Commit Graph

1707 Commits

Author SHA1 Message Date
Eli Friedman 13ffdd86b9 Make the ".*" operator work correctly when the base is a prvalue and the field has a non-trivial copy constructor. PR13097.
llvm-svn: 158578
2012-06-15 23:51:06 +00:00
John McCall d8d00be632 It turns out that implementing the rethrow-on-fallthrough
semantics of a ctor/dtor function-try-block catch handler
by pushing a normal cleanup is not just overkill but actually
actively wrong when the handler contains an explicit return
(which is only legal in a dtor).  Just emit the rethrow as
ordinary code at the fallthrough point.  Fixes PR13102.

llvm-svn: 158488
2012-06-15 05:27:05 +00:00
Charles Davis a81618db44 Grab bag of Microsoft Mangler fixes:
- Support mangling virtual function tables (base tables need work on the
  ManglerContext interface).
- Correct mangling of local scopes (i.e. functions and C++ methods).
- Replace every llvm_unreachable() for actually-reachable code with a
  diagnostic.

llvm-svn: 158376
2012-06-13 00:18:14 +00:00
Rafael Espindola a486f48e5b We were computing the visibility and linkage of template parameters, but
only using the linkage.

Use and test both, documenting that considering the visibility and linkage
of template parameters is a difference from gcc.

llvm-svn: 158309
2012-06-11 14:29:58 +00:00
Richard Smith f76568591c PR13047: Fix various abuses of clang::Type in the MS mangler, to make it work
in the presence of type sugar.

llvm-svn: 158184
2012-06-08 00:37:04 +00:00
Eric Christopher beafa44be8 Reapply "Only emit debug information for methods that are user defined, there's"
As the failing testcase has been fixed.

This reverts commit 0637f407e6ee7fdccde17fbf9a5fcc4853187b3e.

llvm-svn: 158009
2012-06-05 18:16:22 +00:00
John McCall 65485d7694 Revert "Only emit debug information for methods that are user defined, there's"
This reverts r157970, which was not passing on
  clang-x86_64-darwin10-nobootstrap-RA

llvm-svn: 157983
2012-06-05 06:10:39 +00:00
Eric Christopher f5dad49a65 Only emit debug information for methods that are user defined, there's
not much reason to emit for constructors and destructors that aren't
user defined.

rdar://11593099

llvm-svn: 157970
2012-06-05 00:15:06 +00:00
Richard Smith 50fba8f0a9 PR13022: cope with parenthesized function types in MS name mangling.
llvm-svn: 157959
2012-06-04 22:46:59 +00:00
Eric Christopher 3f7c44a131 Add a testcase for C++11 union support.
llvm-svn: 157952
2012-06-04 21:32:12 +00:00
Eric Christopher 31a1fc6654 Support C++11 enum forward declarations.
Part of rdar://11570854

llvm-svn: 157787
2012-06-01 00:22:57 +00:00
Charles Davis 6770dea704 Fix mangling of integral template arguments between 1 and 10. Add a test case
for this. Reported by Timur Iskhodzhanov.

llvm-svn: 157583
2012-05-28 16:53:33 +00:00
Charles Davis 7fb195b683 Test case for 157547. Before that patch, all the digits would be mangled
as zeroes. Now the digits are properly non-zero.

llvm-svn: 157552
2012-05-28 00:43:56 +00:00
Charles Davis 5511dfb399 Mangle template instantiations properly (as of VC 7.x) when compiling for
the Microsoft Visual C++ ABI. Currently limited to type and integral
non-type arguments. Based on a patch by Timur Iskhodzhanov!

llvm-svn: 157524
2012-05-26 23:12:19 +00:00
Nuno Lopes f491fbd715 fix codegen support for alloc_size attribute for static C++ methods
add test case for C++ codegen

llvm-svn: 157500
2012-05-25 21:45:08 +00:00
Rafael Espindola 67a498cc5f Don't ignore linkage when ignoring visibility in the instantiation of a
method template.

llvm-svn: 157486
2012-05-25 17:22:33 +00:00
Rafael Espindola 340941dc16 Don't ignore linkage when ignoring visibility in the instantiation of a
function template.

llvm-svn: 157480
2012-05-25 16:41:35 +00:00
Rafael Espindola a28bf63d9a Consider the linkage for member class templates even when we have to ignore
the visibility.

llvm-svn: 157475
2012-05-25 15:51:26 +00:00
Rafael Espindola 0cf10ac9ab When ignoring visibility in an instantiation, still consider the linkage.
Similar fixes for function and member template to follow as I write the
testcases.

