
ff8cce4395
Make sure to set vtable pointers in the destructors as well.
...
llvm-svn: 95525
2010-02-07 19:45:40 +00:00
6d0e5bd8f3
Use the right type when taking the address of a non-virtual member function pointer. Fixes PR6258.
...
llvm-svn: 95524
2010-02-07 17:37:13 +00:00
cee2d2f016
Use the right linkage for static variables inside C++ inline functions.
...
llvm-svn: 95512
2010-02-07 02:03:08 +00:00
10d369d1a2
Improved handling of the visibility attribute. Declarations now inherit their parent's visibility.
...
(This is kind of a risky change, but I did a self-host build and everything appears to work fine!)
llvm-svn: 95511
2010-02-07 01:44:36 +00:00
9878f9f73b
Make EmitStartEHSpec and EmitEndEHSpec return early when exceptions are disabled.
...
llvm-svn: 95509
2010-02-06 23:59:05 +00:00
fcd764a168
Add support for threadsafe statics, and make them the default (matching gcc).
...
Daniel, I'd appreciate a review of the driver/cc1 parts.
llvm-svn: 95508
2010-02-06 23:23:06 +00:00
5ade5d39b4
Call destructors for constructed bases as well.
...
llvm-svn: 95502
2010-02-06 20:00:21 +00:00
ba63167b64
If a constructor throws an exception we need to execute the destructors for all fully constructed members. Fixes ctor_dtor_count.cpp in the test suite.
...
llvm-svn: 95501
2010-02-06 19:50:17 +00:00
61f6db54e1
Switch CodeGen's "is this variable declaration a definition?" logic
...
over to VarDecl::isThisDeclarationADefinition(), which handles
variables declared with linkage specifications better (among other
things). CMake 2.9 (from CVS) now builds with clang++ and is somewhat
functional.
llvm-svn: 95486
2010-02-06 05:15:45 +00:00
2c4de4c31e
Only append 'L' for internal variable declarations, not all declarations. (Found by the mangle checker, yay)
...
llvm-svn: 95485
2010-02-06 04:52:27 +00:00
9c193965ba
Remove unused variable.
...
llvm-svn: 95476
2010-02-06 03:29:18 +00:00
6710c5351e
Use the correct function info for constructors when applying function attributes. Fixes PR6245.
...
llvm-svn: 95474
2010-02-06 02:44:09 +00:00
cae03b1e6d
Assert when we try to mangle a dependent template name, rather than
...
crashing unceremoniously.
llvm-svn: 95464
2010-02-06 01:09:36 +00:00
ca972cd1f1
Rearrange some checks to avoid call to isCopyConstructor() and clarify path
...
taken for non-trivial constructors.
llvm-svn: 95457
2010-02-06 00:25:16 +00:00
ab26cfa58d
Standardize the parsing of function type attributes in a way that
...
follows (as conservatively as possible) gcc's current behavior: attributes
written on return types that don't apply there are applied to the function
instead, etc. Only parse CC attributes as type attributes, not as decl attributes;
don't accepet noreturn as a decl attribute on ValueDecls, either (it still
needs to apply to other decls, like blocks). Consistently consume CC/noreturn
information throughout codegen; enforce this by removing their default values
in CodeGenTypes::getFunctionInfo().
llvm-svn: 95436
2010-02-05 21:31:56 +00:00
a643780483
Revert r95393, which broke Clang's self-host.
...
llvm-svn: 95430
2010-02-05 21:10:36 +00:00
a16b0cae9a
Implement name mangling for template template parameters
...
llvm-svn: 95427
2010-02-05 20:45:00 +00:00
9c4e4651d5
IRgen: A few more ConvertType cleanups.
...
llvm-svn: 95423
2010-02-05 20:02:42 +00:00
d0bc7b9d13
IRgen: Factor out EmitAggExprToLValue.
...
llvm-svn: 95416
2010-02-05 19:38:31 +00:00
df506b934e
Fix a code gen bug accessing 'isa' field via a message call
...
(Fixes radar 7609722).
llvm-svn: 95406
2010-02-05 19:18:30 +00:00
8848175547
IRgen: Fix some CreateTempAlloca calls to use ConvertTypeForMem when that is
...
conceptually correct. Review appreciated (Chris, Eli, Anders).
llvm-svn: 95401
2010-02-05 18:56:49 +00:00
cb86e1088a
Fix the bug that was breaking self-host, and re-land the static ctor fixes.
...
llvm-svn: 95400
2010-02-05 18:38:45 +00:00
5a5473f0f8
Now that we store calling conventions in the types, use them instead of
...
getting the calling convention from the target function, which may or may not
exist. Fixes PR5280.
llvm-svn: 95399
2010-02-05 18:13:10 +00:00
363589bded
IRgen: Use hasAggregateLLVMType instead of isSingleValueType() for cases that
...
need to deal with aggregates specially; this is consistent with the rest of IRgen.
Also, simplify EmitParmDecl and don't worry about using Decl::getNameAsString.
llvm-svn: 95393
2010-02-05 17:51:33 +00:00
af77cd220a
Revert r95363 and r95375, which broke self-host.
...
llvm-svn: 95389
2010-02-05 16:18:08 +00:00
c6eec40189
Check in a mangle checker that's turned off by default.
...
llvm-svn: 95377
2010-02-05 07:31:37 +00:00
893555ce66
Fix array initialization test.
...
llvm-svn: 95375
2010-02-05 07:00:11 +00:00
ecac409cb0
If a global initializer has a non-trivial destructor it can't be emitted as a constant (even if it has a trivial constructor).
...
llvm-svn: 95363
2010-02-05 05:19:42 +00:00
ae580fede3
Always start tag definitions before completing them. Assert same.
...
Fixes latent and not-so-latent objc++ and blocks++ bugs.
llvm-svn: 95340
2010-02-05 01:33:36 +00:00
67da35c832
Extract a common structure for holding information about the definition
...
of a C++ record. Exposed a lot of problems where various routines were
silently doing The Wrong Thing (or The Acceptable Thing in The Wrong Order)
when presented with a non-definition. Also cuts down on memory usage.
llvm-svn: 95330
2010-02-04 22:26:26 +00:00
824e0613de
When binding an lvalue to a reference, we always need to pop temporaries.
...
With this fix, and the other fixes committed today a make check-all with a clang-built LLVM now gives:
Expected Passes : 6933
Expected Failures : 46
Unsupported Tests : 40
Unexpected Failures: 27
which means that we pass 99.96% of all tests :) The resulting 27 tests are all LLVMC tests and seem to be because of differences in the clang and gcc drivers.
llvm-svn: 95313
2010-02-04 17:32:58 +00:00
9b942c65a5
Fix a bug where we would not mark temporaries as conditional when emitting a conditional operator as an lvalue.
...
llvm-svn: 95311
2010-02-04 17:26:01 +00:00
ae612d22ac
Rename StartConditionalBranch/FinishConditionalBranch to BeginConditionalBranch/EndConditionalBranch.
...
llvm-svn: 95308
2010-02-04 17:18:07 +00:00
c4325486b9
Fix another pointer-to-member function miscompile, this time when trying to call a virtual member function.
...
llvm-svn: 95307
2010-02-04 17:08:48 +00:00
7914dad72d
Calculate offset correctly when taking the address of a virtual member function.
...
llvm-svn: 95305
2010-02-04 16:38:05 +00:00
d061b4429e
Mangle member expressions. Also invented.
...
llvm-svn: 95284
2010-02-04 02:56:29 +00:00
11ef09f8b5
Add a cautionary note about the mangling I just invented.
...
llvm-svn: 95275
2010-02-04 01:48:38 +00:00
09de8ecd88
Add mangling support for calls, sizeof/alignof, constructor calls,
...
float literals, and unresolved lookups (which required hand-wavey extensions).
llvm-svn: 95273
2010-02-04 01:42:13 +00:00
45858d2d59
Revert "Numerous changes to selector handling:", this breaks a whole bunch of
...
working code, for no apparent reason.
llvm-svn: 95244
2010-02-03 20:11:42 +00:00
66498388a7
Handle reference binding in aggregate initializers. Fixes another 47 tests.
...
llvm-svn: 95235
2010-02-03 19:13:55 +00:00
9b92ef90f0
Add a band-aid fix for clang self-hosting. A better fix will follow shortly.
...
llvm-svn: 95232
2010-02-03 18:54:13 +00:00
b247350ef7
More cleanup.
...
llvm-svn: 95226
2010-02-03 17:33:16 +00:00
3b227bd629
Revert the new reference binding code; I came up with a way simpler solution for the reference binding bug that is preventing self-hosting.
...
llvm-svn: 95223
2010-02-03 16:38:03 +00:00
5bb4efdf3c
First pass at adding GC support for GNU runtime. GC ivar maps not yet constructed, GC flag not set.
...
Please don't try using this yet - the runtime support is still very immature and your code will almost certainly crash if you do.
llvm-svn: 95222
2010-02-03 15:59:02 +00:00
7dc5c17d92
When a function or variable somehow depends on a type or declaration
...
that is in an anonymous namespace, give that function or variable
internal linkage.
This change models an oddity of the C++ standard, where names declared
in an anonymous namespace have external linkage but, because anonymous
namespace are really "uniquely-named" namespaces, the names cannot be
referenced from other translation units. That means that they have
external linkage for semantic analysis, but the only sensible
implementation for code generation is to give them internal
linkage. We now model this notion via the UniqueExternalLinkage
linkage type. There are several changes here:
- Extended NamedDecl::getLinkage() to produce UniqueExternalLinkage
when the declaration is in an anonymous namespace.
- Added Type::getLinkage() to determine the linkage of a type, which
is defined as the minimum linkage of the types (when we're dealing
with a compound type that is not a struct/class/union).
- Extended NamedDecl::getLinkage() to consider the linkage of the
template arguments and template parameters of function template
specializations and class template specializations.
- Taught code generation to rely on NamedDecl::getLinkage() when
determining the linkage of variables and functions, also
considering the linkage of the types of those variables and
functions (C++ only). Map UniqueExternalLinkage to internal
linkage, taking out the explicit checks for
isInAnonymousNamespace().
This fixes much of PR5792, which, as discovered by Anders Carlsson, is
actually the reason behind the pass-manager assertion that causes the
majority of clang-on-clang regression test failures. With this fix,
Clang-built-Clang+LLVM passes 88% of its regression tests (up from
67%). The specific numbers are:
LLVM:
Expected Passes : 4006
Expected Failures : 32
Unsupported Tests : 40
Unexpected Failures: 736
Clang:
Expected Passes : 1903
Expected Failures : 14
Unexpected Failures: 75
Overall:
Expected Passes : 5909
Expected Failures : 46
Unsupported Tests : 40
Unexpected Failures: 811
Still to do:
- Improve testing
- Check whether we should allow the presence of types with
InternalLinkage (in addition to UniqueExternalLinkage) given
variables/functions internal linkage in C++, as mentioned in
PR5792.
- Determine how expensive the getLinkage() calls are in practice;
consider caching the result in NamedDecl.
- Assess the feasibility of Chris's idea in comment #1 of PR5792.
llvm-svn: 95216
2010-02-03 09:33:45 +00:00
eb1e610d00
remove a big chunk of #if 0 code.
...
llvm-svn: 95201
2010-02-03 05:28:17 +00:00
db56b919e5
Provide a real fix for PR6199, reverting the old workaround. Here, we
...
realize that CXXConstructExpr is always implicit, so we should just
return its argument (if there is only one) rather than directly
invoking the constructor.
llvm-svn: 95192
2010-02-03 03:01:57 +00:00
92b762e256
Numerous changes to selector handling:
...
- Don't use GlobalAliases with non-0 GEPs (GNU runtime) - this was unsupported and LLVM will be generating errors if you do it soon. This also simplifies the code generated by the GNU runtime a bit.
- Make GetSelector() return a constant (GNU runtime), not a load of a store of a constant.
- Recognise @selector() expressions as valid static initialisers (as GCC does).
- Add methods to GCObjCRuntime to emit selectors as constants (needed for using @selector() expressions as constants. These need implementing for the Mac runtimes - I couldn't figure out how to do this, they seem to require a load.
- Store an ObjCMethodDecl in an ObjCSelectorExpr so that we can get at the type information for the selector. This is needed for generating typed selectors from @selector() expressions (as GCC does). Ideally, this information should be stored in the Selector, but that would be an invasive change. We should eventually add checks for common uses of @selector() expressions. Possibly adding an attribute that can be applied to method args providing the types of a selector so, for example, you'd do something like this:
- (id)performSelector: __attribute__((selector_types(id, SEL, id)))(SEL)
withObject: (id)object;
Then, any @selector() expressions passed to the method will be check to ensure that it conforms to this signature. We do this at run time on the GNU runtime already, but it would be nice to do it at compile time on all runtimes.
- Made @selector() expressions emit type info if available and the runtime supports it.
Someone more familiar with the Mac runtime needs to implement the GetConstantSelector() function in CGObjCMac. This currently just assert()s.
llvm-svn: 95189
2010-02-03 02:09:30 +00:00
2adddcae7e
Remove abstract expression kinds from the StmtClass enum. Update a few users
...
appropriately. Call out a few missing cases in the expression mangler.
llvm-svn: 95176
2010-02-03 00:55:45 +00:00
a71cc15361
Implement promotion for enumeration types.
...
WHAT!?!
It turns out that Type::isPromotableIntegerType() was not considering
enumeration types to be promotable, so we would never do the
promotion despite having properly computed the promotion type when the
enum was defined. Various operations on values of enum type just
"worked" because we could still compute the integer rank of an enum
type; the oddity, however, is that operations such as "add an enum and
an unsigned" would often have an enum result type (!). The bug
actually showed up as a spurious -Wformat diagnostic
(<rdar://problem/7595366>), but in theory it could cause miscompiles.
In this commit:
- Enum types with a promotion type of "int" or "unsigned int" are
promotable.
- Tweaked the computation of promotable types for enums
- For all of the ABIs, treat enum types the same way as their
underlying types (*not* their promotion types) for argument passing
and return values
- Extend the ABI tester with support for enumeration types
llvm-svn: 95117
2010-02-02 20:10:50 +00:00
5dc86337fb
Set the correct vtable pointers _before_ generating code for any member initializers. Fixes about ~2000 clang/LLVM tests in the clang-on-clang build.
...
llvm-svn: 95116
2010-02-02 19:58:43 +00:00
173bfe477b
Use the Arg variable rather than re-computing it. This also silences GCC's
...
unused variable warning.
llvm-svn: 95085
2010-02-02 12:15:55 +00:00
49786a6c31
Codegen CXXConstructExprs with trivial constructors as constants.
...
Eliminates a lot of spurious global initializers, fixing PR6205.
llvm-svn: 95077
2010-02-02 08:02:49 +00:00
75b1f959da
Check in a test case and a nasty workaround for PR6199.
...
llvm-svn: 95076
2010-02-02 07:10:35 +00:00
e8bfe412ec
Improve handling of emitting 'null' pointers to data members.
...
llvm-svn: 95066
2010-02-02 05:17:25 +00:00
259688ce29
Move pointer to data member emission to CodeGenModule and use it in CGExprConstant. Fixes PR5674.
...
llvm-svn: 95063
2010-02-02 03:37:46 +00:00
96ebba5770
ARM/APCS: Fix classification of small complex integer types as "integer like".
...
llvm-svn: 95030
2010-02-01 23:31:26 +00:00
eedf151cb1
ARM/APCS: Pass Complex types following llvm-gcc.
...
llvm-svn: 95029
2010-02-01 23:31:19 +00:00
ab793231c1
Fix FIXME and surrounding comment.
...
llvm-svn: 95023
2010-02-01 22:51:29 +00:00
e8fb4b7fd6
Use appropriate context descriptor in RecordDecl's debug info.
...
llvm-svn: 95016
2010-02-01 22:40:08 +00:00
3efd1470c9
Do not use clang type name to name a local variable, e.g. Decl.
...
llvm-svn: 95010
2010-02-01 21:52:22 +00:00
1c0954cca7
Do not use clang type name to name a local variable, e.g. Decl.
...
llvm-svn: 95009
2010-02-01 21:39:52 +00:00
7b7f46f5a6
Use DeclContext as getContextDescriptor() argument.
...
llvm-svn: 95008
2010-02-01 21:34:11 +00:00
ca5e3eb7f7
NeXT: Add support for -fobjc-legacy-dispatch.
...
llvm-svn: 95005
2010-02-01 21:07:33 +00:00
5b5d2db3f6
Don't explicitly force utf strings into the __TEXT,__ustring
...
by setting the section of the generated global. This is an
optimization done by the code generator, and the code being
removed didn't handle the case when the string contained an
embedded nul (which the code generator does correctly
handle). This is rdar://7589850
llvm-svn: 95003
2010-02-01 20:59:08 +00:00
5ca7984bb4
In C++, an initializer on a variable doesn't necessarily mean it's the definition. With that in mind, rename getDefinition to getAnyInitializer (to distinguish it from getInit) and reimplement it in terms of isThisDeclarationADefinition. Update all code to use this new function.
