Commit Graph

1914 Commits

Author SHA1 Message Date
Nico Weber 1598289c72 Add tests that I wrote for r159060 but forgot to `svn add`.
llvm-svn: 159170
2012-06-25 22:30:41 +00:00
Timur Iskhodzhanov 9137c399bd [Windows] Fix mangling of repeated types in the presence of bool and function pointers PR13176,PR13177
llvm-svn: 159059
2012-06-23 00:54:17 +00:00
Charles Davis de2e5ed4aa MicrosoftMangle: Fix mangling of integral constant non-type template arguments in a class specialization.
llvm-svn: 159056
2012-06-23 00:27:49 +00:00
Nuno Lopes c65221decf revert CodeGen support for the alloc_size attribute until we finish the design of a more generic metadata node
llvm-svn: 159016
2012-06-22 18:01:38 +00:00
Rafael Espindola 25fe3191b3 Disable llvm optimizations in a clang test.
llvm-svn: 159012
2012-06-22 17:28:01 +00:00
Richard Smith 73af5c6dda MS: Mangle rvalue references and nullptr_t, and produce back-references when
appropriate. Patch by João Matos!

llvm-svn: 158895
2012-06-21 02:52:27 +00:00
Alexis Hunt 78e2b91319 Do not crash when we dynamic cast a final type to void*.
llvm-svn: 158763
2012-06-19 23:44:55 +00:00
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
Rafael Espindola 23fd21d4ee Add another testcase that was not being covered.
llvm-svn: 155351
2012-04-23 14:02:53 +00:00
Rafael Espindola cba4879534 Add a testcase that we already get right but was not being tested. Found
by a chrome build on OS X. Thanks to Nico Weber for testing a patch and
providing the .ii file.

llvm-svn: 155326
2012-04-22 22:31:25 +00:00
Rafael Espindola 1e762442c7 Cleanup testcase.
llvm-svn: 155317
2012-04-22 16:03:00 +00:00
Rafael Espindola bbc5cbc410 An attribute in a explicit template installation should take precedence over
the tempale arguments in deciding the visibility.

This agrees with gcc 4.7.

Found by trying to build chrome with component=shared_library with 155314
reverted.

llvm-svn: 155316
2012-04-22 15:31:59 +00:00
Rafael Espindola 6b6e879221 Revert 155102 but add a fixme while I debug what is wrong with chrome's
components build.

llvm-svn: 155314
2012-04-22 12:37:27 +00:00
Richard Smith f3c37e89a1 PR12571: Objects of type clang::ConstantArrayType aren't always emitted with
type llvm::ArrayType -- sometimes we emit them as packed structs. Don't assert
if such a global array has an element type with a non-trivial destructor.

llvm-svn: 155305
2012-04-22 05:51:36 +00:00
Rafael Espindola fae3f908d8 Add another missing testcase.
llvm-svn: 155304
2012-04-22 02:41:22 +00:00
Rafael Espindola 4304101fb2 Add a testcase we were already getting right, but were not testing for.
llvm-svn: 155302
2012-04-22 01:22:25 +00:00
Rafael Espindola 8add48ec9e Fix handling of template parameters. Found by inspection. GCC 4.7 agrees
with this testcase.

llvm-svn: 155301
2012-04-22 00:43:48 +00:00
Timur Iskhodzhanov f32a377a7c Fix bug 12574 - Avoid infinite recursion in constructors and destructors when using Microsoft C++ ABI
llvm-svn: 155189
2012-04-20 08:05:00 +00:00
Rafael Espindola 53cf219367 In mergeVisibility, if we already have an explicit visibility, keep it.
This fixes the included testcase and lets us simplify the code a bit. It
does require using mergeWithMin when merging class information to its
members. Expand the comments to explain why that works.

llvm-svn: 155103
2012-04-19 05:50:08 +00:00
Rafael Espindola 1280fe6245 In mergeVisibilityWithMin, let an implicit hidden symbol take precedence over
an explicit default one. This means that with -fvisibility hidden we
now produce a hidden symbol for

template <typename T>
class DEFAULT foo {
  void bar() {}
};
class zed {};
template class foo<zed>;

This matches the behaviour of gcc 4.7.

llvm-svn: 155102
2012-04-19 05:34:51 +00:00
Richard Smith d372942d77 PR 12586: Fix assert while running libc++ testsuite: deal with exception
specifications on member function templates of class templates and other such
nested beasties. Store the function template from which we are to instantiate
an exception specification rather than trying to deduce it. Plus some
additional test cases.

llvm-svn: 155076
2012-04-19 00:08:28 +00:00
Richard Smith 79a52e5709 PR12569: Instantiate exception specifications of explicit instantiations
and explicit specializations of function templates appropriately.

llvm-svn: 154956
2012-04-17 22:30:01 +00:00
Richard Smith f623c96260 Implement DR1330 in C++11 mode, to support libstdc++4.7 which uses it.
We have a new flavor of exception specification, EST_Uninstantiated. A function
type with this exception specification carries a pointer to a FunctionDecl, and
the exception specification for that FunctionDecl is instantiated (if needed)
and used in the place of the function type's exception specification.

When a function template declaration with a non-trivial exception specification
is instantiated, the specialization's exception specification is set to this
new 'uninstantiated' kind rather than being instantiated immediately.

Expr::CanThrow has migrated onto Sema, so it can instantiate exception specs
on-demand. Also, any odr-use of a function triggers the instantiation of its
exception specification (the exception specification could be needed by IRGen).
In passing, fix two places where a DeclRefExpr was created but the corresponding
function was not actually marked odr-used. We used to get away with this, but
don't any more.

Also fix a bug where instantiating an exception specification which refers to
function parameters resulted in a crash. We still have the same bug in default
arguments, which I'll be looking into next.