llvm-svn: 157470
2012-05-25 14:47:05 +00:00
Eric Christopher 9cc59f66ad Move file to be more representative.
llvm-svn: 157306
2012-05-23 00:18:46 +00:00
Eric Christopher 41f556b709 Check that we have an unknown upper bound range for our flexible array members.
llvm-svn: 157305
2012-05-23 00:18:43 +00:00
Eric Christopher a91af14b8d Emit C++11 enum class information if it exists.
Part of rdar://11496790

llvm-svn: 157304
2012-05-23 00:09:47 +00:00
Eric Christopher 8b3304da56 Test that we emit a subrange type for vlas.
Part of rdar://11457152

llvm-svn: 157269
2012-05-22 18:04:48 +00:00
Rafael Espindola 96dcb8d4bb Function template version of the previous patch.
llvm-svn: 157207
2012-05-21 20:31:27 +00:00
Rafael Espindola 93c289c28a Produce a hidden symbol for zed in
struct HIDDEN foo {
  };
  template <class P>
  struct bar {
  };
  template <>
  struct HIDDEN bar<foo> {
    DEFAULT static void zed();
  };
  void bar<foo>::zed() {
  }

Before we would produce a hidden symbol in

  struct HIDDEN foo {
  };
  template <class P>
  struct bar {
  };
  template <>
  struct bar<foo> {
    DEFAULT static void zed();
  };
  void bar<foo>::zed() {
  }

But adding HIDDEN to the specialization would cause us to produce a default
symbol.

llvm-svn: 157206
2012-05-21 20:15:56 +00:00
Rafael Espindola 85e91ba6f5 Test and document a difference from gcc in the handling of visibility
attributes.

llvm-svn: 157186
2012-05-21 14:22:37 +00:00
Jakob Stoklund Olesen 3923f1d8ca Fix test for 32-bit hosts.
llvm-svn: 157154
2012-05-20 06:43:46 +00:00
Eli Friedman 022900079b Zap a bogus assert for delegating constructors. PR12890, part 2.
I'm pretty sure we are in fact doing the right thing here, but someone who knows the standard better should double-check that we are in fact supposed to zero out the member in the given testcase.

llvm-svn: 157138
2012-05-20 00:46:05 +00:00
Eric Christopher fe0913b026 Update API usage for llvm DIBuilder changes for rvalue reference
types and ensure we are actually creating the type.

rdar://11479676

llvm-svn: 157095
2012-05-19 01:36:50 +00:00
Rafael Espindola 96e6824c31 Fix the visibility of instantiations of static data members.
Fixes pr12835.

llvm-svn: 156897
2012-05-16 02:10:38 +00:00
Rafael Espindola 7f90b7d4c2 Fix our handling of visibility in explicit template instantiations.
* Don't copy the visibility attribute during instantiations. We have to be able
  to distinguish

 struct HIDDEN foo {};
 template<class T>
 DEFAULT void bar() {}
 template DEFAULT void bar<foo>();

from

 struct HIDDEN foo {};
 template<class T>
 DEFAULT void bar() {}
 template void bar<foo>();

* If an instantiation has an attribute, it takes precedence over an attribute
  in the template.

* With instantiation attributes handled with the above logic, we can now
  select the minimum visibility when looking at template arguments.

llvm-svn: 156821
2012-05-15 14:09:55 +00:00
Richard Smith 8e4a3868fe PR12798: Don't drop part of the nested name specifier when instantiating a
pseudo-destructor expression. This can affect whether virtual dispatch for
the destructor call is bypassed.

llvm-svn: 156806
2012-05-15 06:15:11 +00:00
John McCall 8b3f5ffb0a Change the mangling of a ref-qualifier on a function type so that
it is placed in a position which is never ambiguous with a
reference-to-function type.  This follows some recent discussion
and ensuing proposal on cxx-abi-dev.  It is not necessary to
change the mangling of CV-qualifiers because you cannot
apply CV-qualification in the normal sense to a function type.
It is not necessary to change the mangling of ref-qualifiers on
method declarations because they appear in an unambiguous
location.

In addition, mangle CV-qualifiers and ref-qualifiers on function
types when they occur in positions other than member pointers
(that is, when they appear as template arguments).

This is a minor ABI break with previous releases of clang.  It
is not considered critical because (1) ref-qualifiers are
relatively rare, since AFAIK we're the only implementing compiler,
and (2) they're particularly likely to come up in contexts that
do not rely on the ODR for correctness.  We apologize for any
inconvenience;  this is the right thing to do.

llvm-svn: 156794
2012-05-15 02:01:59 +00:00
Richard Smith 48d5418df8 Slightly generalize FileCheck patterns to unbreak -Asserts builds.
llvm-svn: 156783
2012-05-14 22:06:02 +00:00
Richard Smith bb653bd5f9 Implement IRGen for C++11's "T{1, 2, 3}", where T is an aggregate and the
expression is treated as an lvalue.

llvm-svn: 156781
2012-05-14 21:57:21 +00:00
Richard Smith e2648bac3a A union can have a constexpr defaulted default constructor, if it has an
in-class initializer for one of its fields. Value-initialization of such
a type should use the in-class initializer!