...
llvm-svn: 94999
2010-02-01 20:16:42 +00:00
973f2eb0cd
Emit debug info for namespaces.
...
llvm-svn: 94991
2010-02-01 19:16:32 +00:00
a682427e42
Switch expressions like T() and T(1,2) over to new-style initialization. I'm
...
not quite sure what we want to do about the AST representation; comments
welcome.
llvm-svn: 94967
2010-01-31 20:58:15 +00:00
604179775a
Simplify EmitMemberInitializer; no intended functionality change.
...
llvm-svn: 94965
2010-01-31 19:07:50 +00:00
ab0ddb57b1
Start creating CXXBindReferenceExpr nodes when binding complex types to references.
...
llvm-svn: 94964
2010-01-31 18:34:51 +00:00
7ae2d7758f
Rework base and member initialization in constructors, with several
...
(necessarily simultaneous) changes:
- CXXBaseOrMemberInitializer now contains only a single initializer
rather than a set of initialiation arguments + a constructor. The
single initializer covers all aspects of initialization, including
constructor calls as necessary but also cleanup of temporaries
created by the initializer (which we never handled
before!).
- Rework + simplify code generation for CXXBaseOrMemberInitializers,
since we can now just emit the initializer as an initializer.
- Switched base and member initialization over to the new
initialization code (InitializationSequence), so that it
- Improved diagnostics for the new initialization code when
initializing bases and members, to match the diagnostics produced
by the previous (special-purpose) code.
- Simplify the representation of type-checked constructor initializers in
templates; instead of keeping the fully-type-checked AST, which is
rather hard to undo at template instantiation time, throw away the
type-checked AST and store the raw expressions in the AST. This
simplifies instantiation, but loses a little but of information in
the AST.
- When type-checking implicit base or member initializers within a
dependent context, don't add the generated initializers into the
AST, because they'll look like they were explicit.
- Record in CXXConstructExpr when the constructor call is to
initialize a base class, so that CodeGen does not have to infer it
from context. This ensures that we call the right kind of
constructor.
There are also a few "opportunity" fixes here that were needed to not
regress, for example:
- Diagnose default-initialization of a const-qualified class that
does not have a user-declared default constructor. We had this
diagnostic specifically for bases and members, but missed it for
variables. That's fixed now.
- When defining the implicit constructors, destructor, and
copy-assignment operator, set the CurContext to that constructor
when we're defining the body.
llvm-svn: 94952
2010-01-31 09:12:51 +00:00
6276b2506d
When performing a derived-to-base cast that we know will not change the offset, we don't need to null check the input pointer. Fixes PR5965.
...
llvm-svn: 94942
2010-01-31 02:39:02 +00:00
600f737b95
When doing a base-to-derived cast we don't need to null check the derived value if the class offset is 0.
...
llvm-svn: 94939
2010-01-31 01:43:37 +00:00
84673e200f
Some class related cleanup.
...
llvm-svn: 94938
2010-01-31 01:36:53 +00:00
087bc13d96
More asm cleanup.
...
llvm-svn: 94920
2010-01-30 20:05:21 +00:00
c7c5baa482
Yay for more StringRefs.
...
llvm-svn: 94917
2010-01-30 19:12:25 +00:00
bb4820dc8e
Use appropriate context for typedefs.
...
llvm-svn: 94849
2010-01-29 22:29:31 +00:00
92e2541599
Maintain a map of regions (lexical scopes) and use it to find context for a global variable.
...
llvm-svn: 94817
2010-01-29 18:11:03 +00:00
16810ca647
Name mangling for cast expressions, from Matthias Schiffer! Fixes PR5876.
...
llvm-svn: 94811
2010-01-29 16:37:09 +00:00
42c876dff2
Use EmitLValueForFieldInitialization when synthesizing the copy ctor as well.
...
llvm-svn: 94800
2010-01-29 05:41:25 +00:00
db78f0ad04
Add a new EmitLValueForFieldInitialization that will be used for initializing fields (and reference type fields in particular).
...
llvm-svn: 94799
2010-01-29 05:24:29 +00:00
5d8645b150
Simplify EmitLValueForField - we can get whether the field is part of a union or not from the FieldDecl (through its DeclContext).
...
llvm-svn: 94798
2010-01-29 05:05:36 +00:00
45c7ff1d79
ARM/APCS ABI: Fix some problems with bit-fields in structures. After rereading
...
the ABI spec, this turns out to simplify the code. We still have some annoying
code which mismatches the spec with regard to empty structures.
llvm-svn: 94796
2010-01-29 03:22:29 +00:00
ba6c437d52
Add an CXXBindReferenceExpr (not used just yet).
...
llvm-svn: 94791
2010-01-29 02:39:32 +00:00
7bfc596fcf
s/CGDebugInfo::getContext/CGDebugInfo::getContextDescriptor/g to avoid confusion.
...
llvm-svn: 94760
2010-01-28 23:15:27 +00:00
128aa9d176
Fix indentation.
...
llvm-svn: 94758
2010-01-28 21:54:15 +00:00
4c3e7e99ca
Static methods do not need "this" pointer argument.
...
llvm-svn: 94756
2010-01-28 21:43:50 +00:00
946edc1aa6
Emit base classes info first, as expected by the debugger.
...
llvm-svn: 94755
2010-01-28 21:41:35 +00:00
b1ef991097
Fix an incorrect union layout assert. Fixes PR6164.
...
llvm-svn: 94754
2010-01-28 18:22:03 +00:00
0d61eebd81
s/FunctionNames/DebugInfoNames/g
...
llvm-svn: 94753
2010-01-28 18:21:00 +00:00
84033fb240
Emit vtable info.
...
llvm-svn: 94751
2010-01-28 18:11:52 +00:00
abb4413728
While emitting debugging infor for a C++ class, identify the holder of class's vtable, if any.
...
llvm-svn: 94712
2010-01-28 00:54:21 +00:00
3d4e6d9dd0
Include "this" pointer argument while emitting debug info for a C++ method.
...
llvm-svn: 94710
2010-01-28 00:28:01 +00:00
160146eba2
Change the return type of ASTContext::getDeclAlignInBytes() to CharUnits and,
...
now that the "InBytes" part of the name is implied by the return type, rename
it to getDeclAlign().
llvm-svn: 94681
2010-01-27 17:10:57 +00:00
358e751a7b
Unique ObjC strings (GNU Runtime); fix for PR6142. Note: Doing this in the runtime-specific code is a bit ugly. It would be a good idea to hoist all of the string / protocol uniqueing code up into CGObjCRuntime or CodeGenModule and only handle emitting the original versions in the runtime-specific code.
...
llvm-svn: 94676
2010-01-27 12:49:23 +00:00
40446e8a0a
Structs and classes with non-trivial destructors or copy constructors should be passed indirectly in the 32-bit ABI. Fixes PR6094.
...
llvm-svn: 94656
2010-01-27 03:25:19 +00:00
ded0a406b4
Refine the non-virtual this adjustments for thunks by using the offset
...
to the declaring class from the nearest virtual base class. WIP.
llvm-svn: 94606
2010-01-26 22:44:01 +00:00
a04ecfb701
Refine the non-virtual this adjustments for thunks by using the offset
...
to the declaring class from the nearest virtual base class. WIP.
This fixes 40% of all the problems remaining in one of my testcases.
llvm-svn: 94592
2010-01-26 21:35:27 +00:00
7df3cbeb01
Convert one last size variable to CharUnits (follow-on to 94577).
...
llvm-svn: 94579
2010-01-26 19:59:28 +00:00
3eb55cfe8a
Use CharUnits for sizes, offsets, alignments, and padding amounts for values
...
that are in character units.
llvm-svn: 94577
2010-01-26 19:44:24 +00:00
5262b11962
Use CharUnits for alignments in character units.
...
llvm-svn: 94571
2010-01-26 19:13:33 +00:00
a0f99ff230
Use CharUnits for alignment of UTF16 string in GetAddrOfConstantCFString().
...
llvm-svn: 94564
2010-01-26 18:46:23 +00:00
e5402e4749
Use CharUnits for alignment in EmitNullInitializationLValue().
...
llvm-svn: 94563
2010-01-26 18:35:11 +00:00
ca4a5459d8
Use the right definition when emitting a global variable. Fixes PR5564.
...
llvm-svn: 94555
2010-01-26 17:43:42 +00:00
7228117108
Change getUnique to return a GlobalDecl. Fixes PR6147.
...
llvm-svn: 94554
2010-01-26 17:36:47 +00:00
98ca79435a
Introduce CodeGenModule::GetTargetTypeStoreSize() to calculate the store size
...
of LLVM types in character units.
llvm-svn: 94542
2010-01-26 13:48:07 +00:00
d6b21e4842
If a global variable has an initializer with side effects, it can never be deferred (even if it's in an anonymous namespace).
...
llvm-svn: 94525
2010-01-26 06:15:16 +00:00
0b5f1ddf0a
Fix another debug info crash with virtual bases.
...
llvm-svn: 94520
2010-01-26 05:26:39 +00:00
17ed0496c5
Fix the test I broke, and also fix a crash when declaring a virtual destructor. Add debug info support for pure virtual member functions.
...
llvm-svn: 94519
2010-01-26 05:19:50 +00:00
b85f0ab976
Factor creating the DISubprogram for a member function out into a separate function.
...
llvm-svn: 94513
2010-01-26 04:49:33 +00:00
c1821155bf
Simplify CGDebugInfo::CollectCXXMemberFunctions a little. More to come.
...
llvm-svn: 94511
2010-01-26 04:40:11 +00:00
20bbbd489b
Make sure to always mark a global variable as not being constant if it has a C++ initializer.
...
llvm-svn: 94504
2010-01-26 04:02:23 +00:00
77537b136e
Be sure to track the non-virtual part of the vcall offset in complex
...
multiple inheritance cases. WIP.
This fixes 20% of the outstanding problems found by the randomized
tester.
llvm-svn: 94499
2010-01-26 03:42:22 +00:00
90181eb294
Fixup a missing vcall entry. WIP.
...
llvm-svn: 94478
2010-01-26 00:05:04 +00:00
c54353d935
First cut at emitting inheritance info.
...
llvm-svn: 94473
2010-01-25 23:32:18 +00:00
7af03fb4f0
Emit debug info for virtual functions.
...
llvm-svn: 94467
2010-01-25 23:17:15 +00:00
5cda92820f
global variable that binds reference to a non-lvalue reproted
...
as NYI now.
llvm-svn: 94453
2010-01-25 21:40:39 +00:00
210031b79c
Fix a code gen. bug involving compiling global references.
...
(fixes radar 7574896).
llvm-svn: 94434
2010-01-25 18:30:26 +00:00
83b0dd16c3
fix rdar://7556129 a crash in blocks debug info codegen.
...
llvm-svn: 94402
2010-01-25 03:34:56 +00:00
bf78478d35
simplify code.
...
llvm-svn: 94401
2010-01-25 03:29:35 +00:00
0bcc858a2b
-fno-rtti is now the default.
...
llvm-svn: 94379
2010-01-24 20:43:31 +00:00
aecc45cb3c
Mangle static variables with an extra name to distinguish them from non-static variables in the same TU.
...
Fixes PR5966 for real this time; also reverts r92911, which had a incorrect fix.
llvm-svn: 94352
2010-01-24 03:04:27 +00:00
60ddba67a8
Fix a nasty bug where temporaries weren't marked as being conditional in some cases.
...
llvm-svn: 94341
2010-01-24 00:20:05 +00:00
d6ef69a7db
Add bzero builtin; this should help codegen quality for code using this
...
function.
llvm-svn: 94320
2010-01-23 19:00:10 +00:00
8f8f405c71
No need to terminate this buffer.
...
llvm-svn: 94313
2010-01-23 18:16:07 +00:00
481e3a87fe
Created __builtin___NSStringMakeConstantString() builtin, which generates constant Objective-C strings.
...
llvm-svn: 94274
2010-01-23 02:40:42 +00:00
1f49d65235
Finish off fixing up debug information.
...
llvm-svn: 94193
2010-01-22 18:48:47 +00:00
0599df16a6
Eliminate some Clang warnings
...
llvm-svn: 94177
2010-01-22 15:41:14 +00:00
d6b28fc6ad
Simplify code.
...
llvm-svn: 94175
2010-01-22 13:59:13 +00:00
9eb76d47f2
Ensure we output all non-virtual base vtables. Fixes PR5890.
...
llvm-svn: 94163
2010-01-22 06:45:05 +00:00
d2808449e4
Update debugging code.
...
llvm-svn: 94145
2010-01-22 02:51:26 +00:00
f768db7775
Move the VTT related code into its own file, CGVTT.cpp
...
llvm-svn: 94079
2010-01-21 16:50:45 +00:00
3f4a77e082
A Decl->getName() is not unique. However, the debug info descriptors
...
are uniqued. The debug info descriptor describing record's context is
necessary to keep two Decl's descriptor unique if their name match.
There is more work to do to create descriptors for DeclContext.
As a temp. step use type's name in FwdDecl.
llvm-svn: 94050
2010-01-20 23:56:40 +00:00
5e1a83c196
Emit human readable names for operators.
...
llvm-svn: 93837
2010-01-19 02:02:11 +00:00
7a12ad0dc1
First cut at emitting debugging information for C++ member functions.
...
There is lot more work to do in this area.
llvm-svn: 93836
2010-01-19 01:54:44 +00:00
757daca99c
Use llvm::Function name as the linkage name, but strip off leading '01' from display name.
...
llvm-svn: 93822
2010-01-19 00:25:12 +00:00
889ce76f41
Refactor.
...
llvm-svn: 93814
2010-01-19 00:00:59 +00:00
ddf6a57f3e
llvm::Function name is the linkage name. If it has a '01' as a prefix then probably there is a reason.
...
llvm-svn: 93808
2010-01-18 23:27:28 +00:00
07894e8a0e
More VTT builder fixes. With these fixes we now correctly handle the very complex VTT example from the Itanium ABI spec.
...
llvm-svn: 93725
2010-01-18 17:13:59 +00:00
3d1a09552c
Move some common code into BuildVTT.
...
llvm-svn: 93710
2010-01-18 04:45:46 +00:00
1af3df854f
Fix a bunch of VTT layout bugs, add simple tests for VTT layout.
...
llvm-svn: 93709
2010-01-18 04:25:18 +00:00
827b245fd6
Minor VTT builder cleanup, no functionality change.
...
llvm-svn: 93696
2010-01-17 23:53:20 +00:00
42f1376f35
Get the ctor vtable address points directly from the VTT builder.
...
llvm-svn: 93681
2010-01-17 17:10:44 +00:00
dbfd16014a
Add codgen for BI__builtin_llvm_memory_barrier.
...
llvm-svn: 93611
2010-01-16 01:21:14 +00:00
8fdc2cb46a
Fix crash generating debug info for constructor for anonymous struct.
...
llvm-svn: 93601
2010-01-16 00:43:13 +00:00
cab014721b
Fix a couple bugs in copy assignment operator synthesis.
...
llvm-svn: 93546
2010-01-15 20:06:11 +00:00
02990837ad
Convert the type of the LValue offset variable in APValue to CharUnits, moving
...
the LValue-related methods of APValue out of line to avoid header file leaching.
llvm-svn: 93512
2010-01-15 12:37:54 +00:00
2d63010a9c
Emit linkage name even if it matches regular name. The code generator uses linkage name to find subprogram entry for the current function.
...
llvm-svn: 93461
2010-01-14 21:46:57 +00:00
ea84f59f5c
Add comment.
...
llvm-svn: 93434
2010-01-14 18:06:13 +00:00
035ead209f
Made ObjC method name mangling match GCC (which does it in a stupid and broken way that can give conflicts on method names containing underscores, but is needed for gdb to work because gdb does not know how to read ObjC class tables and relies on the mangling).
...
llvm-svn: 93427
2010-01-14 14:08:19 +00:00
5f9a881e1e
Store the address points for constructor vtables directly in the VTT builder, because that's the only time they're needed.
...
llvm-svn: 93412
2010-01-14 02:29:07 +00:00
2a4adbea96
Add a DenseMapInfo specialization for BaseSubobject.
...
llvm-svn: 93399
2010-01-14 01:39:42 +00:00
4274a326be
Because CurLoc is the current source location as far as CGDebugInfo is concerned. It is expected that this is set (usually left bracket location of function body compound statement) before EmitfunctionStart() is used.
...
llvm-svn: 93389
2010-01-14 00:48:09 +00:00
934661ed73
Emit human readable names for c/c++ functions. Avoid emitting linkage name if it matches regular name.
...
llvm-svn: 93383
2010-01-14 00:36:21 +00:00
20871481d2
Add a BaseSubobject class to uniquely identify a base class subobject. Not yet used.
...
llvm-svn: 93345
2010-01-13 20:11:15 +00:00
e3e76e2a69
Use the new isInteger() method in a couple places, some random cleanup, and
...
add a fixme.
llvm-svn: 93179
2010-01-11 19:58:10 +00:00
c8fc573ea0
Fix type mismatch on 64-bit platforms (GNU ObjC).