This, plus a tiny patch to fix libstdc++'s common_type, is enough for clang to
parse (and, in very limited testing, support) all of libstdc++4.7's standard
headers.

llvm-svn: 154886
2012-04-17 00:58:00 +00:00
Rafael Espindola 78158af59c Revert r154749 for now at John McCall's request.
llvm-svn: 154846
2012-04-16 18:46:26 +00:00
Douglas Gregor 3024f07c12 Implement C++11 [expr.prim.general]p3, which permits the use of 'this'
in the declaration of a non-static member function after the
(optional) cv-qualifier-seq, which in practice means in the exception
specification and late-specified return type. 

The new scheme here used to manage 'this' outside of a member function
scope is more general than the Scope-based mechanism previously used
for non-static data member initializers and late-parsesd attributes,
because it can also handle the cv-qualifiers on the member
function. Note, however, that a separate pass is required for static
member functions to determine whether 'this' was used, because we
might not know that we have a static function until after declaration
matching.

Finally, this introduces name mangling for 'this' and for the implicit
'this', which is intended to match GCC's mangling. Independent
verification for the new mangling test case would be appreciated.

Fixes PR10036 and PR12450.

llvm-svn: 154799
2012-04-16 07:05:22 +00:00
Richard Smith 9ec1e48b59 PR12226: don't generate wrong code if a braced string literal is used to
initialize an array of unsigned char. Outside C++11 mode, this bug was benign,
and just resulted in us emitting a constant which was double the required
length, padded with 0s. In C++11, it resulted in us generating an array whose
first element was something like  i8 ptrtoint ([n x i8]* @str to i8).

llvm-svn: 154756
2012-04-15 02:50:59 +00:00
Rafael Espindola 0aec8ece52 Consider visibility attributes last, so that they take precedence.
I am working on a cleaner fix, but this gets the case in PR12552 passing.

llvm-svn: 154749
2012-04-14 15:21:19 +00:00
John McCall 8c38d35b05 Don't enter cleanups for unreachable variables. It's impossible to
jump into these scopes, and the cleanup-entering code sometimes wants
to do some operations first (e.g. a GEP), which can leave us with
unparented IR.

llvm-svn: 154684
2012-04-13 18:44:05 +00:00
John McCall 32233a7c1a Fix a trivial oversight with apple-kext static local destructors
and add a test case.

llvm-svn: 154653
2012-04-13 02:53:27 +00:00
Douglas Gregor 298f43df14 Fix some i1/i8 confusion within _Atomic(bool) in IR generation, both
in general (such an atomic has boolean representation) and
specifically for IR generation of __c11_atomic_init. The latter also
means actually using initialization semantics for this initialization,
rather than just creating a store.

On a related note, make sure we actually put in non-atomic-to-atomic
conversions when performing an implicit conversion sequence. IR
generation is far too kind here, but we still want the ASTs to make
sense.

llvm-svn: 154612
2012-04-12 20:42:30 +00:00
Eric Christopher 5d7fe46ceb Testcase for previous commit.
rdar://11079003

llvm-svn: 154571
2012-04-12 00:36:10 +00:00
Richard Smith be646db000 Update recently-added test to use new __c11_ form of atomic builtins.
llvm-svn: 154514
2012-04-11 18:00:46 +00:00
David Chisnall eb9496efc7 Make __atomic_init() (soon to be __c11_atomic_init()) work with non-scalar types.
llvm-svn: 154507
2012-04-11 17:24:05 +00:00
David Chisnall 28397be059 Allow c++ initialisers to initialise _Atomic fields.
llvm-svn: 154499
2012-04-11 15:29:15 +00:00
Douglas Gregor 11a434ab2e Improve the printing of __PRETTY_FUNCTION__ more provide more
information and more closely match GCC's, from Nikola Smiljanic!

llvm-svn: 154430
2012-04-10 20:14:15 +00:00
Richard Smith 4051ff7650 Don't forget to evaluate the subexpression in a null pointer cast. If we're
converting from std::nullptr_t, the subexpression might have side-effects.

llvm-svn: 154278
2012-04-08 08:02:07 +00:00
Douglas Gregor 31f55dced5 Implement support for null non-type template arguments for non-type
template parameters of pointer, pointer-to-member, or nullptr_t
type in C++11. Fixes PR9700 / <rdar://problem/11193097>.

llvm-svn: 154219
2012-04-06 22:40:38 +00:00
John McCall 76cc43a2a4 Use atexit when __cxa_atexit isn't available instead of adding a
global destructor entry.  For some reason this isn't enabled for
apple-kexts;  it'd be good to have documentation for that.

Based on a patch by Nakamura Takumi!

llvm-svn: 154191
2012-04-06 18:21:06 +00:00
Richard Smith 9380e0ea23 Implement C++11 [temp.arg.nontype]'s permission to use the address of an object
or function with internal linkage as a non-type template argument.

llvm-svn: 154053
2012-04-04 21:11:30 +00:00
Douglas Gregor b61e809b42 Move the computation of the lambda mangling information (mangling
number + context) to the point where we initially start defining the
lambda, so that the linkage won't change when that information is made
available. Fixes the assertion in <rdar://problem/11182962>.

Plus, actually mangle the context of lambdas properly.

llvm-svn: 154029
2012-04-04 17:40:10 +00:00
David Chisnall 5221a947a5 Don't crash (assert failure) when generating blocks for C++ types with a non-const copy constructor.
This was caused by the code deciding the number of fields in the byref structure using a different test to the part of the code creating the GEPs into said structure.  

llvm-svn: 154013
2012-04-04 13:07:13 +00:00
Eric Christopher b7e821a6aa Change location information for synthesized properties to be at the
property file/line rather than the @synthesize file/line. Avoids
some nasty confusing-ness with conflating the file from the scope
and the line from the original declaration. Use    the current scope
location as a separate parameter so that we can    match it up
better in the line table with the beginning of the scope.

Update a couple of testcases accordingly since I had to change
that we actually use the passed in location in EmitFunctionStart
and for the new metadata parameter and add a new testcase to
make sure we've got the right line numbers for synthesized
properties.