The former was just a bug, the latter is a (reported) standard defect.

llvm-svn: 156274
2012-05-07 01:07:30 +00:00
Alexey Samsonov 74a3868dbb This patch adds a new Clang compiler flag "-gline-tables-only".
It reduces the amount of emitted debug information:
1) DIEs in .debug_info have types DW_TAG_compile_unit, DW_TAG_subprogram,
DW_TAG_inlined_subroutine (for opt builds) and DW_TAG_lexical_block only.
2) .debug_str contains only function names.
3) No debug data for types/namespaces/variables is emitted.
4) The data in .debug_line is enough to produce valid stack traces with
function names and line numbers.

Reviewed by Eric Christopher.

llvm-svn: 156160
2012-05-04 07:39:27 +00:00
Chandler Carruth b41171b70b Fix non-deterministic iteration order when walking the specializations
of templates by using the newly introduce FoldingSetVector. This
preserves insertion order for all iteration of specializations.

I've also included a somewhat terrifying testcase that rapidly builds up
a large number of functions. This is enough that any system with ASLR
will have non-deterministic debug information generated for the test
case without the fix here as the debug information is generated in part
by walking these specializations.

llvm-svn: 156133
2012-05-03 23:49:05 +00:00
John McCall 26ebbb748b Merge x86-64-abi-sret-vs-2word-struct-param.cpp into the generic
x86_64-arguments.cpp test file and be sure to test the coerced
case as well.  Thanks to Wei-Ren Chen for bringing this test
to my attention.

llvm-svn: 156047
2012-05-03 01:34:46 +00:00
John McCall c7c96d5ea3 Restructure some of the member-pointer tests, implicitly
removing some that were added without a triple (and moving
their tests into files that appropriately test multiple
architectures).

llvm-svn: 155920
2012-05-01 17:32:20 +00:00
NAKAMURA Takumi 4ad81df818 clang/test/CodeGenCXX: Fix two tests, destructors.cpp and microsoft-abi-array-cookies.cpp, for -Asserts.
llvm-svn: 155913
2012-05-01 11:13:04 +00:00
John McCall c84ed6a336 Abstract the emission of global destructors into ABI-specific code
and only consider using __cxa_atexit in the Itanium logic.  The
default logic is to use atexit().

Emit "guarded" initializers in Microsoft mode unconditionally.
This is definitely not correct, but it's closer to correct than
just not emitting the initializer.

Based on a patch by Timur Iskhodzhanov!

llvm-svn: 155894
2012-05-01 06:13:13 +00:00
David Blaikie 7afed5e5bf PR12710 - broken default argument handling for templates.
I broke this in r155838 by not actually instantiating non-dependent default arg
expressions. The motivation for that change was to avoid producing duplicate
conversion warnings for such default args (we produce them once when we parse
the template - there's no need to produce them at each instantiation) but
without actually instantiating the default arg, things break in weird ways.

Technically, I think we could still get the right diagnostic experience without
the bugs if we instantiated the non-dependent args (for non-dependent params
only) immediately, rather than lazily. But I'm not sure if such a refactoring/
change would be desirable so here's the conservative fix for now.

llvm-svn: 155893
2012-05-01 06:05:57 +00:00
John McCall 7f0245d930 Remove some not-very-stable assumptions from this testcase.
llvm-svn: 155889
2012-05-01 05:29:32 +00:00
John McCall b91cd6687c Refactor the C++ ABI code a little bit to take advantage of
what I'm going to treat as basically universal properties of
array-cookie code.  Implement MS array cookies on top of that.
Based on a patch by Timur Iskhodzhanov!

llvm-svn: 155886
2012-05-01 05:23:51 +00:00
John McCall 23dfaa1cef When mangling a synthetic function declaration, we might not have
type-source information for its parameters.  Don't crash when
mangling them in the MS C++ ABI.  Patch by Timur Iskhodzhanov!

llvm-svn: 155879
2012-05-01 02:33:44 +00:00
Richard Smith d494c500a0 PR12224 (sort of): Diagnose inheriting constructor declarations in C++11 mode.
We do not support IRGen for these, and get some parts of the semantic analysis
wrong.

llvm-svn: 155728
2012-04-27 19:33:05 +00:00
Eli Friedman f927b8bc81 Add a missing check in CodeGen of packed classes with vtables. <rdar://problem/11324125>.
llvm-svn: 155689
2012-04-27 02:34:46 +00:00
Eric Christopher 3cc207bbcc Forward declarations should take a context. This helps the debugger
find forward declarations in the context that the actual definition
will occur.

rdar://11291658

llvm-svn: 155381
2012-04-23 19:00:24 +00:00
Rafael Espindola b522a5f246 Fix visibility when we have two types with explicit visibility in a template
argument list.

llvm-svn: 155368
2012-04-23 17:51:55 +00:00