...
llvm-svn: 93169
2010-01-11 19:02:35 +00:00
40775003e6
Roll out ASTContext::getTypeSizeInChars(), replacing instances of
...
"ASTContext::getTypeSize() / 8". Replace [u]int64_t variables with CharUnits
ones as appropriate.
Also rename RawType, fromRaw(), and getRaw() in CharUnits to QuantityType,
fromQuantity(), and getQuantity() for clarity.
llvm-svn: 93153
2010-01-11 17:06:35 +00:00
b322c3a34b
Try to make cmake happy
...
llvm-svn: 93119
2010-01-10 13:06:34 +00:00
55bcea1e65
Generalize target weirdness handling having proper layering in mind:
...
1. Add helper class for sema checks for target attributes
2. Add helper class for codegen of target attributes
As a proof-of-concept - implement msp430's 'interrupt' attribute.
llvm-svn: 93118
2010-01-10 12:58:08 +00:00
89b14a4309
Rename file to generalization in next commits
...
llvm-svn: 93117
2010-01-10 12:57:30 +00:00
bb8976e36e
implement codegen support for preinc as an lvalue, PR5514.
...
llvm-svn: 93076
2010-01-09 21:44:40 +00:00
116ce8f172
refactor pre/postinc logic into CGF and require the caller to pass in the
...
lvalue to poke, no functionality change.
llvm-svn: 93075
2010-01-09 21:40:03 +00:00
34c72dd683
revert r92749, which is just dead code.
...
llvm-svn: 93074
2010-01-09 21:14:47 +00:00
5866fe306d
Fix for PR5967: Make const-marking for LLVM globals correct for cases requiring
...
run-time initialization, and emit run-time initializers aggresively to avoid
ordering issues with deferred globals.
llvm-svn: 92976
2010-01-08 00:50:11 +00:00
df34917a38
Export a public symbol for classes with the GNU runtime.
...
llvm-svn: 92973
2010-01-08 00:14:31 +00:00
950d5b2cd3
fix PR5869: mangle static symbols like gcc does to make it easier to diff symbol tables
...
llvm-svn: 92911
2010-01-07 09:36:51 +00:00
2a34df3264
Fix linkage for RTTI names by re-using the logic for computing the
...
linkage of vtables. Before this, we were emitting RTTI names for
template instantiations with strong external linkage rather than with
weak ODR linkage.
llvm-svn: 92857
2010-01-06 22:00:56 +00:00
ccecc1bb43
Fix marking of virtual members for nested classes whose first non-pure virtual function has a body inlined in the class
...
llvm-svn: 92855
2010-01-06 20:27:16 +00:00
9a2073c035
Fix for PR5691
...
llvm-svn: 92846
2010-01-06 18:02:59 +00:00
a671c2a9fe
Fix (invalid) may-be-uninitialized warning.
...
llvm-svn: 92828
2010-01-06 07:13:04 +00:00
d01472beee
Revert my available_externally vtables experiment. It's breaking the LLVM-with-Clang build with linker errors that I have yet to investigate.
...
llvm-svn: 92822
2010-01-06 04:50:56 +00:00
81478719a7
Reorder to avoid invalidating idx which is an accessor into VCall.
...
llvm-svn: 92819
2010-01-06 03:09:19 +00:00
26751ffc53
Silence bogus GCC warning
...
llvm-svn: 92787
2010-01-05 21:55:26 +00:00
1a68ab6c83
Make use of available_externally linkage for vtables when the
...
non-inline key function of a class template instantiation, when no key
function is present, the class template instantiation itself was
instantiated with an explicit instantiation declaration (aka extern
template). I'm fairly certain that the C++0x specification gives us
this lattitude, although GCC doesn't take advantage of it.
llvm-svn: 92779
2010-01-05 21:40:05 +00:00
a318efd1f2
Improve key-function computation for templates. In particular:
...
- All classes can have a key function; templates don't change that.
non-template classes when computing the key function.
- We always mark all of the virtual member functions of class
template instantiations.
- The vtable for an instantiation of a class template has weak
linkage.
We could probably use available_externally linkage for vtables of
classes instantiated by explicit instantiation declarations (extern
templates), but GCC doesn't do this and I'm not 100% that the ABI
permits it.
llvm-svn: 92753
2010-01-05 19:06:31 +00:00
79ed590c85
Add code to skip the emission of available externally functions at -O0. WIP.
...
llvm-svn: 92749
2010-01-05 18:48:04 +00:00
a0b025e40f
When emitting member function pointers, use the canonical decl if the member function is virtual. Fixes PR5940.
...
llvm-svn: 92680
2010-01-05 05:04:05 +00:00
d443c0a02a
Revert r92431, this code isn't dead and broke the ntfs build.
...
llvm-svn: 92493
2010-01-04 18:47:06 +00:00
5b9019515a
Fix -Asserts warning.
...
llvm-svn: 92491
2010-01-04 18:02:28 +00:00
6257e08a3d
Get rid of more dead code.
...
llvm-svn: 92439
2010-01-03 01:11:47 +00:00
c52a9367b3
Get rid of some unnecessary code.
...
llvm-svn: 92432
2010-01-02 23:43:59 +00:00
14e4005851
Delete impossible case.
...
llvm-svn: 92431
2010-01-02 23:28:56 +00:00
c92a37e917
Speculative MSVC fix.
...
llvm-svn: 92421
2010-01-02 18:46:23 +00:00
93a1884fde
Move address points to CGVtableInfo, no functionality change.
...
llvm-svn: 92420
2010-01-02 18:02:32 +00:00
e36a6b3e44
Correctly pass VTT parameters to constructors and destructors. The VTTs aren't yet used in the ctors/dtors, but that will follow.
...
llvm-svn: 92409
2010-01-02 01:01:18 +00:00
27da15ba8e
Move a few more functions away from CGCXX and to CGClass and CGExprCXX.
...
llvm-svn: 92399
2010-01-01 20:29:01 +00:00
9d44fdb397
Fix last patch, catch of reference to non-pointer.
...
llvm-svn: 92386
2010-01-01 03:20:32 +00:00
5ee041873b
Fix catching a reference to a pointer.
...
llvm-svn: 92385
2010-01-01 02:51:52 +00:00
2d21394190
If the key function of a record is inline, then the RTTI data should have weak_odr linkage.
...
llvm-svn: 92371
2009-12-31 19:36:25 +00:00
7ae81bb711
More RTTI builder cleanup.
...
llvm-svn: 92360
2009-12-31 17:43:53 +00:00
a442499072
Fix a bunch of bugs with VMI RTTI building, and add a whole bunch of tests.
...
llvm-svn: 92319
2009-12-30 23:47:56 +00:00
ac2f681777
More RTTI cleanup, test that RTTI classes have the correct vtables.
...
llvm-svn: 92284
2009-12-30 01:00:12 +00:00
79c184e975
Match gcc and treat vector types as fundamental types.
...
llvm-svn: 92278
2009-12-29 22:30:11 +00:00
ef88695860
Handle enum types as well.
...
llvm-svn: 92276
2009-12-29 22:13:01 +00:00
0e4151c685
Test linkage of RTTI descriptors of array types.
...
llvm-svn: 92274
2009-12-29 21:58:32 +00:00
26cf4ab8e2
Fix function type RTTI linkage and add tests.
...
llvm-svn: 92266
2009-12-29 20:20:19 +00:00
3f8e5dd898
strength reduce this call away.
...
llvm-svn: 92253
2009-12-29 07:49:13 +00:00
88ea2455f8
Get rid of FixedWidthIntType, as suggested by Chris and Eli.
...
llvm-svn: 92246
2009-12-29 07:07:36 +00:00
9f021fd05d
adjust for llvm api changes.
...
llvm-svn: 92236
2009-12-28 23:41:39 +00:00
5e124bf913
this form of SetDebugLocation is about to go away, add some #includes that
...
are about to not come in implicitly.
llvm-svn: 92228
2009-12-28 21:44:41 +00:00
73d50b9141
Promote arguments of frameaddr / returnaddr builtins to i32 type, when needed.
...
This is needed for the platforms, where bitwidth of "int" is not 32 bits
(e.g. 16 on msp430).
llvm-svn: 92176
2009-12-27 14:27:22 +00:00
1e26278b3b
Add a sanity assertion so that we don't silently generate bad code; I'll file
...
a bug with a testcase hitting this assertion in a moment.
llvm-svn: 92175
2009-12-27 07:02:50 +00:00
af76ad73ed
Fix obvious mistake.
...
llvm-svn: 92172
2009-12-27 03:44:17 +00:00
d20ef75b91
Remove some dead variables clang-analyzer found.
...
llvm-svn: 92162
2009-12-25 15:43:36 +00:00
b8841af8bd
Minor optimization; emit proper unsupported messages for a couple of cases.
...
llvm-svn: 92153
2009-12-25 06:17:05 +00:00
2e06e8bbcc
Some small improvements to dead code elimination; helps a bit on
...
LLVM-Code-Symbols test.
llvm-svn: 92152
2009-12-25 05:29:40 +00:00
fddc26cc64
Make copy constructor elimination work in more cases; the case in question
...
here affects clang-on-clang.
llvm-svn: 92151
2009-12-24 23:33:34 +00:00
fb404888af
Move a bunch of class related functions to CGClass.cpp, no functionality change.
...
llvm-svn: 92148
2009-12-24 22:46:43 +00:00
bfb3671b25
Pass the return value slot to all call exprs.
...
llvm-svn: 92145
2009-12-24 21:13:40 +00:00
1749083e2e
Fill in the return value slot in CGExprAgg::VisitCallExpr. This takes us halfway towards fixing PR5824.
...
llvm-svn: 92142
2009-12-24 20:40:36 +00:00
61a401caec
Pass ReturnValueSlot to EmitCall. No functionality change yet.
...
llvm-svn: 92138
2009-12-24 19:25:24 +00:00
0435ed5875
Add a ReturnValueSlot class. Change the argument order in EmitCall to match the other overload better.
...
llvm-svn: 92136
2009-12-24 19:08:58 +00:00
5c13090bd4
Fix double-destruction assertion to account for temporaries in conditionals
...
llvm-svn: 92134
2009-12-24 18:16:21 +00:00
4cd5d53ebf
Assert that we aren't trying to push the same C++ temporary onto the live temporary stack twice. A little insurance against PR5867 surprising us again
...
llvm-svn: 92132
2009-12-24 17:48:05 +00:00
d538a6d364
Remove some dead code.
...
llvm-svn: 92123
2009-12-24 07:29:41 +00:00
4a779b93e4
Cleanup some dead code.
...
llvm-svn: 92122
2009-12-24 06:52:05 +00:00
bb2f57dd39
Don't set hidden for a non-external symbol as that would make it extenal.
...
Refine codegen for visibility and hidden. WIP.
llvm-svn: 92118
2009-12-24 02:33:48 +00:00
3a509cd8cc
Fix for bug 5691.
...
This fixes throwing exceptions inside @catch blocks nested inside outer @try blocks and also fixes jumping from an inner @finally to an outer @finally (via any relevant @catch blocks).
The code exhibiting this bug was based on code from CGObjCMac. I believe that this bug may still be present on the Mac runtimes, although the test case in the bug contains a few GNUisms and won't compile without some minor tweaks with Apple's libobjc.
llvm-svn: 92117
2009-12-24 02:26:34 +00:00
a8a2a33af9
Fix recent regression caught by g++.old-deja/g++.mike/eh19.C.
...
llvm-svn: 92109
2009-12-24 01:10:27 +00:00
f3c23c0cd4
Fix regression found by g++.dg/eh/alias1.C.
...
llvm-svn: 92072
2009-12-23 22:48:20 +00:00
a88d197284
Mangle block pointer types. Fixes PR5858.
...
llvm-svn: 92069
2009-12-23 22:31:44 +00:00
247894b361
There is no such thing as typeinfo for a cv-qualified type. Assert
...
that this is true when mangling, then fix up the various places in
Sema and/or CodeGen that need to remove qualifiers. Addresses a
linking issue when building LLVM with Clang.
llvm-svn: 92064
2009-12-23 22:04:40 +00:00
f0a9ba37d4
simplify my previous patch.
...
llvm-svn: 92057
2009-12-23 21:33:41 +00:00
4e1a323b85
fix opencl extvector element extraction on rvalues. We previously
...
error_unsupported on test10 and crashed on test11.
llvm-svn: 92056
2009-12-23 21:31:11 +00:00
910847c149
Mangle template template parameters. Fixes PR5861.
...
llvm-svn: 92030
2009-12-23 19:30:55 +00:00
c87915629b
Update for the intrinsic changes in llvm: the object size intrinsic
...
only takes a boolean second argument now. Update tests accordingly.
Currently the builtin still accepts the full range for compatibility.
llvm-svn: 91983
2009-12-23 03:49:37 +00:00
8c89d59c82
Move the implementation of ASTContext::getTypeSizeInChars() to the .cpp file to
...
avoid #including CharUnits.h in ASTContext.h.
llvm-svn: 91903
2009-12-22 14:23:30 +00:00
8a35c7982c
Make sure that we mangle overloaded operators that are member functions correctly, giving them the correct arity.
...
With this seemingly insignificant fix, we are now able to build and link clang using clang itself! (LLVM still has to be built with gcc for the time being).
llvm-svn: 91893
2009-12-22 06:36:32 +00:00
3780f0b680
x86_64: Structures with no fields but which have padding should be classified as
...
integer.
- This is consistent, but may not be correct. I will revisit x86_64 ABI handling for C++ as a whole at some point.
- PR5831.
llvm-svn: 91874
2009-12-22 01:19:25 +00:00
507eb87f05
Eliminate the ASTContext argument to CXXConstructorDecl::isCopyConstructor, since the context is available in the Decl
...
llvm-svn: 91862
2009-12-22 00:34:07 +00:00
0f139a1df1
Improve on my previous fix for debug information. Rather than
...
recursing in CGDebugInfo::CreateTypeNode, teach
CanonicalizeTypeForDebugInfo---now called UnwrapTypeForDebugInfo---to
keep unwrapping the type until we hit something that can be
represented by debug information. Thanks to Anders for pointing this out!
llvm-svn: 91840
2009-12-21 20:18:30 +00:00
0915b43ac2
Teach debug info generation to handle TemplateSpecializationType,
...
ElaboratedType, QualifiedNameType, and SubstTemplateTypeParmType type
nodes. Also, produce an "unsupported" diagnostic for C++0x type nodes
and "typeof" nodes, rather than asserting nondescriptly.
llvm-svn: 91837
2009-12-21 19:57:21 +00:00
2b7f444366
Incomplete structs should also have internal linkage.
...
llvm-svn: 91805
2009-12-21 00:41:42 +00:00
18e6ee1b20
Correcly handle pointers to member pointer types where the class or the pointee is incomplete.
...
llvm-svn: 91804
2009-12-20 23:37:55 +00:00
d008129f34
Rework the way pointer types are handled by the RTTI builder. We now get the right linkage for indirect pointers to incomplete structs.
...
llvm-svn: 91799
2009-12-20 22:30:54 +00:00
e017ecc9e2
Remove ';' after method definition. Noticed by clang++, which one would think
...
would have a higher respect for its own code. This is getting old, is this
warning really adding value?
llvm-svn: 91779
2009-12-19 17:50:07 +00:00
357e8c94d6
Fix for PR5524: make reference binding in default argument work correctly.
...
llvm-svn: 91733
2009-12-19 00:20:10 +00:00
70724ad2ff
Work in progress for setting the vtable pointers for all bases correctly in
...
the constructor. This doesn't handle cases requiring the VTT at the moment,
and generates unnecessary stores, but I think it's essentially correct.
llvm-svn: 91731
2009-12-18 23:47:41 +00:00
2b677cd4ff
Cleanup switch so it doesn't have a default case.
...
llvm-svn: 91725
2009-12-18 23:28:34 +00:00
303e45750e
Fix for PR5830: fix the computation of the offset to a virtual base.
...
llvm-svn: 91724
2009-12-18 23:27:44 +00:00
91b640abfc
Fix a few MSVC warnings.
...
llvm-svn: 91714
2009-12-18 20:58:47 +00:00
28a133d602
CK_UserDefinedConversion is a valid kind when doing copy ctor elision.
...
llvm-svn: 91695
2009-12-18 17:29:21 +00:00
01e620efaa
Change the return type of ASTContext::getTypeSizeInChars() from uint64_t to the
...
new opaque value type, CharUnits. This will help us avoid accidentally mixing
quantities that are in bit and character units.
llvm-svn: 91689
2009-12-18 15:55:54 +00:00
acfe6aa83a
Rename getByteSize() and getTypeSizeInBytes() in ASTContext to getCharWidth()
...
and getTypeSizeInChars() to reflect their basis in character type units, not
that of a possibly independent architecture-specific byte.
llvm-svn: 91688
2009-12-18 15:24:05 +00:00
0087bc851c
Fix regression I introduced when dynamic_cast-ing to a reference type.