Part of rdar://11026482

llvm-svn: 153917
2012-04-03 00:44:15 +00:00
John McCall b88a566cc1 Make sure we unique static-local decls across multiple emissions of
the function body, but do so in a way that doesn't make any assumptions
about the static local actually having a proper, unique mangling,
since apparently we don't do that correctly at all.

llvm-svn: 153776
2012-03-30 21:00:39 +00:00
Chandler Carruth 8453795255 Revert r153723, and its follow-ups r153728 and r153733.
These patches cause us to miscompile and/or reject code with static
function-local variables in an extern-C context. Previously, we were
papering over this as long as the variables are within the same
translation unit, and had not seen any failures in the wild. We still
need a proper fix, which involves mangling static locals inside of an
extern-C block (as GCC already does), but this patch causes pretty
widespread regressions. Firefox, and many other applications no longer
build.

Lots of test cases have been posted to the list in response to this
commit, so there should be no problem reproducing the issues.

llvm-svn: 153768
2012-03-30 19:44:53 +00:00
John McCall 87590e60c0 Do the static-locals thing properly in the face of unions and
other things which might mess with the variable's type.

llvm-svn: 153733
2012-03-30 07:09:50 +00:00
John McCall 947c1830e4 When emitting a static local variable in C++, handle
the case that the variable already exists.  Partly this is just
protection against people making crazy declarations with custom
asm labels or extern "C" names that intentionally collide with
the manglings of such variables, but the main reason is that we
can actually emit a static local variable twice with the
requirement that it match up.  There may be other cases with
(e.g.) the various nested functions, but the main exemplar is
with constructor variants, where we can be forced into
double-emitting the function body under certain circumstances
like (currently) the presence of virtual bases.

llvm-svn: 153723
2012-03-30 04:25:14 +00:00
Eli Friedman a154dd5b33 ConstStructBuilder: fix offset math for base classes so it works correctly in general. Found by inspection.
llvm-svn: 153720
2012-03-30 03:55:31 +00:00
Chad Rosier 615ed1a3a6 Revert r153613 as it's causing large compile-time regressions on the nightly testers.
llvm-svn: 153660
2012-03-29 17:37:10 +00:00
John McCall 1a0877f99d When we can't prove that the target of an aggregate copy is
a complete object, the memcpy needs to use the data size of
the structure instead of its sizeof() value.  Fixes PR12204.

llvm-svn: 153613
2012-03-28 23:30:44 +00:00
Chandler Carruth 8b4140d712 Move the emission of strict enum range metadata behind a flag (the same
flag as GCC uses: -fstrict-enums). There is a *lot* of code making
unwarranted assumptions about the underlying type of enums, and it
doesn't seem entirely reasonable to eagerly break all of it.

Much more importantly, the current state of affairs is *very* good at
optimizing based upon this information, which causes failures that are
very distant from the actual enum. Before we push for enabling this by
default, I think we need to implement -fcatch-undefined-behavior support
for instrumenting and trapping whenever we store or load a value outside
of the range. That way we can track down the misbehaving code very
quickly.

I discussed this with Rafael, and currently the only important cases he
is aware of are the bool range-based optimizations which are staying
hard enabled. We've not seen any issue with those either, and they are
much more important for performance.

llvm-svn: 153550
2012-03-27 23:58:37 +00:00
Rafael Espindola d0decdec57 Fix copy and pasto.
llvm-svn: 153385
2012-03-24 16:57:10 +00:00
Rafael Espindola 5c0034a7c6 Add back r153360 with a fix for enums that cover all the 32 bit values.
Thanks to NAKAMURA Takumi for finding it!

llvm-svn: 153383
2012-03-24 16:50:34 +00:00
NAKAMURA Takumi 2681efcc95 Revert r153360 (and r153380), "Second part of PR12251. Produce the range metadata in clang for booleans and".
For i686 targets (eg. cygwin), I saw "Range must not be empty!" in verifier.

It produces (i32)[0x80000000:0x80000000) from (uint64_t)[0xFFFFFFFF80000000ULL:0x0000000080000000ULL), for signed i32 on MDNode::Range.

llvm-svn: 153382
2012-03-24 14:43:42 +00:00
Benjamin Kramer e2c1e64909 On i386 the alignment of i64 is 4, not 8.
llvm-svn: 153380
2012-03-24 13:22:50 +00:00
Rafael Espindola 54355820e8 Second part of PR12251. Produce the range metadata in clang for booleans and
c++ enums.

llvm-svn: 153360
2012-03-24 00:28:06 +00:00
Chandler Carruth 13332ee518 Relax the FileCheck assertions in this test a touch. Previously the test
relied on an artifact of how the inliner and subsequent passes in
clang's -O3 mode happen to treat basic blocks and the labels for the
basic blocks. In my work on the inliner, and changed this fundamental
assumption, and the label that was being checked on the entry basic
block will no longer appear in opt builds. There was no reason to expect
the label to always be present anyways, much to my regret.

I've changed the test to just ensure that we return an immediate
constant. If there are intervening instructions, that's bad, but not
really that relevant to the test.

I'd love it if others have a better way of checking that a function body
contains only a 'ret' instruction that isn't dependent on whether or not
the entry block receives a label...

llvm-svn: 153243
2012-03-22 06:22:13 +00:00
John McCall b6c4a7ef21 For the annals of subtle but terrible bugs: fix a longstanding bug
in vtable layout where virtual methods inherited from virtual bases
could be assigned the same vcall adjustment slot if they shared
a name and parameter signature but differed in their
cv-qualification.  The code was already trying to handle this
case, but unfortunately used the ordinary type qualifiers
(which are always empty here) instead of the method qualifiers.
This seems like something that the API should discourage, but
I don't know how to carry that principle out in this instance.

Eliminate this function's need for an ASTContext while we're at it.

This bug affects the ABI, and fixing it brings us into accord with
the Itanium ABI (and GCC's implementation of it), but, obviously,
technically breaks full compatibility with previous releases of Clang.
Just letting you know.

llvm-svn: 153168
2012-03-21 06:57:19 +00:00
Benjamin Kramer e6b4a16ddd Fix a use-after-free in thunk emission. EmitThunk may call RAUW on Init, invalidating the pointer.
Fixes PR12284. The test case only triggered under asan/valgrind, but it's better than nothing.

llvm-svn: 153120
2012-03-20 20:18:13 +00:00
Benjamin Kramer 8cc491a601 Debug info: Tighten up uses of plain MDNode pointers which don't survive replaceOperandWith.
TrackingVH notices when it gets RAUW'd. Fixes PR12305 and PR12315.

llvm-svn: 153115
2012-03-20 19:49:14 +00:00
Bill Wendling 44ab2103e2 Make the storing variable's name generic.
llvm-svn: 152963
2012-03-16 23:37:23 +00:00
Bill Wendling f6af30f4c7 When "low alignment" is specified, then set the alignment of the aggregate's
store to 1. This allows code-gen to select a more appropriate alignment. If left
to zero, an alignment greater than the alignment of the pointer may be selected,
causing code-gen to use instructions which require an alignment greater than the
pointer guarantees.
<rdar://problem/11043589>

llvm-svn: 152951
2012-03-16 21:45:12 +00:00
Eric Christopher d36ad66785 Add support to mangle templated member function names with template
args.

Fixes rdar://11042577

llvm-svn: 152691
2012-03-14 00:25:46 +00:00
Eric Christopher 947bb5a1f1 When adding member functions to a class add any specializations of
function templates as well.

A future commit will mangle the added name with the template args
like classes are mangled.

Fixes rdar://10986010

llvm-svn: 152683
2012-03-13 23:40:48 +00:00
James Molloy e943003c09 Ensure that default arguments are handled correctly in sub scopes. For example:
void f () {
  int g (int a, int b=4);
  {
    int g(int a, int b=5);
  }
}

should compile.

llvm-svn: 152621
2012-03-13 08:55:35 +00:00
Rafael Espindola 1f71a23985 The type of a definition should not increase its visibility. Fixes PR12221.
llvm-svn: 152493
2012-03-10 13:01:40 +00:00
Douglas Gregor d2f7007e09 Only make a call to a copy constructor elidable if in fact we are
doing a copy. Fixes PR12139.

llvm-svn: 152485
2012-03-10 06:53:13 +00:00
John McCall 7133505936 Unify the BlockDeclRefExpr and DeclRefExpr paths so that
we correctly emit loads of BlockDeclRefExprs even when they
don't qualify as ODR-uses.  I think I'm adequately convinced
that BlockDeclRefExpr can die.

llvm-svn: 152479
2012-03-10 03:05:10 +00:00
Eli Friedman 9135902478 Make sure the accessors for overridden methods don't return inherited constructors. Fixes PR12219.
llvm-svn: 152470
2012-03-10 01:39:01 +00:00
Richard Smith bcc22fc4e1 Support for raw and template forms of numeric user-defined literals,
and lots of tidying up.

llvm-svn: 152392
2012-03-09 08:00:36 +00:00
Eli Friedman 1c8d2ca829 Make sure we update the static local decl address map when we are forced to rebuild a global because of the initializer. <rdar://problem/10957867>.
llvm-svn: 152372
2012-03-09 03:27:46 +00:00
Eli Friedman effcd47eef Make sure constant emission handles initializer lists with strings correctly. Part of <rdar://problem/10957867>.
llvm-svn: 152370
2012-03-09 03:06:56 +00:00
Rafael Espindola f5140611fc Also test pr12104 with pch.
llvm-svn: 152317
2012-03-08 15:07:42 +00:00
Richard Smith 39570d0020 Add support for cooked forms of user-defined-integer-literal and
user-defined-floating-literal. Support for raw forms of these literals
to follow.

llvm-svn: 152302
2012-03-08 08:45:32 +00:00
Richard Smith 75b67d6dc5 User-defined literal support for character literals.
llvm-svn: 152277
2012-03-08 01:34:56 +00:00
Richard Smith c67fdd4eb9 AST representation for user-defined literals, plus just enough of semantic
analysis to make the AST representation testable. They are represented by a
new UserDefinedLiteral AST node, which is a sugared CallExpr. All semantic
properties, including full CodeGen support, are achieved for free by this
representation.

UserDefinedLiterals can never be dependent, so no custom instantiation
behavior is required. They are mangled as if they were direct calls to the
underlying literal operator. This matches g++'s apparent behavior (but not its
actual mangling, which is broken for literal-operator-ids).

User-defined *string* literals are now fully-operational, but the semantic
analysis is quite hacky and needs more work. No other forms of user-defined
literal are created yet, but the AST support for them is present.

This patch committed after midnight because we had already hit the quota for
new kinds of literal yesterday.

llvm-svn: 152211
2012-03-07 08:35:16 +00:00
Richard Smith 8c029611a6 Don't even try to directly emit the value of a DeclRefExpr if that declaration
is not usable in a constant expression. ~2.5% speedup on 403.gcc / combine.c.

llvm-svn: 152193
2012-03-07 01:58:44 +00:00
Sebastian Redl 249dee5433 If the element type of an initializer list has a destructor, make sure we check it. Fixes PR12178.
llvm-svn: 152048
2012-03-05 19:35:43 +00:00
Rafael Espindola 189fa748ec Fix a small difference in sema and codegen views of what needs to be output.
In the included testcase, soma thinks that we already have a definition after we
see the out of line decl. Codegen puts it in a deferred list, to be output if
a use is seen. This would break when we saw an explicit template instantiation
definition, since codegen would not be notified.

This patch adds a method to the consumer interface so that soma can notify
codegen that this decl is now required.

llvm-svn: 152024
2012-03-05 10:54:55 +00:00
James Molloy 7583ccdc1f Fix a bug in the mangler where in 'namespace std { extern "C" {X;} }', X would not be seen to be in ::std::.
Migrate two other places where the same logic is used to use the helper function that already exists.

llvm-svn: 152022
2012-03-05 09:59:43 +00:00
Richard Smith 522fa53703 Add a pile of tests for unrestricted unions, and advertise support for them.
llvm-svn: 151992
2012-03-03 23:51:05 +00:00
Richard Smith bc638767f8 Reinstate r151879, r151880, reverted in r151922, along with a bugfix for
scalar emission of DeclRefExprs to const bools: emit scalar bools as i1,
not as i8.