...
llvm-svn: 91687
2009-12-18 14:55:04 +00:00
b396f43520
It's perfectly fine to see UserDefinedConversion casts when emitting scalar expressions.
...
llvm-svn: 91686
2009-12-18 14:42:03 +00:00
357b6fd3e5
Address some of Anders comments
...
llvm-svn: 91670
2009-12-18 05:19:44 +00:00
e1314a64b8
Switch the initialization required by return statements over to the
...
new InitializationSequence. This fixes some bugs (e.g., PR5808),
changed some diagnostics, and caused more churn than expected. What's
new:
- InitializationSequence now has a "C conversion sequence" category
and step kind, which falls back to
- Changed the diagnostics for returns to always have the result type
of the function first and the type of the expression second.
CheckSingleAssignmentConstraints to peform checking in C.
- Improved ASTs for initialization of return values. The ASTs now
capture all of the temporaries we need to create, but
intentionally do not bind the tempoary that is actually returned,
so that it won't get destroyed twice.
- Make sure to perform an (elidable!) copy of the class object that
is returned from a class.
- Fix copy elision in CodeGen to properly see through the
subexpressions that occur with elidable copies.
- Give "new" its own entity kind; as with return values and thrown
objects, we don't bind the expression so we don't call a
destructor for it.
Note that, with this patch, I've broken returning move-only types in
C++0x. We'll fix it later, when we tackle NRVO.
llvm-svn: 91669
2009-12-18 05:02:21 +00:00
24c7f5efc5
Add -dwarf-debug-flags, which provides a way to embed the cc1 level options used
...
to compile a translation unit into the debug info for that file.
- Used by parts of Darwin build process to check compiler flags, etc.
- <rdar://problem/7256886> clang does not emit AT_APPLE_flags
llvm-svn: 91661
2009-12-18 02:43:17 +00:00
0da53228e6
Handle case when DestPtr is 0.
...
llvm-svn: 91658
2009-12-18 02:14:27 +00:00
3f4336cb1f
Rename GetAddrOfRTTI to GetAddrOfRTTIDescriptor. Remove the overload that takes a CXXRecordDecl since we were just creating a QualType from it anyway.
...
llvm-svn: 91590
2009-12-17 07:09:17 +00:00
cd8faeb0b1
Ensure we run cleanups for CXXTemporaries on the exceptional edge. WIP.
...
llvm-svn: 91588
2009-12-17 06:08:47 +00:00
1fd734276d
Simplify RTTIBuilder::finish.
...
llvm-svn: 91585
2009-12-17 05:10:59 +00:00
e5a94105eb
Move the Info vector into the RTTIBuilder struct. No functionality change.
...
llvm-svn: 91583
2009-12-17 05:06:03 +00:00
7f4933f6c7
Fix for PR5801: codegen memcpy, memmove, memset directly to LLVM intrinsics.
...
llvm-svn: 91573
2009-12-17 00:14:28 +00:00
4f4b186215
When value-initializing a class with no user-defined constructors but
...
with a non-trivial default constructor, zero-initialize the storage
and then call the default constructor. Fixes PR5800.
llvm-svn: 91548
2009-12-16 18:50:27 +00:00
d681a29ac0
Baby steps towards fixing PR5589. If a class needs a vtable pointer, add one.
...
llvm-svn: 91545
2009-12-16 17:27:20 +00:00
07863b8b6e
Use GetAddrOfRTTI everywhere and remove GenerateRTTI and GenerateRTTIRef. With this change, we can now compile and link TableGen.
...
llvm-svn: 91520
2009-12-16 07:05:41 +00:00
18ada9856f
Handle ImplicitValueInitExpr in AggExprEmitter.
...
llvm-svn: 91519
2009-12-16 06:57:54 +00:00
b7e93e268f
Mangle CXXOperatorCallExprs, fixes PR5796.
...
llvm-svn: 91507
2009-12-16 05:48:46 +00:00
b85ffb689e
Add an addition check for undefined behavior for when we hit a
...
__builtin_unreachable. WIP.
llvm-svn: 91499
2009-12-16 03:07:12 +00:00
3f6f9febad
Implement additional undefined checks for additional loads and stores. WIP.
...
llvm-svn: 91498
2009-12-16 02:57:00 +00:00
2b9fc83db5
Fixes a code gen bug related to accessing a now
...
non-existing 'isa' field of a non-existing struct type
all related to legacy type definition for 'id' which we have
dropped in clang in favor of a built-in type.
(fixes radar 7470820).
llvm-svn: 91455
2009-12-15 21:34:52 +00:00
6e997b2993
ShouldDestroyTemporaries? I don't think so.
...
llvm-svn: 91450
2009-12-15 20:51:39 +00:00
2f3da9b205
update to match LLVM API change:
...
Remove isPod() from DenseMapInfo, splitting it out to its own
isPodLike type trait. This is a generally useful type trait for
more than just DenseMap, and we really care about whether something
acts like a pod, not whether it really is a pod.
llvm-svn: 91422
2009-12-15 07:26:51 +00:00
2151725325
Fix a small bug in ComputeMethodVtableIndices.
...
llvm-svn: 91411
2009-12-15 03:31:17 +00:00
d777cccb47
Don't force the emission of destructor definitions.
...
llvm-svn: 91394
2009-12-15 02:06:15 +00:00
cf16d2cc42
Fix spacing.
...
llvm-svn: 91386
2009-12-15 01:22:35 +00:00
9a4e0128a8
Ensure we preserve line information for each trap for
...
-fcatch-undefined-behavior if we aren't optimizing. WIP.
llvm-svn: 91382
2009-12-15 00:59:40 +00:00
e8c3b3eca4
Switch codegen for -fcatch-undefined-bahavior over to __builtin_trap
...
instead of abort to improve codesize and codegen.
llvm-svn: 91374
2009-12-15 00:35:12 +00:00
590d18f103
We have to allow one to form an address for one past the end. WIP.
...
llvm-svn: 91347
2009-12-14 22:14:31 +00:00
ba6a0c40c4
Add support for detecting undefined shift behavior. WIP.
...
llvm-svn: 91341
2009-12-14 21:58:14 +00:00
9ac0a04b29
Patch to fix 32-bit @try failure with internal assertion when compiling
...
an Objective-C rethrow nested inside another try/catch block. (fixes radar 7466728).
llvm-svn: 91335
2009-12-14 21:35:01 +00:00
f8858af909
Simplifiy.
...
llvm-svn: 91324
2009-12-14 20:52:00 +00:00
67dfd4236a
Revert mmx palignr to use an intrinsic, since mmx shuffle patterns are missing.
...
llvm-svn: 91269
2009-12-14 05:15:02 +00:00
72ec6bc6f4
Support x86's PALIGNR instruction without the use of a palignr intrinsic.
...
llvm-svn: 91264
2009-12-14 04:57:03 +00:00
e66e2942fc
Mangle unary, binary and ternary expressions correctly.
...
llvm-svn: 91257
2009-12-14 01:45:37 +00:00
c67974ccb1
Simplify BuildSimpleType in the RTTBuilder to avoid using an std::vector.
...
llvm-svn: 91255
2009-12-13 23:47:29 +00:00
399f499f0c
Don't use a cookie if the global placement new function is used.
...
llvm-svn: 91251
2009-12-13 20:34:34 +00:00
adbe4249d6
Fix regression in my last commit - if a struct has a trivial destructor but no usual deallocation function we don't need a cookie.
...
llvm-svn: 91249
2009-12-13 20:10:12 +00:00
21122cf657
More array cookie improvements.
...
llvm-svn: 91247
2009-12-13 20:04:38 +00:00
f5af8dead7
If the usual array deallocation function of a class takes two arguments, we need to pass the number of bytes allocated to the deallocation function, just not the number of bytes in a single element.
...
llvm-svn: 91246
2009-12-13 18:48:07 +00:00
e7025c55a4
Change the RTTIBuilder's finish member function to take a pointer to the constants array + the length of the array.
...
llvm-svn: 91241
2009-12-13 16:38:01 +00:00
8abbc657a9
Silence some release build warnings.
...
llvm-svn: 91234
2009-12-13 01:37:04 +00:00
751c1e7b24
patch to add a property from a protocol to a class that adopts the protocol.
...
(fixes radar 7466494).
llvm-svn: 91227
2009-12-12 21:26:21 +00:00
1615d45daa
Un-namespace-qualify llvm_unreachable. It's a macro, so the qualification gave
...
no extra safety anyway.
llvm-svn: 91207
2009-12-12 05:05:38 +00:00
d954638f02
Implement runtime checks for undefined behavior. WIP.
...
This implements a new flag -fcatch-undefined-behavior. The flag turns
on additional runtime checks for:
T a[I];
a[i] abort when i < 0 or i >= I.
Future stuff includes shifts by >= bitwidth amounts.
llvm-svn: 91198
2009-12-12 01:27:46 +00:00
ffe912c647
Patch to allow C-style cast from 'void *' to block pointer type.
...
(fixes radar 7465023).
llvm-svn: 91171
2009-12-11 22:40:48 +00:00
b095e15af4
Fix for PR5714: make sure globals that will be modified aren't marked const.
...
llvm-svn: 91156
2009-12-11 21:23:03 +00:00
3ace52b3a7
Fix linkage of type info and vtable for classes without linkage.
...
llvm-svn: 91152
2009-12-11 20:48:18 +00:00
75c9b97123
Make sure mangling doesn't crash in another case. Add some more tests.
...
llvm-svn: 91149
2009-12-11 20:21:38 +00:00
30e94d047e
Fix for PR5706: let mangleName deal with mangling names without identifiers
...
correctly.
llvm-svn: 91136
2009-12-11 18:00:57 +00:00
1d6ad50826
Move info vectors and add assertions in preparation of moving the vector directly into RTTIBuilder.
...
llvm-svn: 91129
2009-12-11 16:41:51 +00:00
a0b98f083a
Use GetAddrOfRTTI when getting the RTTI pointer for a base class.
...
llvm-svn: 91127
2009-12-11 16:37:06 +00:00
b11118bd0b
Random string-related cleanups.
...
llvm-svn: 91119
2009-12-11 13:33:18 +00:00
68396b16f6
Move the code for converting a member pointer to a bool so that it is usable
...
for logical not.
llvm-svn: 91112
2009-12-11 09:26:29 +00:00
1762cf2ad0
Fix for PR5718: implement equality comparisons for member function pointers.
...
llvm-svn: 91108
2009-12-11 07:36:43 +00:00
fd7dfeb779
Improve linkage of RTTI data structures. Introduce CodeGenModule::GetAddrOfRTTI which figures out the right linkage of the RTTI information for the given type and whether it should be defined or not. I will migrate clients over to GetAddrOfRTTI in subsequent commits (with tests).
...
llvm-svn: 91098
2009-12-11 02:46:30 +00:00
c988201417
Use named flags in RTTIBuilder::BuildPointerType.
...
llvm-svn: 91088
2009-12-11 01:27:37 +00:00
0c6a7d8f42
When extending the lifetime of a temporary, make sure to emit a branch to the cleanup exit block. This fixes a broken module error in LLVMCConfigurationEmitter.cpp.
...
llvm-svn: 91086
2009-12-11 01:00:09 +00:00
afd1edb52e
When an exception needs to be freed by calling __cxa_exception_free, make sure to stash away the exception pointer somewhere.
...
This fixes an "Instruction does not dominate all uses!" verification error when compiling TableGen.
llvm-svn: 91084
2009-12-11 00:32:37 +00:00
37077bf651
Fix spacing.
...
llvm-svn: 91078
2009-12-11 00:02:10 +00:00
d70bbfd7d8
Support unary type traits in a scalar context. Not that I've actually seen
...
this construct, but might as well for completeness.
llvm-svn: 91071
2009-12-10 22:40:32 +00:00
781161dc55
Make sure that explicitly instantiated functions get the right linkage.
...
llvm-svn: 91069
2009-12-10 22:25:34 +00:00
bf56946073
Mangle static variables inside Objective-C methods in Objective-C++. We currently mangle them the same way as gcc does.
...
llvm-svn: 91042
2009-12-10 03:14:39 +00:00
d5bc94e2eb
Get rid of static variable.
...
llvm-svn: 91041
2009-12-10 02:21:21 +00:00
fefc7b8e7b
Actually store the pointer into the global variable, spotted by Eli.
...
llvm-svn: 91036
2009-12-10 01:58:33 +00:00
3c6a12ca65
Setup cleanup on eh edge out of the construction of the cleanup object
...
during throw to deallocate the exception object. WIP.
llvm-svn: 91035
2009-12-10 01:52:30 +00:00
a72ddd4609
Handle emitting static variables that have reference type.
...
llvm-svn: 91027
2009-12-10 01:05:11 +00:00
364051c6a6
Separate generating code for static variables and global variables.
...
llvm-svn: 91026
2009-12-10 00:57:45 +00:00
633c6f6f36
Move more decl related functions to CGDeclCXX.cpp
...
llvm-svn: 91012
2009-12-10 00:30:05 +00:00
bc49cfed84
Add CGDeclCXX.cpp and move EmitCXXGlobalVarDeclInit there.
...
llvm-svn: 91006
2009-12-10 00:16:00 +00:00
75546b8c88
Fixup spacing.
...
llvm-svn: 91001
2009-12-10 00:06:18 +00:00
2600491817
Fix spacing.
...
llvm-svn: 91000
2009-12-10 00:05:14 +00:00
f5cbb08c4b
Ensure we only generate one terminate handler.
...
llvm-svn: 90998
2009-12-10 00:02:42 +00:00
3b60f034f9
Remove obsolete FIXME and some dead fallback code.
...
llvm-svn: 90997
2009-12-09 23:52:09 +00:00
102530719f
Add terminate handler to cleanups on exceptional edges.
...
llvm-svn: 90996
2009-12-09 23:49:53 +00:00
531c16f9fb
Code gen for ObjCIsaExpr AST used as lvalue.
...
(fixes radar 7457534).
llvm-svn: 90995
2009-12-09 23:35:29 +00:00
25b20fc2ae
Add terminate handler for copy constructors for thrown objects. WIP.
...
llvm-svn: 90994
2009-12-09 23:31:35 +00:00
2b488873bf
Break out code for reuse. WIP.
...
llvm-svn: 90991
2009-12-09 22:59:31 +00:00
f16b8c30d9
Add throw support. WIP.
...
llvm-svn: 90982
2009-12-09 19:24:08 +00:00
a5fee26d28
Codegen. support for ObjCIsaExpr AST which until now
...
was not needed (fixes radar 7453430).
llvm-svn: 90981
2009-12-09 19:05:56 +00:00
aff69af918
Add cleanups for exceptional edges. WIP.
...
llvm-svn: 90940
2009-12-09 03:35:49 +00:00
4f678f3de1
Fix for PR5709: use the computed type of the declaration instead of the
...
type of the builtin when generating the function declaration for a builtin
library call.
llvm-svn: 90936
2009-12-09 03:05:59 +00:00
859c415567
More detailed analysis of typecast to an objective-c pointer
...
in objective-c++ mode without being too lenient.
llvm-svn: 90895
2009-12-08 23:09:15 +00:00
4a6b337cfd
Add fixme.
...
llvm-svn: 90884
2009-12-08 22:12:48 +00:00
e6d54ca553
Remove some old code. WIP.
...
llvm-svn: 90882
2009-12-08 21:46:41 +00:00
17c4f8d746
Update CGExprConstant for change to emit padding values as undef.
...
- This fixes 2003-05-21-BitfieldHandling.
llvm-svn: 90876
2009-12-08 21:12:32 +00:00
ba8451b243
Use StringRef in CGDebugInfo::EmitFunctionStart.
...
llvm-svn: 90856
2009-12-08 14:04:35 +00:00
43c7c5befd
Use a Twine to concatenate the name instead of going through std::string.
...
llvm-svn: 90854
2009-12-08 13:07:37 +00:00
1d20564b8d
Use Path.makeAbsolute() and make a constant std::string a const char*.
...
llvm-svn: 90851
2009-12-08 11:02:29 +00:00
ce54657e95
DeclRefExpr stores a ValueDecl internally.
...
Template instantiation can re-use DeclRefExprs.
llvm-svn: 90848
2009-12-08 09:08:17 +00:00
80888c7b38
Build fix.
...
llvm-svn: 90841
2009-12-08 06:54:20 +00:00
bb5008a32a
Fix for PR5707: make sure implicit copy constructors initialize the vtable
...
pointer.
llvm-svn: 90840
2009-12-08 06:46:18 +00:00
48c6b0d14f
Slight tweak to vtable linkage.
...
llvm-svn: 90832
2009-12-08 04:09:14 +00:00
f2c79b6b9c
Misc key function fixes.
...
llvm-svn: 90831
2009-12-08 03:56:49 +00:00
8aaff69e86
Fix some direct checks of expressions which might be surrounded by parentheses.