In addition to the extra unit testing, this has successfully bootstrapped.

llvm-svn: 151955
2012-03-02 23:27:11 +00:00
Daniel Dunbar d01281fe0d Revert r151879, r151880, "PR12145: Avoid emitting loads of constexpr variables in contexts where there" and "Fix buildbot: make this test less dependent on the value names in the produced IR."
They broke bootstrap.

llvm-svn: 151922
2012-03-02 16:24:25 +00:00
Richard Smith 545f38c3d2 Fix buildbot: make this test less dependent on the value names in the produced IR.
llvm-svn: 151880
2012-03-02 03:22:38 +00:00
Richard Smith a541a3bbee PR12145: Avoid emitting loads of constexpr variables in contexts where there
is no odr-use of the variable. Go slightly beyond what the standard requires
for variables of reference type.

llvm-svn: 151879
2012-03-02 03:16:32 +00:00
Eric Christopher 5bc1f47b00 Update test for llvm assembly output change. Also add a fixme that this
shouldn't be relying on assembly emission.

For the record we should check the metadata output from the front
end and then check in the backend that such a thing emits a pubtypes
entry.

llvm-svn: 151851
2012-03-01 22:49:10 +00:00
Eric Christopher 4b6753cf10 Reapply r151702 with a small fix for a failure to cut and paste
correctly.

Still rdar://10900684

llvm-svn: 151838
2012-03-01 21:36:52 +00:00
Daniel Dunbar c2f9dc91c1 Revert r151702, "Add support for handling captured variables in lambda debug
info.", which broke some -O0 -g tests.

llvm-svn: 151730
2012-02-29 16:28:29 +00:00
Eric Christopher 3203f6b9da Add support for handling captured variables in lambda debug info.
This currently doesn't handle capturing the 'this' pointer for any
enclosing class.

Steal the lambda-expressions.cpp testcase and debugify it and try
to use more variables to proof it against random changes.

Part of rdar://10900684

llvm-svn: 151702
2012-02-29 03:25:30 +00:00
Eli Friedman 576cbd03b4 Make sure list-initialization of arrays works correctly in explicit type conversions. PR12121.
llvm-svn: 151674
2012-02-29 00:00:28 +00:00
Sebastian Redl a235e2d8ad Hack in a loud error for PR12086. Better than a silent miscompile.
llvm-svn: 151586
2012-02-27 23:20:01 +00:00
Eric Christopher 5bd52c1b3e Make this test a bit more robust for debug info changes.
llvm-svn: 151518
2012-02-27 08:22:57 +00:00
Richard Smith 0d1f3cb1b5 Special members which are defaulted or deleted on their first declaration are
trivial if the implicit declaration would be. Don't forget to set the Trivial
flag on the special member as well as on the class. It doesn't seem ideal that
we have two separate mechanisms for storing this information, but this patch
does not attempt to address that.

This leaves us in an interesting position where the has_trivial_X trait for a
class says 'yes' for a deleted but trivial X, but is_trivially_Xable says 'no'.
This seems to be what the standard requires.

llvm-svn: 151465
2012-02-26 00:31:33 +00:00
Sebastian Redl aeac3ff82a Richard Smith pointed out that there already is a proposal for init list mangling.
llvm-svn: 151462
2012-02-25 22:59:28 +00:00
Sebastian Redl 4a7eab2545 CodeGen support for global variables of type std::initializer_list<X>.
This emits a backing array with internal linkage and fills it with data,
then has the initializer_list point at the array. Dynamic initialization
and global destructors are correctly supported.

What doesn't work is nested initializer_lists. I have no idea how to
get them to work, either. However, these should be very rare, and so
I'll just call it a known bug and declare generalized initializers
DONE!

llvm-svn: 151457
2012-02-25 20:51:20 +00:00
Sebastian Redl dadc06494d Better mangling for new-expressions. Also, although we can't mangle arbitrary initializer lists yet (we will need this), turn the crash into a controlled error.
llvm-svn: 151455
2012-02-25 20:51:07 +00:00
Douglas Gregor 4f283c8b20 For the purposes of building LLVM types, a forward-declared
enumeration type with a fixed underlying type is complete. Fixes
<rdar://problem/10916155>.

llvm-svn: 151403
2012-02-24 22:40:36 +00:00
Douglas Gregor 3a08c1cd3b Two minor, related fixes for template instantiation with blocks:
- Make sure that the block expression is instantiation-dependent if the
    block is in a dependent context
  - Make sure that the C++ 'this' expression gets captured even if we
  don't rebuild the AST node during template instantiation. This would
  also have manifested as a bug for lambdas.

Fixes <rdar://problem/10832617>.

llvm-svn: 151372
2012-02-24 17:41:38 +00:00
Chad Rosier f62290a19b Reapply r151172 - Unwind path cleanup for array new list initializers - with a
test case that only runs on debug builds.

llvm-svn: 151311
2012-02-24 00:13:55 +00:00
Richard Smith c899892485 PR12067: When emitting an evaluated constant structure in C++11 mode, don't
forget the vptrs.

llvm-svn: 151245
2012-02-23 08:33:23 +00:00
Rafael Espindola 8093fdfab1 Two fixes to how we compute visibility:
* Handle some situations where we should never make a decl more visible,
  even when merging in an explicit visibility.

* Handle attributes in members of classes that are explicitly specialized.

Thanks Nico for the report and testing, Eric for the initial review, and dgregor
for the awesome test27 :-)

llvm-svn: 151236
2012-02-23 04:17:32 +00:00
Chad Rosier 1a3c3d4537 Revert r151172: Unwind path cleanup for array new list initializers.
llvm-svn: 151203
2012-02-22 21:36:06 +00:00
Sebastian Redl 674c77e555 Unwind path cleanup for array new list initializers.
llvm-svn: 151172
2012-02-22 17:37:59 +00:00
Sebastian Redl f862eb6a9f CodeGen for array new list initializers. Doesn't correctly clean up in the face of exceptions yet.
llvm-svn: 151171
2012-02-22 17:37:52 +00:00
NAKAMURA Takumi 0716b50db6 test/CodeGenCXX/thiscall-struct-return.cpp: Relax expressions for -Asserts.
llvm-svn: 151126
2012-02-22 03:36:54 +00:00
Aaron Ballman 3c42441b2c Adding support for Microsoft's thiscall calling convention. Clang side of the patch.
llvm-svn: 151122
2012-02-22 03:04:13 +00:00
Douglas Gregor 6f88e5e0d7 Implement non-internal linkage for lambda closure types that need a
stable mangling, since these lambdas can end up in multiple
translation units. Sema is responsible for deciding when this is the
case, because it's already responsible for choosing the mangling
number.

llvm-svn: 151029
2012-02-21 04:17:39 +00:00
Douglas Gregor 926410d2db Implement name mangling for lambda expressions that occur within the
initializers of data members (both static and non-static).

llvm-svn: 151017
2012-02-21 02:22:07 +00:00
Douglas Gregor 7fcbd902b4 Implement name mangling for lambda expressions that occur within the
default arguments of function parameters. This simple-sounding task is
complicated greatly by two issues:

  (1) Default arguments aren't actually a real context, so we need to
  maintain extra state within lambda expressions to track when a
  lambda was actually in a default argument.
  (2) At the time that we parse a default argument, the FunctionDecl
  doesn't exist yet, so lambda closure types end up in the enclosing
  context. It's not clear that we ever want to change that, so instead
  we introduce the notion of the "effective" context of a declaration
  for the purposes of name mangling.

llvm-svn: 151011
2012-02-21 00:37:24 +00:00
Nick Lewycky 4506204640 Emit the exact size for the invariant intrinsics.
llvm-svn: 151010
2012-02-21 00:26:58 +00:00
Douglas Gregor 43c3f28c23 Make sure that we set up the right declaration contexts when creating
and introducing the lambda closure type and its function call
operator. Previously, we assumed that the lambda closure type would
land directly in the current context, and not some parent context (as
occurs with linkage specifications). Thanks to Richard for the test case.

llvm-svn: 150987
2012-02-20 20:47:06 +00:00
Douglas Gregor d5c4844e02 When we resolve the type of an 'auto' variable, clear out the linkage
of that variable; it will need to be recomputed with the resolved
type.

llvm-svn: 150984
2012-02-20 20:05:29 +00:00
Douglas Gregor 6379854457 Basic support for name mangling of C++11 lambda expressions. Because
name mangling in the Itanium C++ ABI for lambda expressions is so
dependent on context, we encode the number used to encode each lambda
as part of the lambda closure type, and maintain this value within
Sema.

Note that there are a several pieces still missing:
  - We still get the linkage of lambda expressions wrong
  - We aren't properly numbering or mangling lambda expressions that
  occur in default function arguments or in data member initializers.
  - We aren't (de-)serializing the lambda numbering tables

llvm-svn: 150982
2012-02-20 19:44:39 +00:00
Eric Christopher 66562a4a54 Remove the type retaining from the clang frontend. This is now
handled by the caching and rauw. Also fix one cache that wasn't
being added to highlighted by this patch. Update all testcases
accordingly.

This should fix the deall failure.

llvm-svn: 150977
2012-02-20 18:05:24 +00:00
Benjamin Kramer 8558da1930 Harden test against for operator new(unsigned int).
llvm-svn: 150932
2012-02-19 16:20:58 +00:00
Sebastian Redl d026dc499c Make heap-allocation of std::initializer_list 'work'.
llvm-svn: 150931
2012-02-19 16:03:09 +00:00
Sebastian Redl 4e04dd1979 Make std::initializer_list member initializers 'work'.
llvm-svn: 150930
2012-02-19 15:41:54 +00:00
Sebastian Redl 8eb351d72e Get recursive initializer lists to work and add a test. Codegen of std::initializer_list is now complete. Onward to array new.
llvm-svn: 150926
2012-02-19 12:28:02 +00:00
Sebastian Redl 99f6616470 Add a testcase for using objects with list-constructors, and fix a Sema crash by repeating an old hack.
llvm-svn: 150925
2012-02-19 12:27:56 +00:00
Sebastian Redl 1d197174a2 Add a testcase to show that temporaries from the initializer list are destroyed correctly.
llvm-svn: 150924
2012-02-19 12:27:51 +00:00
Sebastian Redl ca89d6841a Add a testcase for start+end implementations of std::initializer_list.
llvm-svn: 150923
2012-02-19 12:27:47 +00:00
Eric Christopher 1c3785aae9 Add in a caching mechanism so that forward declarations are replaced
with full types if they exist.

rdar://10809898 and rdar://10209967 and rdar://10400981

llvm-svn: 150858
2012-02-18 00:50:17 +00:00
Richard Smith 132bea96ad Don't emit optimization-specific intrinsic at -O0.
llvm-svn: 150838
2012-02-17 20:12:52 +00:00
Sebastian Redl c83ed8248e Basic code generation support for std::initializer_list.
We now generate temporary arrays to back std::initializer_list objects
initialized with braces. The initializer_list is then made to point at
the array. We support both ptr+size and start+end forms, although
the latter is untested.

Array lifetime is correct for temporary std::initializer_lists (e.g.
call arguments) and local variables. It is untested for new expressions
and member initializers.

Things left to do:
Massively increase the amount of testing. I need to write tests for
start+end init lists, temporary objects created as a side effect of
initializing init list objects, new expressions, member initialization,
creation of temporary objects (e.g. std::vector) for initializer lists,
and probably more.
Get lifetime "right" for member initializers and new expressions. Not
that either are very useful.
Implement list-initialization of array new expressions.

llvm-svn: 150803
2012-02-17 08:42:25 +00:00
Richard Smith 08a5144621 The clang half of r150794: after the construction of a global or static const
variable ends, if the variable has a trivial destructor and no mutable
subobjects then emit an llvm.invariant.start call for it. globalopt knows to
make the variable const when evaluating this.

llvm-svn: 150798
2012-02-17 07:31:37 +00:00
Richard Smith e070fd2ac8 Bug fix: do not emit static const local variables with mutable members
as constants.