...
llvm-svn: 90825
2009-12-08 02:09:46 +00:00
cd6a50f3e5
Make copy assignment operator synthesis not explode for classes with complex
...
or non-record aggregate members.
It might be worth spending some time to optimize this code (and the parallel
code for copy constructors) to memcpy in larger chunks, rather than copying
one member at a time. Not sure exactly how beneficial that would be, but
it seems like could help for large classes with, for example, a vtable pointer
forcing the generation of a copy constructor.
llvm-svn: 90823
2009-12-08 01:57:53 +00:00
567ae47b4a
Remove several .c_str() to be forward-compatible with StringRef.
...
llvm-svn: 90822
2009-12-08 01:46:24 +00:00
02c23d68cb
Fixup catch parameters with class reference type. WIP.
...
llvm-svn: 90821
2009-12-08 01:29:31 +00:00
220bf4fc5a
No need to add tail padding if the resulting LLVM struct type will have the same size as the final record size.
...
llvm-svn: 90820
2009-12-08 01:24:23 +00:00
31bc3ad275
A bunch more thunk fixes from misc testing.
...
(Yes, I do intend to commit some tests for this.)
llvm-svn: 90818
2009-12-07 23:56:34 +00:00
1d849219a5
Add codegen support for exception specifications. WIP.
...
llvm-svn: 90817
2009-12-07 23:38:24 +00:00
f6c175b745
Correctly handle conditional operators involving throw.
...
llvm-svn: 90800
2009-12-07 20:25:53 +00:00
5445f6e5b6
Make the comma operator consistently call EnsureInsertPoint.
...
llvm-svn: 90799
2009-12-07 20:18:11 +00:00
62afe99814
Work around emitters that can't deal with dead code contexts yet.
...
llvm-svn: 90796
2009-12-07 20:12:14 +00:00
1aaecfa02d
Mangle basic_ostream and basic_iostream specializations.
...
llvm-svn: 90794
2009-12-07 19:56:42 +00:00
52d7141186
Remove obsolete fallback code for objectsize.
...
llvm-svn: 90780
2009-12-07 18:58:11 +00:00
7251327d75
implement PR5274: mark 'restrict' parameters as noalias
...
llvm-svn: 90778
2009-12-07 18:30:06 +00:00
e1b3e6292a
It's OK to try to emit a vtable definition more than once. Fixes PR5697.
...
llvm-svn: 90751
2009-12-07 07:59:52 +00:00
c0dacee1e0
Erm, revert for the moment; I didn't test this as thoroughly as I should have
...
(although it does pass regression tests).
llvm-svn: 90747
2009-12-07 04:45:50 +00:00
c5d3a6ae3d
Tweak the formula for non-virtual offsets to something which appears a bit
...
more accurate.
llvm-svn: 90746
2009-12-07 04:38:34 +00:00
5ebf8b44e9
Move key functions to a separate map.
...
llvm-svn: 90745
2009-12-07 04:35:11 +00:00
f2eda5efd2
Be a bit more complete about accumulating SavedThisAdjustments.
...
llvm-svn: 90723
2009-12-06 22:33:51 +00:00
8174f2c23c
Work-in-progess rewrite of thunks: move thunk generation outside of vtable
...
generation, and make sure we generate thunks when the function is defined
rather than when the vtable is defined.
llvm-svn: 90722
2009-12-06 22:01:30 +00:00
3efc6e6f46
Add rudimentary support for member pointers to CGDebugInfo.
...
llvm-svn: 90711
2009-12-06 18:00:51 +00:00
72ecc68e11
use new helpers to simplify code.
...
llvm-svn: 90709
2009-12-06 17:22:42 +00:00
fe5f7d916f
Set the correct linkage for VTTs as well.
...
llvm-svn: 90689
2009-12-06 01:09:21 +00:00
232324c5c0
More linkage improvements.
...
llvm-svn: 90687
2009-12-06 00:53:22 +00:00
0911ae8ef2
Pass the desired vtable linkage to GenerateVtable directly. Only call MaybeMarkVirtualImplicitMembersReferenced for non-inline functions.
...
llvm-svn: 90686
2009-12-06 00:23:49 +00:00
7e28c5f2e2
Make GenerateVtable a private member function of CGVtableInfo.
...
llvm-svn: 90684
2009-12-06 00:01:05 +00:00
4ed44eb7d9
Only emit the vtable definition if the class has a key function and we're emitting it, or if the class doesn't have a key function and we are emitting the complete constructor.
...
llvm-svn: 90681
2009-12-05 22:42:54 +00:00
0ab79e22ab
If a class does not have a key function, its linkage should be weak_odr.
...
llvm-svn: 90680
2009-12-05 22:24:38 +00:00
58b271d450
Use createGlobalVariable for creating vtable variables too.
...
llvm-svn: 90679
2009-12-05 22:19:10 +00:00
b3f54b748d
CodeGenModule::GenerateVtable now returns a pointer directly to the vtable and not to the address point.
...
llvm-svn: 90676
2009-12-05 21:28:12 +00:00
a95d4c51dc
Make sure that hte vtable always has an i8* array type.
...
llvm-svn: 90675
2009-12-05 21:09:05 +00:00
c8e39ec79f
Add a function for getting the vtable address point of a class.
...
llvm-svn: 90674
2009-12-05 21:03:56 +00:00
b694f05ffb
Simplify BuildVTT.
...
llvm-svn: 90673
2009-12-05 21:02:25 +00:00
b78d093fcc
Let the VTT builder pretend that getVtable returns a pointer to the vtable and not to the vtable address point.
...
llvm-svn: 90672
2009-12-05 20:58:49 +00:00
78c2898e02
Change getCtorVtable to not return the address point offset, but the global variable.
...
llvm-svn: 90671
2009-12-05 20:42:53 +00:00
5a1a84f9fa
Simplify some code.
...
llvm-svn: 90670
2009-12-05 18:38:15 +00:00
21bbc1e705
Factor vtable related GlobalVariable creation out into a separate function. Add vtable linkage test.
...
llvm-svn: 90667
2009-12-05 17:04:47 +00:00
e99c110d06
implement rdar://7346691 by codegen'ing struct/array initializers
...
to a memset or a memcpy from a global when possible.
llvm-svn: 90658
2009-12-05 08:22:11 +00:00
e08ea7874a
various cleanups.
...
llvm-svn: 90657
2009-12-05 08:21:30 +00:00
ffcd06ea7c
simplify a condition and add a testcase.
...
llvm-svn: 90652
2009-12-05 06:49:57 +00:00
6c08ce7706
Skip actually generating the vtable unless we are defining it. This avoids
...
the side-effects of generating the methods in the vtable (including PR5676).
llvm-svn: 90635
2009-12-05 01:05:03 +00:00
a30e1750cd
Don't emit explicit specializations of static member variable declarations.
...
llvm-svn: 90624
2009-12-04 23:50:01 +00:00
bee78dd31b
Add support for function try blocks.
...
llvm-svn: 90622
2009-12-04 23:26:17 +00:00
b96ec56871
Fix "using typename" and the instantiation of non-dependent using declarations.
...
llvm-svn: 90614
2009-12-04 22:46:56 +00:00
62ab616666
Return bool as a bool instead of a uint64_t.
...
llvm-svn: 90610
2009-12-04 22:45:27 +00:00
9b3a395bf3
Put in FIXME that this mangling is not official in.
...
llvm-svn: 90594
2009-12-04 21:11:13 +00:00
bf2f0c2e34
Switch mangling of literal operator names to a string that's
...
a) legal
b) likely to be chosen as the official mangling
This will break ABI compatibility with all literal operator names,
so you may need to recompile any such code. Sorry.
llvm-svn: 90587
2009-12-04 21:01:37 +00:00
afe4764fa5
Be sure to instantiate rtti for non-polymorphic bases.
...
llvm-svn: 90586
2009-12-04 20:55:13 +00:00
ce460529c0
When generating a virtual destructor, don't try to make a virtual call to the base class destructor because then we'll just re-enter the same destructor!
...
This was done to fix PR5619, so I went ahead and passed a dummy VTT pointer for now.
llvm-svn: 90578
2009-12-04 19:33:17 +00:00
875912ac13
Fixup nested try blocks.
...
llvm-svn: 90576
2009-12-04 19:21:57 +00:00
5c820758f5
Fix warning and be sure to set up the rethrow pointer if we fall off
...
the end of all the catches.
llvm-svn: 90574
2009-12-04 19:03:47 +00:00
bad80eb792
Rename 'Class' to 'MostDerivedClass'
...
llvm-svn: 90573
2009-12-04 18:36:22 +00:00
19462d681a
Store the vtable components in a SmallVector.
...
llvm-svn: 90571
2009-12-04 16:24:46 +00:00
4e1d75f3ed
Use a different ConstantArray::get.
...
llvm-svn: 90570
2009-12-04 16:22:27 +00:00
472404f8a6
Move the vtable vector directly into the Vtable builder.
...
llvm-svn: 90569
2009-12-04 16:19:30 +00:00
7bb7076b51
Change getIndex to return false if the index wasn't found. Avoids an extra hash lookup.
...
llvm-svn: 90568
2009-12-04 15:49:02 +00:00
65d87222b2
Minor cleanup; no functionality change.
...
llvm-svn: 90565
2009-12-04 08:52:11 +00:00
81fb0d2656
Remove unused argument.
...
llvm-svn: 90563
2009-12-04 08:40:51 +00:00
c07aebee90
Eliminate submethods vector.
...
llvm-svn: 90562
2009-12-04 08:36:58 +00:00
3d2e9de337
Eliminate the inner loop in VtableBuilder::OverrideMethod.
...
llvm-svn: 90561
2009-12-04 08:34:14 +00:00
5c9e7b16a5
Correctly mangle the 'std' namespace inside extern "C++" blocks.
...
llvm-svn: 90544
2009-12-04 06:23:23 +00:00
f3935b4d4b
Make sure that overridden method decls are always canonical.
...
llvm-svn: 90542
2009-12-04 05:51:56 +00:00
94bc2dade6
Fix regression in vtable improvements.
...
llvm-svn: 90540
2009-12-04 04:30:03 +00:00
8c1253b4c2
Try/catch statements seem to be working well enough to turn on.
...
llvm-svn: 90538
2009-12-04 03:57:07 +00:00
2ab55765f7
Disable for now, doesn't play nice with the temporary code.
...
llvm-svn: 90537
2009-12-04 03:55:53 +00:00
b05eb963e0
Rename method to something easier to search for.
...
llvm-svn: 90536
2009-12-04 03:54:56 +00:00
a93e980fbc
More work in preparation of getting rid of the submethods loop.
...
llvm-svn: 90535
2009-12-04 03:52:52 +00:00
f2f31f4371
Remove an unused member variable.
...
llvm-svn: 90534
2009-12-04 03:46:21 +00:00
e609636e0d
Add a way to get the index of a method. Assert that we have the same index for now.
...
llvm-svn: 90533
2009-12-04 03:41:37 +00:00
b07567c315
Add a premature optimization.
...
llvm-svn: 90532
2009-12-04 03:07:26 +00:00
50f147460a
Add the method directly to the vtable.
...
llvm-svn: 90531
2009-12-04 03:06:03 +00:00
79bce40365
Get rid of the PureVirtualMethods map.
...
llvm-svn: 90530
2009-12-04 02:58:12 +00:00
ddf42c8d80
Move handling of pure virtual methods to AppendMethods (and rename it to AppendMethodsToVtable).
...
llvm-svn: 90529
2009-12-04 02:56:03 +00:00
5b3ea9bf1c
Move covariant thunk handling to AppendMethods too.
...
llvm-svn: 90528
2009-12-04 02:52:22 +00:00
86809cd8ab
Move 'this' pointer adjustment thunks to AppendMethods.
...
llvm-svn: 90527
2009-12-04 02:43:50 +00:00
4b1942cb8b
Make functions returning a struct indirectly evaluate the returned struct
...
directly into the sret pointer. This is an optimization in C, but is required
for correctness in C++ for classes with a non-trivial copy constructor.
llvm-svn: 90526
2009-12-04 02:43:40 +00:00
495634e5ac
Factor appending methods to a vtable out into a separate function.
...
llvm-svn: 90525
2009-12-04 02:39:04 +00:00
4c837d258f
Get rid of the Thunks struct too.
...
llvm-svn: 90524
2009-12-04 02:26:15 +00:00
c521f952d6
Remove the CovariantThunk struct.
...
llvm-svn: 90523
2009-12-04 02:22:02 +00:00
597c776c41
Remove the GlobalDecl from the Thunk as well.
...
llvm-svn: 90522
2009-12-04 02:14:12 +00:00
8c889abc1a
Remove the GlobalDecl from the CovariantThunk struct, we can just look it up in the Methods table now.
...
llvm-svn: 90519
2009-12-04 02:11:21 +00:00
cdf1898086
Start populating the VtableMembers structure.
...
llvm-svn: 90517
2009-12-04 02:08:24 +00:00
a84b6e85f0
Add a data structure for efficient storing of vtable methods. Not used yet.
...
llvm-svn: 90515
2009-12-04 02:01:07 +00:00
114ab9f695
Fixup reference binding for catch parameters.
...
Fixup throws and rethrows to use invoke as appropriate.
llvm-svn: 90513
2009-12-04 01:51:45 +00:00
6d11ec8cb8
Update chunk of #if 0'ed code to remove fixed FIXME and make it compile. We
...
probably want to do some sort of performance assessment before enabling it,
though.
llvm-svn: 90510
2009-12-04 01:30:56 +00:00
5f141b75f9
These are done.
...
llvm-svn: 90494
2009-12-03 22:55:42 +00:00
7398ff0f07
Improve catch parameter bindings for scalar non-pointers. WIP.
...
llvm-svn: 90492
2009-12-03 22:38:15 +00:00
ce2cd01647
Handle static_assert inside functions.
...
llvm-svn: 90461
2009-12-03 17:26:31 +00:00
cdeb800152
Eli, I copied my code from this code... Let's fix the souce of the bad idea!
...
Thanks.
llvm-svn: 90458
2009-12-03 16:55:20 +00:00
c339081c24
Fix for PR5659: correct a rather nasty oversight in the type conversion for
...
member pointer types.
llvm-svn: 90450
2009-12-03 12:44:31 +00:00
7230fa5630
Convert ABIArgInfo::dump to raw_ostream.
...
llvm-svn: 90440
2009-12-03 09:13:49 +00:00
c6089fda20
Attempt to fix the MSVC build.
...
llvm-svn: 90427
2009-12-03 07:30:40 +00:00
4039f35344
Fix thunk generation for thunks with a parameter with reference type.
...
llvm-svn: 90412
2009-12-03 04:49:52 +00:00
551fe84d0e
Minor cleanup.
...
llvm-svn: 90411
2009-12-03 04:27:05 +00:00
e2d4a2c3a8
Add support for thunking dtors. Oh why does this make my head hurt?
...
llvm-svn: 90409
2009-12-03 03:47:56 +00:00
2842b4cf86
Reflow.
...
llvm-svn: 90407
2009-12-03 03:40:14 +00:00
dabfa3cd8b
Revert r90402 for now, virt.cpp is failing.
...
llvm-svn: 90406
2009-12-03 03:28:24 +00:00
e692f7c309
Use Eli's ComputeThunkAdjustment for calculating the return adjustment.
...
llvm-svn: 90402
2009-12-03 03:15:31 +00:00
5f91fd64b6
Add CodeGenModule::ComputeThunkAdjustment, which Eli wrote.
...
llvm-svn: 90401
2009-12-03 03:06:55 +00:00
80bc5d5d04
Remove the index from the Thunk struct.
...
llvm-svn: 90400
2009-12-03 02:41:55 +00:00
29a1f751b3
Change the Thunks map to use the vtable index as the key.
...
llvm-svn: 90399
2009-12-03 02:39:59 +00:00
0e1e7632bc
Add the global decl to the Thunk struct.
...
llvm-svn: 90398
2009-12-03 02:36:40 +00:00
9f98f7a9e5
Remove unused struct fields.
...
llvm-svn: 90397
2009-12-03 02:34:59 +00:00
657f139abe
Delay computing the return adjustments for covariant thunks until when they are added to the vtable.
...
llvm-svn: 90396
2009-12-03 02:32:59 +00:00
2ca285fc5c
No need to create the covariant thunk in both places now.
...
llvm-svn: 90394
2009-12-03 02:22:59 +00:00
1157e8fc68
Whoops, forgot to save :)
...
llvm-svn: 90393
2009-12-03 02:20:26 +00:00
06c14b6f65
Remove the index field from the CovariantThunk structure.
...
llvm-svn: 90392
2009-12-03 02:16:14 +00:00
73295f96c5
Change the CovariantThunk map to use the vtable index as its key.
...
llvm-svn: 90391
2009-12-03 02:12:03 +00:00
c38b40a709
Store a GlobalDecl in the return adjustment.
...
llvm-svn: 90387
2009-12-03 02:03:29 +00:00
2bd3c0fa34
Do not include the 'this' pointer adjustment in the covariant return type. Instead, store it in the (now oddly named) Thunks map.