Refactor and simplify all the separate checks for whether a type can be
emitted as a constant.

llvm-svn: 150793
2012-02-17 06:48:11 +00:00
Richard Smith ae819500a1 When performing IRGen on a global, emit it as a constant if:
1) It has a const-qualified type, and
 2) It has no mutable members, and
 3) It has no dynamic initialization, and
 4) It has trivial destruction.
Remove the unnecessary requirement that the type be POD. This allows us to
mark all constexpr objects with no mutable members as 'constant'.

llvm-svn: 150792
2012-02-17 04:54:50 +00:00
Richard Smith 14a941380a Make sure all remaining parts of the constant evaluator are aware that an array
can be represented by an LValue, and use that to simplify the code a little.

llvm-svn: 150789
2012-02-17 03:35:37 +00:00
Douglas Gregor 355efbb2e0 Rework the Sema/AST/IRgen dance for the lambda closure type's
conversion to function pointer. Rather than having IRgen synthesize
the body of this function, we instead introduce a static member
function "__invoke" with the same signature as the lambda's
operator() in the AST. Sema then generates a body for the conversion
to function pointer which simply returns the address of __invoke. This
approach makes it easier to evaluate a call to the conversion function
as a constant, makes the linkage of the __invoke function follow the
normal rules for member functions, and may make life easier down the
road if we ever want to constexpr'ify some of lambdas.

Note that IR generation is responsible for filling in the body of
__invoke (Sema just adds a dummy body), because the body can't
generally be expressed in C++.

Eli, please review!

llvm-svn: 150783
2012-02-17 03:02:34 +00:00
Richard Smith 5d108606cc PR12012: Fix a regression in r150419 where we would try (and fail) to
zero-initialize class types with virtual bases when constant-evaluating an
initializer.

llvm-svn: 150770
2012-02-17 00:44:16 +00:00
Eric Christopher 4c006e557d Reapply r150631:
"Add a completed/incomplete type difference. This allows us to have
    partial types for contexts and forward decls while allowing us to
    complete types later on for debug purposes.

    This piggy-backs on the metadata replacement and rauw changes
    for temporary nodes and takes advantage of the incremental
    support I added in earlier. This allows us to, if we decide,
    to limit adding methods and variables to structures in order
    to limit the amount of debug information output into a .o file.

    The caching is a bit complicated though so any thoughts on
    untangling that are welcome."

with a fix:

 - Remove all RAUW during type construction by adding stub versions
   of types that we later complete.

and some TODOs:

 - Add an RAUW cache for forward declared types so that we can replace
   them at the end of compilation.
 - Remove the code that updates on completed types because we no
   longer need to have that happen. We emit incomplete types on
   purpose and only want to know when we want to complete them.

llvm-svn: 150752
2012-02-16 22:54:45 +00:00
Eli Friedman b971d49765 Elide copy construction in new expressions. PR11757.
llvm-svn: 150738
2012-02-16 22:45:48 +00:00
Richard Smith 4297375561 C++11 allows unions to have static data members. Remove the corresponding
restriction and add some tests.

llvm-svn: 150721
2012-02-16 20:41:22 +00:00
Eric Christopher 9a89705d03 Revert "Add a completed/incomplete type difference. This allows us to have"
This reverts commit 9a68d4584afcd0853b930bd80235b58736e785b4.

llvm-svn: 150637
2012-02-15 23:51:20 +00:00
Eric Christopher 7381907e65 Add a completed/incomplete type difference. This allows us to have
partial types for contexts and forward decls while allowing us to
complete types later on for debug purposes.

This piggy-backs on the metadata replacement and rauw changes
for temporary nodes and takes advantage of the incremental
support I added in earlier. This allows us to, if we decide,
to limit adding methods and variables to structures in order
to limit the amount of debug information output into a .o file.

The caching is a bit complicated though so any thoughts on
untangling that are welcome.

llvm-svn: 150631
2012-02-15 23:25:18 +00:00
John McCall c62bb39142 Split reinterpret_casts of member pointers out from CK_BitCast; this
is general goodness because representations of member pointers are
not always equivalent across member pointer types on all ABIs
(even though this isn't really standard-endorsed).

Take advantage of the new information to teach IR-generation how
to do these reinterprets in constant initializers.  Make sure this
works when intermingled with hierarchy conversions (although
this is not part of our motivating use case).  Doing this in the
constant-evaluator would probably have been better, but that would
require a *lot* of extra structure in the representation of
constant member pointers:  you'd really have to track an arbitrary
chain of hierarchy conversions and reinterpretations in order to
get this right.  Ultimately, this seems less complex.  I also
wasn't quite sure how to extend the constant evaluator to handle
foldings that we don't actually want to treat as extended
constant expressions.

llvm-svn: 150551
2012-02-15 01:22:51 +00:00
Richard Smith 1bfe068e71 PR11650: Implement resolution of core issue 1301. Value initialization can't be
used to construct an object of union type with a deleted default constructor
(plus fixes for some related value-initialization corner cases).

llvm-svn: 150502
2012-02-14 21:14:13 +00:00
Eli Friedman 4412b23e7e Add a coverage test for lambda expression IRGen.
llvm-svn: 150458
2012-02-14 03:07:59 +00:00
Richard Smith 6331c408b5 Deal with a horrible C++11 special case. If a non-literal type has a constexpr
constructor, and that constructor is used to initialize an object of static
storage duration such that all members and bases are initialized by constant
expressions, constant initialization is performed. In this case, the object
can still have a non-trivial destructor, and if it does, we must emit a dynamic
initializer which performs no initialization and instead simply registers that
destructor.

llvm-svn: 150419
2012-02-13 22:16:19 +00:00
Eric Christopher d752e3d280 Testcase for previous commit.
PR11970.

llvm-svn: 150389
2012-02-13 15:04:15 +00:00
Eric Christopher 8a41bd88d7 Temporarily walk back a few of my recent debug info limiting changes
while reworking how we handle wanting to emit only parts of structures.