...
llvm-svn: 90386
2009-12-03 01:58:20 +00:00
ca1bf68cfb
Move VtableBuilder::OverrideMethod out of line in preparation of other changes to it. No functionality change.
...
llvm-svn: 90382
2009-12-03 01:54:02 +00:00
26ada867ba
Remove untrue statement.
...
llvm-svn: 90377
2009-12-03 01:34:13 +00:00
59daccda08
Note a failure I saw from the g++ testsuite:
...
FAIL: g++.old-deja/g++.mike/eh23.C (test for excess errors)
llvm-svn: 90374
2009-12-03 01:22:05 +00:00
37e4c41283
Work-in-progress: teach mangler how to mangle thunks for destructors.
...
llvm-svn: 90360
2009-12-03 00:03:05 +00:00
2d52ce5f94
Cleanups on exceptional edges don't work at all, yet. This doesn't
...
catch very many of them and if we caught all of them, the errors would
be annoying. I'm working on this next. WIP.
llvm-svn: 90358
2009-12-02 23:46:02 +00:00
90990965e0
Add a cleanup scope for each catch clause.
...
llvm-svn: 90357
2009-12-02 23:37:16 +00:00
a68e2a2967
Add missing branch to exit. Seemingly obvious when I look at the
...
code, but to track this down was laborious.
llvm-svn: 90356
2009-12-02 23:28:08 +00:00
d4df7753d2
Fix for PR5522 and PR5666: fix a bunch of mangling issues with extern variables
...
and funcctions declared locally within a function.
llvm-svn: 90344
2009-12-02 20:32:49 +00:00
9732915bf9
Pull the terminate handler up so that we can use it for the catch
...
parameter setup code and set up the catch parameter setup code to
protect that code with terminate.
llvm-svn: 90340
2009-12-02 19:53:57 +00:00
8a96d3a355
Turn off for now.
...
llvm-svn: 90339
2009-12-02 19:50:41 +00:00
ae1b85d984
Put the Builder classes into the anonymous namespace.
...
llvm-svn: 90335
2009-12-02 19:07:44 +00:00
c01c2b87e8
Change rtti/Rtti to RTTI, as it is an acronym.
...
llvm-svn: 90334
2009-12-02 18:57:08 +00:00
d6eb364b31
Rename CGRtti.cpp to CGRTTI.cpp.
...
llvm-svn: 90331
2009-12-02 18:32:08 +00:00
a9a590cd12
Avoid warning for getTerminateFn defined but not used.
...
llvm-svn: 90330
2009-12-02 18:20:18 +00:00
a510597ef9
terminate doesn't throw.
...
llvm-svn: 90308
2009-12-02 08:18:09 +00:00
33270211a0
More exception handling improvements... WIP.
...
Highlights include:
Add a helper to generate __cxa_free_exception and _ZSt9terminatev.
Add a region to handle EH object deallocation for ctor failures for throw.
Add a terminate handler for __cxa_end_catch.
A framework for adding cleanup actions for the exceptional edges only.
llvm-svn: 90305
2009-12-02 07:41:41 +00:00
037bcb5be3
Added an assert.
...
llvm-svn: 90285
2009-12-01 23:35:33 +00:00
b2b120f664
Make EmitStoreOfScalar generate a more sane representation of boolean stores.
...
"Fixes" PR5645.
llvm-svn: 90272
2009-12-01 22:31:51 +00:00
a6227fd896
(objc2 nonfragile-abi specific). If the translation unit includes an implementation
...
of a subclass (direct or indirect) of a weak_import root class, emit a weak reference
for the root class's metaclass (should complete radar 6815425).
llvm-svn: 90249
2009-12-01 18:25:24 +00:00
c377c81dcf
Update for llvm intrinsics change.
...
llvm-svn: 90208
2009-12-01 05:00:51 +00:00
5406614be9
Much work on try/catch statement. WIP.
...
Highlights include:
Fix __cxa_begin_catch so it now returns a value.
Added getUnwindResumeOrRethrowFn helper to build up calls to
_Unwind_Resume_or_Rethrow.
Broke out object copying code into CopyObject from EmitCXXThrowExpr.
Built up another version of CopyObject that can copy from memory for
use in the catch parameter code.
RTTI generation for type matching for catch.
Code to check for the type match for catch.
Code to generate the catch parameter, WIP, need make sure references
and pointers and copy ctors work.
llvm-svn: 90205
2009-12-01 03:41:18 +00:00
b1d3f7c909
Have ASTRecordLayout keep track of the key function, in preparation of fixing a synthetic ctor/dtor bug.
...
llvm-svn: 90168
2009-11-30 23:41:22 +00:00
be64149749
More support for virtual destructor calls.
...
Fixes pr5619
llvm-svn: 90158
2009-11-30 22:07:18 +00:00
e5311b0d79
Reflow comments and fix 80-col violation.
...
llvm-svn: 90152
2009-11-30 20:08:49 +00:00
c920fa2a74
Fix an assert.
...
llvm-svn: 90149
2009-11-30 19:43:26 +00:00
2c11cd156e
Don't try to generate common globals for C++ files, instead of depending on the FE to set NoCommon, and simplify CodeGenOptions initialization.
...
llvm-svn: 90119
2009-11-30 08:40:34 +00:00
03aa2f1262
Minor cleanup (no functionality change).
...
llvm-svn: 90105
2009-11-30 01:19:33 +00:00
c76493a65d
Don't pass false (default) for isVolatile parameter to CreateLoad.
...
llvm-svn: 90098
2009-11-29 21:23:36 +00:00
1dd0164bbe
Explicitly use setVolatile instead of extra IRBuilder argument.
...
llvm-svn: 90095
2009-11-29 21:11:47 +00:00
dacbe6bdcd
Don't pass false (default) for isVolatile parameter to CreateStore.
...
llvm-svn: 90094
2009-11-29 21:11:41 +00:00
ba9fd986eb
Simplify code. No functionality change.
...
llvm-svn: 90087
2009-11-29 19:51:45 +00:00
3d221f2fce
Add DeclarationName support for C++0x operator literals. They should now work as
...
function names outside of templates - they'll probably cause some damage there as
they're largely untested.
llvm-svn: 90064
2009-11-29 07:34:05 +00:00
75fa84ebc6
Change LangOptions::ObjCConstantStringClass to an std::string to avoid worrying about the lifetime.
...
llvm-svn: 90052
2009-11-29 02:38:47 +00:00
9eac065e67
Move MainFileName option variable into CodeGenOptions instead of LangOptions.
...
llvm-svn: 90051
2009-11-29 02:38:34 +00:00
337e3a5fea
Remove remaining VISIBILITY_HIDDEN from anonymous namespaces.
...
llvm-svn: 90044
2009-11-28 19:45:26 +00:00
c9e3e6ff29
Tests now pass with the assertion.
...
llvm-svn: 90026
2009-11-28 03:35:45 +00:00
78910a508a
Add a much more thorough test of casts to virtual bases, and fix
...
GetCXXBaseClassOffset to actually pass the test.
llvm-svn: 90025
2009-11-28 03:31:34 +00:00
548cc9d143
We always want to create a virtual function pointer entry if the path from the derived return value to the base overridden return value contains a virtual base class.
...
llvm-svn: 90024
2009-11-28 03:03:52 +00:00
d598850203
Move the vtable builder to an anonymous namespace.
...
llvm-svn: 90021
2009-11-27 22:21:51 +00:00
f942ee0297
Don't build the entire vtable when all we want is the index of a virtual method.
...
llvm-svn: 90017
2009-11-27 20:47:55 +00:00
7eb0ccaa58
Add case for CK_DerivedToBaseMemberPointer cast kind to
...
AggExprEmitter::VisitCastExpr.
llvm-svn: 89988
2009-11-27 04:46:20 +00:00
e96f1d3b2f
More work on ScalarExprEmitter::EmitCastExpr: for every cast kind, either
...
implement it explicitly or assert that it doesn't make sense for a scalar.
This caught a couple interesting issues: one, CK_BaseToDerivedMemberPointer
casts were getting silently miscompiled, and two, Sema was constructing some
strange implicit casts of type CK_UserDefinedConversion.
The change in SemaExprCXX makes sure the cast kinds are getting set correctly.
llvm-svn: 89987
2009-11-27 04:41:50 +00:00
0dfc6804ff
Cleanup for ScalarExprEmitter::EmitCastExpr; no functional changes.
...
llvm-svn: 89986
2009-11-27 02:07:44 +00:00
250534cf9a
Fix for PR5594: use EmitGlobalDefinition instead of EmitCXXDestructor so that
...
we check whether the vtable needs to be generated.
llvm-svn: 89984
2009-11-27 01:42:12 +00:00
323bb04ef4
Lazily create the __cxa_pure_virtual reference.
...
llvm-svn: 89965
2009-11-26 19:54:33 +00:00
d420a31fe9
Use the new CovariantThunkAdjustment in the vtable builder.
...
Make the pure virtual methods map a set instead.
llvm-svn: 89961
2009-11-26 19:32:45 +00:00
334af99964
Remove dead forward declaration. Fix struct/class mismatch. Remove trailing whitespace.
...
llvm-svn: 89947
2009-11-26 13:09:03 +00:00
4907c6f2ed
Remove unused variable.
...
llvm-svn: 89945
2009-11-26 07:48:44 +00:00
ffc066f4b9
Shortcut explicit calls to a trivial copy assignment operator.
...
llvm-svn: 89944
2009-11-26 07:45:48 +00:00
84a7e347bf
Simplify and fix up the handling of implicit constructors, copy assignment
...
operators, and destructors. Avoids generating declarations/definitions of
trivial constructors/destructors, and makes sure the trivial copy assignment
operator is generated when necessary.
llvm-svn: 89943
2009-11-26 07:40:08 +00:00
d15eb34d4b
Implement IRGen for MemberExpr referring to static member function.
...
llvm-svn: 89938
2009-11-26 06:08:14 +00:00
6d771bcd09
Add a Thunk struct to the vtable builder.
...
llvm-svn: 89935
2009-11-26 03:25:13 +00:00
cb422f1dbe
Use the more general Evaluate infrastructure for folding constant DeclRefs
...
instead of checking explicitly for an EnumConstantDecl. This folds references
to constant integer VarDecls, which has two benefits:
1. Slightly smaller emitted code from emitting a constant instead of a load,
and skipping emitting some constant globals.
2. Some code forgets to define static constant member variables; emitting a
load instead of the value in that case leads to a link error. (Such programs
are technically not well-formed, but in practice build with g++.)
llvm-svn: 89934
2009-11-26 03:22:21 +00:00
2f87c4f1a5
Add a CovariantThunkAdjustment struct that represents the adjustments needed for a covariant thunk.
...
llvm-svn: 89933
2009-11-26 03:09:37 +00:00
8a5f75ed5d
Use new getLinkage() method to correctly compute whether a variable has
...
internal linkage. Fixes PR5433.
llvm-svn: 89931
2009-11-26 02:52:12 +00:00
abe274a8b0
Move the mangler into the CodeGen namespace. Change mangleThunk to take a ThunkAdjustment.
...
llvm-svn: 89930
2009-11-26 02:49:32 +00:00
c778540f9a
Add a ThunkAdjustment struct which holds a non-virtual and a virtual adjustment offset. Start using it. General cleanup in Mangle.cpp.
...
llvm-svn: 89925
2009-11-26 02:32:05 +00:00
58bf6e1885
Use StringRef (again) in DebugInfo interface.
...
llvm-svn: 89867
2009-11-25 17:37:31 +00:00
82ba57c8f0
Add VTT parameter to base ctors/dtors with virtual bases. (They aren't used yet).
...
llvm-svn: 89835
2009-11-25 03:15:49 +00:00
3d9e90ae3e
Implement proper cleanup semantics for condition variables in for
...
statements, e.g.,
for(; X x = X(); ) { ... }
Daniel or Anders, please review!
llvm-svn: 89832
2009-11-25 01:51:31 +00:00
7bab5ff8e7
Eliminate CXXConditionDeclExpr with extreme prejudice.
...
All statements that involve conditions can now hold on to a separate
condition declaration (a VarDecl), and will use a DeclRefExpr
referring to that VarDecl for the condition expression. ForStmts now
have such a VarDecl (I'd missed those in previous commits).
Also, since this change reworks the Action interface for
if/while/switch/for, use FullExprArg for the full expressions in those
expressions, to ensure that we're emitting
Note that we are (still) not generating the right cleanups for
condition variables in for statements. That will be a follow-on
commit.
llvm-svn: 89817
2009-11-25 00:27:52 +00:00
680f861d74
Clean up the AST for while loops and fix several problems with
...
cleanups for while loops:
1) Make sure that we destroy the condition variable of a while statement each time through the loop for, e.g.,
while (shared_ptr<WorkInt> p = getWorkItem()) {
// ...
}
2) Make sure that we always enter a new cleanup scope for the body of the while loop, even when there is no compound expression, e.g.,
while (blah)
RAIIObject raii(blah+1);
llvm-svn: 89800
2009-11-24 21:15:44 +00:00
3c9beab48a
Handle references correctly when synthesizing copy constructors.
...
With this change, the clang-on-clang test result is now
Expected Passes : 224
Unexpected Failures: 37
Which means that we can compile over 80% of clang with clang! :)
llvm-svn: 89799
2009-11-24 21:08:10 +00:00
3a202f601e
Handle cases where we're constructing an array of objects and the constructor has default arguments.
...
llvm-svn: 89783
2009-11-24 18:43:52 +00:00
dcf1962405
Explicitly store the condition variable within switch statements, and
...
make sure that this variable is destroyed when we exit the switch
statement.
llvm-svn: 89776
2009-11-24 17:07:59 +00:00
b9f96a3286
We always need to emit the base expression of a member expression, even when the member decl refers to an enum. Thanks to Eli for pointing this out!
...
llvm-svn: 89775
2009-11-24 16:57:33 +00:00
965f450d42
Introduce cleanup scopes for "if" statements in two places:
...
- Outside the "if", to ensure that we destroy the condition variable
at the end of the "if" statement rather than at the end of the
block containing the "if" statement.
- Inside the "then" and "else" branches, so that we emit then- or
else-local cleanups at the end of the corresponding block when the
block is not a compound statement.
To make adding these new cleanup scopes easier (and since
switch/do/while will all need the same treatment), added the
CleanupScope RAII object to introduce a new cleanup scope and make
sure it gets cleaned up.
llvm-svn: 89773
2009-11-24 16:43:22 +00:00
48a409eb39
Rename CleanupScope -> DelayedCleanupBlock. No functionality change.
...
llvm-svn: 89769
2009-11-24 16:21:10 +00:00
59486a2ddb
Get rid of the ugly CGCXX names and replace them with CGClass, CGExprCXX and CGTemporaries.
...
llvm-svn: 89742
2009-11-24 05:51:11 +00:00
bd97548f35
When mangling a ctor/dtor we need to take into consideration whether it's a member template.
...
llvm-svn: 89741
2009-11-24 05:36:32 +00:00
0ade9715f2
Ignore constructor member templates in CodeGenModule::EmitTopLevelDecl.
...
llvm-svn: 89737
2009-11-24 05:16:24 +00:00
6445773279
It is common for vtables to contain pointers to functions that have either incomplete return types or incomplete argument types.
...
Handle this by returning the llvm::OpaqueType for those cases, which CodeGenModule::GetOrCreateLLVMFunction knows about, and treats as being an "incomplete function".
llvm-svn: 89736
2009-11-24 05:08:52 +00:00
633caca353
Explicitly track the condition variable within an "if" statement,
...
rather than burying it in a CXXConditionDeclExpr (that occassionally
hides behind implicit conversions). Similar changes for
switch, while, and do-while will follow, then the removal of
CXXConditionDeclExpr. This commit is the canary.
llvm-svn: 89717
2009-11-23 23:44:04 +00:00
f2e4f72f03
Don't try to treat an enum constant as an lvalue.
...
llvm-svn: 89705
2009-11-23 21:56:41 +00:00
7fa434c2b8
Handle converting member pointers to bool.
...
llvm-svn: 89692
2009-11-23 20:04:44 +00:00
8c793172c2
Handle base-to-derived casts. Will land test case shortly.
...
llvm-svn: 89678
2009-11-23 17:57:54 +00:00
ed6c744091
Centralize and complete the computation of value- and type-dependence for DeclRefExprs
...
llvm-svn: 89649
2009-11-23 11:41:28 +00:00
785406bd19
Teach CodeGenFunction::EmitDecl to ignore Using and UsingShadow decls.
...
llvm-svn: 89633
2009-11-23 00:07:06 +00:00
e1cd0152a9
x86_64, PR5582: Layout bases for C++ records.
...
- Ideally we would have an single iteration interface for this, but this works
for now.
llvm-svn: 89632
2009-11-22 23:01:23 +00:00
5838108317
Use EmitStoreOfScalar when copying the scalar to the space allocated by 'new'.
...
llvm-svn: 89613
2009-11-22 19:25:33 +00:00
e18e29eeb9
Support emitting aggregate class initializers. Fixes PR5581.
...
llvm-svn: 89569
2009-11-21 23:56:04 +00:00
252ba5fb6f
This patch implements objective-c's 'SEL' type as a built-in
...
type and fixes a long-standing code gen. crash reported in
at least two PRs and a radar. (radar 7405040 and pr5025).
There are couple of remaining issues that I would like for
Ted. and Doug to look at:
Ted, please look at failure in Analysis/MissingDealloc.m.
I have temporarily added an expected-warning to make the
test pass. This tests has a declaration of 'SEL' type which
may not co-exist with the new changes.
Doug, please look at a FIXME in PCHWriter.cpp/PCHReader.cpp.
I think the changes which I have ifdef'ed out are correct. They
need be considered for in a few Indexer/PCH test cases.
llvm-svn: 89561
2009-11-21 19:53:08 +00:00
88ad4c5d02
Mangler: Sketch mangling for TemplateArgument::Declaration kind.
...
- Several important FIXMEs related to whether arguments are expressions or external names, and the mangling of extern "C" names (c.f., PR5522).
llvm-svn: 89556
2009-11-21 09:17:15 +00:00
e949e6cced
Mangler: Lift shouldMangleDeclName predicate out of CXXNameMangler::mangle.
...
- Sometimes we have to mangle things we wouldn't normally (e.g., because they appear in a template expression).
- This also tidies up the predicate to be more obvious what is getting mangled.
llvm-svn: 89555
2009-11-21 09:14:52 +00:00
161ade4455
Mangler: Strengthen invariants, MangleContext::mangleName should only be called on var or function decls.
...
llvm-svn: 89554
2009-11-21 09:14:44 +00:00
8483d211bd
Mangler: Inline a bunch of functions into their sole caller, not that I don't love typing for typings sake, but...
...
llvm-svn: 89553
2009-11-21 09:12:13 +00:00
1110dc06a8
Remove dead variable.
...
llvm-svn: 89551
2009-11-21 09:11:59 +00:00
fdae6f7cd3
Mangler: Split isStdNamespace for when the caller already has a NamespaceDecl.
...
llvm-svn: 89550
2009-11-21 09:11:45 +00:00
4118ec5571
IRgen: Eliminate CXXNameMangler::mangleCXX{C,D}tor.
...
llvm-svn: 89549
2009-11-21 09:06:31 +00:00
e128dd18ab
IRgen: Switch the C++ mangler interfaces to take the SmallVector to write into,
...
instead of requiring clients to make a raw_svector_ostream, which is just an
implementation detail.
llvm-svn: 89548
2009-11-21 09:06:22 +00:00
ef5d75afeb
Sink free mangle* methods into MangleContext.
...
llvm-svn: 89547
2009-11-21 09:06:10 +00:00
d614e326db
Delete trailing space.
...
llvm-svn: 89545
2009-11-21 09:05:47 +00:00
96d5c76498
Added rudimentary C++0x attribute support.
...
The following attributes are currently supported in C++0x attribute
lists (and in GNU ones as well):
- align() - semantics believed to be conformant to n3000, except for
redeclarations and what entities it may apply to
- final - semantics believed to be conformant to CWG issue 817's proposed
wording, except for redeclarations
- noreturn - semantics believed to be conformant to n3000, except for
redeclarations
- carries_dependency - currently ignored (this is an optimization hint)
llvm-svn: 89543
2009-11-21 08:43:09 +00:00
58ef18b38c
Checkpoint current work. WIP.
...
llvm-svn: 89513
2009-11-20 23:44:51 +00:00
0a66c26191
Fix lifetime of conditional temporaries. Patch by Victor Zverovich!
...
llvm-svn: 89467
2009-11-20 17:27:56 +00:00
bc8bdea183
Fixed crash when using undefined protocols (GNU runtime).
...
llvm-svn: 89457
2009-11-20 14:50:59 +00:00
564048e692
Fix a couple minor memory leaks.
...
llvm-svn: 89450
2009-11-20 05:53:06 +00:00
d8d26d06a5
Implement throw d, where d is a class type that requires copy
...
construction. WIP.
llvm-svn: 89442
2009-11-20 02:31:07 +00:00
4d7a07bdaf
Handle throw d, where d is a class type but only has a trivial copy
...
constructor. WIP.
llvm-svn: 89438
2009-11-20 01:57:39 +00:00
d1782cc478
Add suport for throw;. WIP.
...
llvm-svn: 89424
2009-11-20 00:56:31 +00:00
18fc44dfd3
Fix rtti generation for throws. WIP.
...
llvm-svn: 89420
2009-11-20 00:43:57 +00:00
101f052faf
Simplify rtti building code a little. Prep for reuse for throw rtti
...
generation.
llvm-svn: 89416
2009-11-20 00:31:50 +00:00
a7a1b7e780
Fixup key function calculations.
...
llvm-svn: 89412
2009-11-20 00:02:19 +00:00
8cd7813ca3
Draw a brighter line between "unresolved" expressions, where we have done the
...
appropriate lookup and simply can't resolve the referrent yet, and
"dependent scope" expressions, where we can't do the lookup yet because the
entity we need to look into is a dependent type.
llvm-svn: 89402
2009-11-19 22:55:06 +00:00
cd2b821241
Fixup address point computations. WIP.
...
llvm-svn: 89386
2009-11-19 20:52:19 +00:00
aa51ad6815
Refine vtable, rtti and rtti name instantiation so that they follow
...
the key function. All the code is wired up, but won't work yet, as I
had to turn off key function calculation as it doesn't work yet.
Also, we refine visibility of the vtable, rtti and rtti name to match
the class, as well as as arrange for all the symbols to be internal
for anonymous namespace entities.
llvm-svn: 89309
2009-11-19 04:04:36 +00:00
beab381d2d
This doesn't work yet.
...
llvm-svn: 89307
2009-11-19 03:53:54 +00:00
1a139f8fb5
Improve instantiation control for rtti data and allow key functions to
...
instantiate a class. WIP.
llvm-svn: 89289
2009-11-19 01:08:19 +00:00
1960b20403
Refine linkage on thunks. WIP.
...
llvm-svn: 89287
2009-11-19 00:49:05 +00:00
ea950e249f
Reflow to fit 80-col.
...
llvm-svn: 89222
2009-11-18 18:57:56 +00:00
643818bdd8
Trim whitespace.
...
llvm-svn: 89219
2009-11-18 18:52:31 +00:00
acfb1dfa4d
Minor cleanup to member pointer handling.
...
llvm-svn: 89201
2009-11-18 09:41:26 +00:00
928a567af1
Clean up EmitPointerToDataMemberBinaryExpr a bit.
...
llvm-svn: 89197
2009-11-18 05:01:17 +00:00
c0f632d6b2
Set up vtable visibility appropriately.
...
llvm-svn: 89193
2009-11-18 04:00:48 +00:00
83d5e00a96
Propagate hidden to the _ZTS symbols appropriately. WIP.
...
llvm-svn: 89191
2009-11-18 03:46:51 +00:00
794290ca29
Add an assertion to catch bad calls to EmitCallArgs.
...
llvm-svn: 89190
2009-11-18 03:42:04 +00:00
c5d2ed73ff
Be sure to set visibility for ZTI symbols for classes from the class.
...
llvm-svn: 89186
2009-11-18 03:21:29 +00:00
f77aaa0089
Make _ZTI symbols hidden. This speeds up the dynamic linker.
...
llvm-svn: 89185
2009-11-18 02:53:28 +00:00
24f55430c8
Missing piece of r89173.
...
llvm-svn: 89174
2009-11-18 00:57:03 +00:00
fe81e3fb3c
Refactor emitting call to delete operator into common function EmitDeleteCall.
...
llvm-svn: 89173
2009-11-18 00:50:08 +00:00
5bbbb13765
Fix one last gotcha with typeid.
...
llvm-svn: 89172
2009-11-18 00:40:12 +00:00
4aaf79a942
Cleanup.
...
llvm-svn: 89166
2009-11-17 23:51:10 +00:00
103a084cbe
Add rtti support for arrays, functiond without prototypes, vectors and
...
enums.
llvm-svn: 89165
2009-11-17 23:45:57 +00:00
96b96d5f2e
Add rtti info for function prototypes and refactor. This allows
...
pointer to member functions to work. WIP.
llvm-svn: 89161
2009-11-17 23:11:22 +00:00
95ace55701
This patch finalizes implementatin of weak_import
...
objective-c2 classes (radar 6815425).
llvm-svn: 89157
2009-11-17 22:42:00 +00:00
6fdfea67ad
Add rtti support for pointer to data members.
...
llvm-svn: 89155
2009-11-17 22:33:00 +00:00
3ad8dcfd55
More cases for weak_import objective-c2 classes.
...
(still radar 6815425).
llvm-svn: 89152
2009-11-17 22:02:21 +00:00
db72c89eb1
Add rtti support for non-member pointers. WIP.
...
llvm-svn: 89148
2009-11-17 21:44:24 +00:00
6726055375
More support for weak_import objective-c2 class.
...
(radar 6815425).
llvm-svn: 89146
2009-11-17 21:37:35 +00:00
950a9518b1
Added block type introspection support.
...
As per Fariborz's suggestion, committed now but can be reverted later if the used flag is problematic for Apple.
llvm-svn: 89134
2009-11-17 19:33:30 +00:00
cb1b7bf79d
Fixed bug where ivar offsets were being initialized as 0 with the fragile GNU ABI.
...
llvm-svn: 89133
2009-11-17 19:32:15 +00:00
1e9de059cc
Skip over shadow using decls during codegen.
...
llvm-svn: 89079
2009-11-17 09:33:40 +00:00
dee9a30204
Unify the way destructor epilogues are generated for synthesized and regular destructors. Also fix PR5529.
...
llvm-svn: 89034
2009-11-17 04:44:12 +00:00
cfd3012756
More const is always good.
...
llvm-svn: 89033
2009-11-17 03:57:07 +00:00
3afea1d8fe
Fix typo. Thanks Ted.
...
llvm-svn: 89031
2009-11-17 03:01:03 +00:00
8f5e6770e7
Add typeinfo support for T* and const T* for all builtin types T.
...
llvm-svn: 89030
2009-11-17 02:57:13 +00:00
3f75d552a3
Add typeid for the builtin types. WIP.
...
llvm-svn: 89028
2009-11-17 02:16:21 +00:00
c2c03349f3
Ensure we peer through () when handling typeid(*p).
...
llvm-svn: 89015
2009-11-17 00:45:21 +00:00
4d0e9095e4
Since we always have 2 edges, we don't need to reserve 3 slot for the
...
PHI node.
llvm-svn: 89005
2009-11-17 00:10:05 +00:00
e8cdcc9ab0
Minor CFG refinements for typeid and dynamic_cast.
...
llvm-svn: 89004
2009-11-17 00:08:50 +00:00
e85ef718d5
Fix up EmitMemberInitializer to handle many more cases.
...
llvm-svn: 88999
2009-11-16 23:53:01 +00:00
5e7d969526
Reorganize EmitMemberInitializer to put anonymous unions on the common codepath.
...
llvm-svn: 88995
2009-11-16 23:34:11 +00:00
c1daba3ec8
Make member initializers for union members work correctly.
...
llvm-svn: 88989
2009-11-16 22:58:01 +00:00
6ca0e21de4
Implement dynamic_cast<void*>(E).
...
llvm-svn: 88988
2009-11-16 22:52:20 +00:00
c2ef215bda
Implement a few more cases for copy constructor synthesis.
...
llvm-svn: 88971
2009-11-16 21:47:41 +00:00
1b8fe5b716
First part of changes to eliminate problems with cv-qualifiers and
...
sugared types. The basic problem is that our qualifier accessors
(getQualifiers, getCVRQualifiers, isConstQualified, etc.) only look at
the current QualType and not at any qualifiers that come from sugared
types, meaning that we won't see these qualifiers through, e.g.,
typedefs:
typedef const int CInt;
typedef CInt Self;
Self.isConstQualified() currently returns false!
Various bugs (e.g., PR5383) have cropped up all over the front end due
to such problems. I'm addressing this problem by splitting each
qualifier accessor into two versions:
- the "local" version only returns qualifiers on this particular
QualType instance
- the "normal" version that will eventually combine qualifiers from this
QualType instance with the qualifiers on the canonical type to
produce the full set of qualifiers.
This commit adds the local versions and switches a few callers from
the "normal" version (e.g., isConstQualified) over to the "local"
version (e.g., isLocalConstQualified) when that is the right thing to
do, e.g., because we're printing or serializing the qualifiers. Also,
switch a bunch of
Context.getCanonicalType(T1).getUnqualifiedType() == Context.getCanonicalType(T2).getQualifiedType()
expressions over to
Context.hasSameUnqualifiedType(T1, T2)
llvm-svn: 88969
2009-11-16 21:35:15 +00:00
c08bdea63d
Clean up scalar cast kind handling; make cast kind handling explicitly handle
...
more cases. No intended visible change.
llvm-svn: 88968
2009-11-16 21:33:53 +00:00
bf44533846
Audit done, all the required casts are already done.
...
llvm-svn: 88966
2009-11-16 21:22:19 +00:00
37b4b8b1da
revert r88963.
...
llvm-svn: 88965
2009-11-16 21:17:07 +00:00
b71c28db22
Use TrackingVH to hold forward decl. This one is for RecordType.
...
llvm-svn: 88963
2009-11-16 21:06:35 +00:00
00dbf4c3c6
Parallel fix to r88951: use TrackingVH to hold forward decl.
...
llvm-svn: 88962
2009-11-16 21:04:30 +00:00
b05d0824c6
Fix valgrind uninitialized error.
...
llvm-svn: 88952
2009-11-16 20:33:31 +00:00
10909d5faf
Use TrackingVH to hold forward decl.
...
llvm-svn: 88951
2009-11-16 20:09:38 +00:00
b3b44ce433
Fixed two minor differences between clang and GCC-generated runtime structures for the GNU runtime.
...
llvm-svn: 88937
2009-11-16 19:05:54 +00:00
0d82fa66a5
The ssp and sspreq function attributes should only be applied to function definitions, not declarations or calls.
...
llvm-svn: 88915
2009-11-16 16:56:03 +00:00
7876dad8e7
Pass a value for the isSigned parameter of CreateIntCast, rather than
...
passing the name (an exotic way of specifying that the result is signed!).
llvm-svn: 88909
2009-11-16 13:11:21 +00:00
65511704f7
Implement most of dynamic_cast. WIP.
...
llvm-svn: 88901
2009-11-16 06:50:58 +00:00
49ddc5fb94
Make GetAddrOfConstantStringFromLiteral return a constant of the correct type.
...
This doesn't have any visible effects at the moment because normally the
implicit cast code forces the type to the expected type.
llvm-svn: 88896
2009-11-16 05:55:46 +00:00
8c98dffd1f
Some minor cleanup for EmitCastLValue.
...
llvm-svn: 88894
2009-11-16 05:48:01 +00:00
e6ce354795
Fix a couple of cases where we weren't generating the right kind of call
...
for a call to a virtual function.
llvm-svn: 88891
2009-11-16 05:31:29 +00:00
011c4c341b
Implement two-argument form of delete operator.
...
llvm-svn: 88890
2009-11-16 05:16:40 +00:00
d787204e6d
When generating the deleting ctor, emit a call to delete.
...
llvm-svn: 88878
2009-11-15 23:03:25 +00:00
4ac39ef6f9
Peer through refernces for typeid. WIP.
...
llvm-svn: 88871
2009-11-15 20:30:39 +00:00
1bf924b121
Finish off zero check for typeid(*p) so that it will do a __cxa_bad_typeid.
...
llvm-svn: 88852
2009-11-15 16:52:53 +00:00
c9b231c8d1
Implement typeid for class types.
...
llvm-svn: 88843
2009-11-15 08:09:41 +00:00
4c808dfc58
Finish off support for typeinfo generation for classes.
...
llvm-svn: 88828
2009-11-15 03:28:10 +00:00
f5b2869b75
Finisgh off rest of class_type_info rtti generation.
...
llvm-svn: 88823
2009-11-14 23:32:21 +00:00
50cb321fdc
Handle CK_BitCast in EmitCastLValue.
...
llvm-svn: 88810
2009-11-14 21:21:42 +00:00
0acee6e0d7
Canonicalize the type before trying to create a debug type.
...
llvm-svn: 88808
2009-11-14 21:08:12 +00:00
6037e78149
Have CGDebugInfo::getOrCreateType cache the QualType instead of having every ConvertType overload do it.
...
llvm-svn: 88807
2009-11-14 20:52:05 +00:00
1acec6a41a
Build up more of the rtti info for a class. WIP.
...
llvm-svn: 88795
2009-11-14 15:55:18 +00:00
14718425fa
Add the name to the rtti data structure.
...
llvm-svn: 88792
2009-11-14 14:25:18 +00:00
274ab904d0
Avoid assert-crash in a case where the expression passed to EmitConstantExpr
...
legitimately has side-effects (and needs to be generated as a non-constant).
llvm-svn: 88767
2009-11-14 08:51:33 +00:00
b210fc598f
Make __func__ and friends work correctly within the initializer for a static
...
local variable.
llvm-svn: 88766
2009-11-14 08:37:13 +00:00
b572c92674
PR5483: Generate missing form of destructor when it is virtual. (Someone
...
more familiar with this stuff should double-check that there isn't some more
general rule; this is purely from inspecting g++ output.)
llvm-svn: 88755
2009-11-14 04:19:37 +00:00
da2efb091d
Generate the old API when sending message to super
...
in a category implementation (objc 32bit api related).
llvm-svn: 88741
2009-11-14 02:18:31 +00:00
52ce3bbf57
Handle CXXDefaultArgExprs in EmitLValue. Fixes PR5484.
...
llvm-svn: 88735
2009-11-14 01:51:50 +00:00
2ec5dd7160
Mangling support for typeinfo names.
...
llvm-svn: 88726
2009-11-14 00:14:13 +00:00
c5a332cefb
Handle descructor printing better.
...
llvm-svn: 88723
2009-11-13 23:45:53 +00:00
559387fe8b
Also track address points for primaries bases.
...
llvm-svn: 88717
2009-11-13 23:13:20 +00:00
ebea005812
Code gen. For virtual destructor call on array objects
...
(still part of pr5472).
llvm-svn: 88712
2009-11-13 22:29:45 +00:00
6f04b1e605
Obvious fix for PR5474.
...
llvm-svn: 88696
2009-11-13 21:23:46 +00:00
f75c1f7a73
Fixes a code gen. bug for array delete operator call
...
int 32bit abi (pr5472 related).
-This line, and those below, will be ignored--
M lib/CodeGen/CGCXXExpr.cpp
llvm-svn: 88695
2009-11-13 21:20:14 +00:00
6814eaa2cc
Code gen for arrady delete operator. Fixes pr5472.
...
llvm-svn: 88680
2009-11-13 19:27:47 +00:00
b40f295037
Do not store DIDescriptor directly into a container. Store MDNode directly, through TrackingVH.
...
llvm-svn: 88677
2009-11-13 19:10:24 +00:00
88fc7d4202
This falls into the category of stupid pet tricks. I hate to do this,
...
but this is necessary to continue work on virtual vtables. We don't
want to penalize virtual table building testcases, just because
complex virtual conversions don't yet work.
llvm-svn: 88676
2009-11-13 18:53:35 +00:00
fb4dda4fed
Instead of storing CXXMethodDecls in the vtable builder, store GlobalDecls so we can represent both the complete and deleting destructors. Also, when encountering a destructor decl, emit entries for both the complete and deleting destructors. Mike, please review.
...
With this change, FileCheck builds and runs the clang test suite without failures!
llvm-svn: 88663
2009-11-13 17:08:56 +00:00
1b4441915a
Wherein the TargetInfo argument to Preprocessor is made 'const' and propogated.
...
llvm-svn: 87087
2009-11-13 05:51:54 +00:00
1721ef7a53
Include header for printf.
...
llvm-svn: 87086
2009-11-13 05:46:16 +00:00
e828c36933
Add a special BuildVirtualCall that's going to be used for building calls to destructors. This is needed because when compiling:
...
struct A {
virtual ~A();
};
void f(A* a) {
delete a;
}
A's deleting destructor should be called.
llvm-svn: 87083
2009-11-13 04:45:41 +00:00
78cfaa9e56
Fix two bugs with temporaries:
...
1. For
A f() {
return A();
}
we were incorrectly calling the A destructor on the returned object.
2. For
void f(A);
void g() {
A a;
f(a);
}
we were incorrectly not calling the copy constructor.
llvm-svn: 87082
2009-11-13 04:34:45 +00:00
87f84c1e72
Move GlobalDecl to its own file. Also add DenseMapInfo traits.
...
llvm-svn: 87081
2009-11-13 04:25:07 +00:00
fa81808a11
Refine construction vtables; they don't include bits that don't have
...
virtual bases unless they are morally virtual.
llvm-svn: 87071
2009-11-13 02:35:38 +00:00
653d0b99df
Refine the construction vtables with respect to offsets. WIP.
...
llvm-svn: 87067
2009-11-13 02:13:54 +00:00
83066c8dee
Allow the tracking of address points for construction vtables as well.
...
llvm-svn: 87063
2009-11-13 01:54:23 +00:00
2b34bc5a96
Refine which vtbl is refernced in VTTs.
...
llvm-svn: 87043
2009-11-12 23:36:21 +00:00
ca0de33113
Refine offsets into vtables for the VTT.
...
llvm-svn: 87041
2009-11-12 23:14:15 +00:00
8677bc27bf
Refine vtable pointers for secondary vtables inside VTTs to point to
...
the right base vtable. WIP.
llvm-svn: 87039
2009-11-12 22:56:32 +00:00
2cefe38365
Fix the offset calculations for non-virtual bases with overrides.
...
Refine the VTT entries for virtual bases to refer to the complete
object's vtable instead of constructor vtables.
Refine the AddressPoint calculations for VTT entries for virtual bases.
llvm-svn: 87021
2009-11-12 20:47:57 +00:00
eb80c98a72
Fix a code gen bug in i386-apple-darwin (objc fragile abi), sending
...
message to 'super'. Fixes radar 7205866.
llvm-svn: 87017
2009-11-12 20:14:24 +00:00
94f798c079
"Attach debug info with llvm instructions" mode was enabled a month ago. Now make it permanent and remove old way of inserting intrinsics to encode debug info for locations and types.
...
llvm-svn: 87007
2009-11-12 18:21:39 +00:00
6e98d7af22
Use getNameAsCString() instead of getName().data()
...
llvm-svn: 87001
2009-11-12 17:49:47 +00:00
bc55fe26c6
Move CompileOptions -> CodeGenOptions, and sink it into the CodeGen library.
...
This resolves the layering violation where CodeGen depended on Frontend.
llvm-svn: 86998
2009-11-12 17:24:48 +00:00
7bdf096f8d
Do not use StringRef while using DebugInfo interface.
...
llvm-svn: 86915
2009-11-12 00:51:46 +00:00
fd84d6d808
Fix ctor vtable name. WIP.
...
llvm-svn: 86879
2009-11-11 20:32:03 +00:00
eac4559790
Push ctor vtable construction down further. WIP.
...
llvm-svn: 86878
2009-11-11 20:26:26 +00:00
5348515193
Do not eagerly set stop point for arguments. This misleads the debugger in identifying beginning of function body. Instead, create new location to attach with llvm.dbg.declare. This location is only used to find the context of the variable by the code generator, and it is not used to emit line number info.
...
llvm-svn: 86862
2009-11-11 19:10:19 +00:00
03f62ed9bb
Value initialize non-class array members in ctor's
...
initializer list. Fixes PR5463.
llvm-svn: 86849
2009-11-11 17:55:25 +00:00
c7b9f5e762
More VTT work. We now track offsets and use the ctor vtable builder
...
interface. WIP.
llvm-svn: 86793
2009-11-11 03:08:24 +00:00
e6c8112572
Avoid generating additional destructor(s) for initialized constructed
...
objects.
llvm-svn: 86778
2009-11-11 01:13:34 +00:00
8b2d2d0707
Stub out most of the VTT building code. WIP.
...
llvm-svn: 86772
2009-11-11 00:35:07 +00:00
af993bf5cc
Attach location info with llvm.dbg.declare.
...
llvm-svn: 86750
2009-11-10 23:07:24 +00:00
d76f438369
Revert r86741; it doesn't handle virtual bases correctly.
...
llvm-svn: 86745
2009-11-10 22:48:10 +00:00
0960059c0c
Minor optimization to skip null check in generated code where it isn't
...
necessary.
llvm-svn: 86741
2009-11-10 22:42:26 +00:00
3ad2645430
Fix for PR5454: make sure to use the right block as the predecessor in the
...
generated PHI node for the null check of a new operator.
llvm-svn: 86738
2009-11-10 22:39:09 +00:00
1254a0978c
This patch implements Code gen. for destruction of
...
global array of objects.
llvm-svn: 86701
2009-11-10 19:24:06 +00:00
b47222144b
Only generate a VTT for classes that need a VTT.
...
llvm-svn: 86699
2009-11-10 19:13:04 +00:00
d846d0825b
Add vtable caching to prevent multiple vtables for the same class from
...
being generated.
Add the most derived vtable pointer to the VTT.
llvm-svn: 86671
2009-11-10 07:44:33 +00:00
9f23a149cb
Be sure to clear out VCall when we clear out VCalls.
...
Start implementing VTTs. WIP.
llvm-svn: 86650
2009-11-10 02:30:51 +00:00
3b9176985d
Finish off mangling for the VTT.
...
llvm-svn: 86647
2009-11-10 01:58:37 +00:00
ef15744092
Add mangling for the construction vtable.
...
llvm-svn: 86643
2009-11-10 01:41:59 +00:00
612e38026a
Fix clang's use of DenseMap iterators after r86636 fixed their constness.
...
Patch by Victor Zverovich!
llvm-svn: 86638
2009-11-10 01:17:45 +00:00
68ec7ee0e8
Enable the use of the new llvm objectsize intrinsic.
...
llvm-svn: 86607
2009-11-09 22:40:09 +00:00
4ef1d400d9
Make sure that Type::getAs<ArrayType>() (or Type::getAs<subclass of
...
ArrayType>()) does not instantiate. Update all callers that used this
unsafe feature to use the appropriate ASTContext::getAs*ArrayType method.
llvm-svn: 86596
2009-11-09 22:08:55 +00:00
cc522d9ab3
Rearrange function to avoid recursive use-after-free.
...
llvm-svn: 86516
2009-11-09 05:07:37 +00:00
a72bf0fae8
Explicitly note that pre-inc/dec lvalues are not supported yet, so that it
...
doesn't crash. (Such expressions are valid in C++, but not in C.)
llvm-svn: 86513
2009-11-09 04:20:47 +00:00
11c08c8e5b
Rename areAllFields32Or64BitBasicType to canExpandIndirectArgument to closer match what it is semantically used for.
...
Also, fix a major bug where fields from a C++ struct might be dropped -- the expand action doesn't handle them correctly yet.
llvm-svn: 86502
2009-11-09 01:33:53 +00:00
d7686ef31c
Unify the codepaths used to verify base and member initializers for explicitly
...
and implicitly defined constructors. This has a number of benefits:
1. Less code.
2. Explicit and implicit constructors get the same diagnostics.
3. The AST explicitly contains constructor calls from implicit default
constructors. This allows handing some cases that previously weren't handled
correctly in IRGen without any additional code. Specifically, implicit default
constructors containing calls to constructors with default arguments are now
handled correctly.
llvm-svn: 86500
2009-11-09 01:05:47 +00:00
7e215ea165
Add clarifying parens.
...
llvm-svn: 86457
2009-11-08 09:46:46 +00:00
5bbdc9f857
Handle member expressions where the member declaration is actually a static variable. Fixes PR5392.
...
llvm-svn: 86414
2009-11-07 23:16:50 +00:00
ea4c30b39a
More LValue related code cleanup.
...
llvm-svn: 86413
2009-11-07 23:06:58 +00:00
2ff6395ddc
More cleanup, the code is much easier to follow now.
...
llvm-svn: 86412
2009-11-07 22:53:10 +00:00
6eee97276b
Reduce nesting, no functionality change.
...
llvm-svn: 86411
2009-11-07 22:46:42 +00:00
03f83e58f7
We only need to call SetObjCNonGC for local variables. No functionality change.
...
llvm-svn: 86410
2009-11-07 22:43:34 +00:00
509850ebec
Change EmitPointerToDataMemberLValue to take a FieldDecl. No intended functionality change.
...
llvm-svn: 86407
2009-11-07 22:00:15 +00:00
15275e55ce
add missing #include
...
llvm-svn: 86368
2009-11-07 09:22:46 +00:00
c0a3561f0f
Always mangle functions with special names. Fixes PR5420.
...
llvm-svn: 86350
2009-11-07 07:15:03 +00:00
ce21431fc4
When looking up and adding substitutions to the substitution table, make sure to always use the canonical declaration. With tihs change, FileCheck compiles and links but crashes during startup.
...
llvm-svn: 86339
2009-11-07 04:26:04 +00:00
3d888e4548
Revert r86315 and add Type::FixedWidthInt to the FIXME cases.
...
llvm-svn: 86320
2009-11-07 01:19:37 +00:00
5b39e2382f
Do not assert if debug info for certain type is not generated.
...
llvm-svn: 86315
2009-11-07 00:29:05 +00:00
eef461c230
Do not emit linkage name for global variables. It confuses gdb, because it picks up AT_MIPS_linkage_name and ignores AT_name.
...
llvm-svn: 86308
2009-11-07 00:10:18 +00:00
cd6f9ed906
Refine the non-virtual this adjustment. Optimize out virtual this
...
adjustments of zero.
llvm-svn: 86300
2009-11-06 23:27:42 +00:00
e8e631c912
Prevent a code gen. crash on empty unions - pr5408.
...
llvm-svn: 86287
2009-11-06 20:47:40 +00:00
443f677302
Simplify the debug info code, handle lvalue references and template specializations.
...
llvm-svn: 86277
2009-11-06 19:19:55 +00:00
c2d71b531c
This patch fixes code gen. part of pr5333 (Conversion
...
using elipsis conversion).
llvm-svn: 86276
2009-11-06 18:47:57 +00:00
8a6512961f
Handle QualifiedNameType and SubstTemplateTypeParmType types in CGDebugInfo::CreateTypeNode.
...
llvm-svn: 86274
2009-11-06 18:45:16 +00:00
25ed5c2f42
Don't assert when trying to generate debug info for vector types. This needs to be fixed eventually...
...
llvm-svn: 86268
2009-11-06 18:24:04 +00:00
353fc1e447
update comment, pointed out by Gabor
...
llvm-svn: 86263
2009-11-06 18:10:47 +00:00
4f24225e8e
indirectbr seems to work! Rip out the old code.
...
llvm-svn: 86256
2009-11-06 17:38:18 +00:00
eb9bc2ba82
Instead of returning a null DIType for unhandled types, assert.
...
llvm-svn: 86254
2009-11-06 17:01:39 +00:00
421a5c9651
If a member variable of reference type is bound to a temporary in its member initializer it needs to be destroyed at the end of the constructor.
...
llvm-svn: 86230
2009-11-06 04:19:02 +00:00
5852b13827
If a member initializer create temporaries we need to destroy them. Fixes PR5077.
...
llvm-svn: 86225
2009-11-06 04:11:09 +00:00
a7cb98b202
More cleanup.
...
llvm-svn: 86224
2009-11-06 03:23:06 +00:00
438cf92533
Cleanup ctor/dtor emission.
...
llvm-svn: 86222
2009-11-06 02:55:43 +00:00
9e4e0237e4
Handle ParenExprs in mangleExpression.
...
llvm-svn: 86218
2009-11-06 02:50:19 +00:00
72431bd9ea
Refine the vcall for a function that is defined in a virtual base
...
class that is overridden in a base that isn't morally virtual.
llvm-svn: 86217
2009-11-06 02:38:24 +00:00
ba3c6681f0
Enable debug info for global variables at -O1+
...
llvm-svn: 86156
2009-11-05 19:13:29 +00:00
4127b8ef97
Added support for static variables which require
...
initialization before main. Fixes pr5396.
llvm-svn: 86145
2009-11-05 18:03:03 +00:00
c550768a11
References can't be zero; omit zero check for return value adjustments
...
in covariant thunks that return references.
llvm-svn: 86121
2009-11-05 06:32:02 +00:00
b8da7a046d
Refine covariant return value adjustments for thunks when null
...
pointers are returned.
llvm-svn: 86120
2009-11-05 06:12:26 +00:00
18bc195473
Add code gen for pointer-to-member function in
...
ctor's initializer. Fixes pr5178.
llvm-svn: 86040
2009-11-04 19:13:37 +00:00
09cc10f9f8
This patch extends CleanupScope to support destruction
...
of array objects on block exit. Patch is by Anders Calrsson.
llvm-svn: 86032
2009-11-04 17:57:40 +00:00
615ac67fb8
Store the unresolved class type in MemberPointerType's Class field,
...
from Peter Collingbourne!
llvm-svn: 86030
2009-11-04 16:49:01 +00:00
f9d41df093
Merge constant array and structures. This will create a global variables for arrays and structs that are constant and their initializer is constant. It is on by default but can be disable with the flag -fno-merge-all-constants.
...
llvm-svn: 85991
2009-11-04 01:18:09 +00:00
b9c9b351e9
Fix 80-col violations.
...
llvm-svn: 85990
2009-11-04 01:11:15 +00:00
71609a2874
Split out return adjustments in thunks from this adjustment in thunks
...
so the optimizer can tailcall into the return value adjustment thunk.
This improves codesize for complex hierarchies.
llvm-svn: 85988
2009-11-04 00:53:51 +00:00
Devang Patel