Fixes PR11970.

llvm-svn: 150388
2012-02-13 14:56:11 +00:00
Eric Christopher 39e39c8efe Don't cache the artificial type for the this pointer, there's no
difference in the qual type. This is a workaround for the fact that
the type isn't artificial but the this decl is, however, we don't
have any way of representing it in the current metadata. For now,
however, just don't cache the full type.

Fixes rdar://10831526 and probably a couple of others.

llvm-svn: 150159
2012-02-09 07:26:21 +00:00
Eric Christopher f76dd56dd3 Whitespace.
llvm-svn: 150026
2012-02-08 00:23:21 +00:00
Eric Christopher 75b90c4a86 Use the new forward declaration scheme for records. Also add more
caching of results after we create them.

Fixes rdar://10809898

llvm-svn: 150025
2012-02-08 00:23:18 +00:00
Craig Topper 33aa76c88a Test for r149855.
llvm-svn: 149862
2012-02-06 08:31:19 +00:00
Richard Smith 574e896976 Implement name mangling for scalar value initialization. Reported on IRC by Xeo.
llvm-svn: 149854
2012-02-06 02:54:51 +00:00
Devang Patel fb04ece198 Update tests so that they don't rely upon LLVMDebugVersion number.
llvm-svn: 149726
2012-02-03 23:58:39 +00:00
Douglas Gregor 3dd5fe2006 Make sure that the layout-override parser grabs the size, not the data
size. Otherwise, we can end up with bogus layouts.

llvm-svn: 149703
2012-02-03 19:31:51 +00:00
Douglas Gregor 1f79ca839c When a pack expansion occurs in the template argument list of an alias
template without a corresponding parameter pack, don't immediately
substitute the alias template. This is under discussion in the C++
committee, and may become ill-formed, but for now we match GCC.

llvm-svn: 149697
2012-02-03 17:16:23 +00:00
Rafael Espindola 6d65d7b63d Reject mismatched "#pragma GCC visibility push" and "#pragma GCC visibility pop".
llvm-svn: 149559
2012-02-01 23:24:59 +00:00
Eric Christopher ee4ab93906 For pass-by-value record arguments to functions emit a forward decl
instead of the entire class definition.

llvm-svn: 149474
2012-02-01 06:07:23 +00:00
Bob Wilson 8e5acc5cc1 Use the new Triple::getMacOSXVersion function in another place.
I removed support for "*-darwin*-iphoneos" triples, since we now have
iOS listed as a separate OS in the triples.

llvm-svn: 149455
2012-01-31 23:52:58 +00:00
Richard Smith 440e9db9e7 Extend for-range temporary cleanups codegen test to catch the bug which Eli
fixed in r149440.

llvm-svn: 149450
2012-01-31 23:43:25 +00:00
John McCall a023b6924a Per discussion on cxx-abi-dev, don't drop leading zeroes from the
mangling of floating-point literals.  I just went ahead and
reimplemented toString() here;  if someone wants to generalize
the library routine to do this, or feels strongly that we should
be post-processing, please feel free.

llvm-svn: 149256
2012-01-30 18:36:31 +00:00
John McCall 6e1c012385 Get a little bit smarter about killing off the ReturnValue alloca
in the presence of straight-line cleanups.  This is a simple but
important case, particularly for ARC.

llvm-svn: 149190
2012-01-29 02:35:02 +00:00
John McCall eed64c77d2 Complain about attempts to use 'protected' visibility on targets
like Darwin that don't support it.  We should also complain about
invalid -fvisibility=protected, but that information doesn't seem
to exist at the most appropriate time, so I've left a FIXME behind.

llvm-svn: 149186
2012-01-29 01:20:30 +00:00
Richard Smith 253c2a390a constexpr: Implement the [dcl.constexpr]p5 check for whether a constexpr
function definition can produce a constant expression. This also provides the
last few checks for [dcl.constexpr]p3 and [dcl.constexpr]p4.

llvm-svn: 149108
2012-01-27 01:14:48 +00:00
Douglas Gregor e9fc377a44 Extend the ExternalASTSource interface to allow the AST source to
provide the layout of records, rather than letting Clang compute
the layout itself. LLDB provides the motivation for this feature:
because various layout-altering attributes (packed, aligned, etc.)
don't get reliably get placed into DWARF, the record layouts computed
by LLDB from the reconstructed records differ from the actual layouts,
and badness occurs. This interface lets the DWARF data drive layout,
so we don't need the attributes preserved to get the answer write.

The testing methodology for this change is fun. I've introduced a
variant of -fdump-record-layouts called -fdump-record-layouts-simple
that always has the simple C format and provides size/alignment/field
offsets. There is also a -cc1 option -foverride-record-layout=<file>
to take the output of -fdump-record-layouts-simple and parse it to
produce a set of overridden layouts, which is introduced into the AST
via a testing-only ExternalASTSource (called
LayoutOverrideSource). Each test contains a number of records to lay
out, which use various layout-changing attributes, and then dumps the
layouts. We then run the test again, using the preprocessor to
eliminate the layout-changing attributes entirely (which would give us
different layouts for the records), but supplying the
previously-computed record layouts. Finally, we diff the layouts
produced from the two runs to be sure that they are identical.

Note that this code makes the assumption that we don't *have* to
provide the offsets of bases or virtual bases to get the layout right,
because the alignment attributes don't affect it. I believe this
assumption holds, but if it does not, we can extend
LayoutOverrideSource to also provide base offset information.

Fixes the Clang side of <rdar://problem/10169539>.

llvm-svn: 149055
2012-01-26 07:55:45 +00:00
Chris Lattner e8c94f7d98 fix to go along with an llvm change: VMCore now returns an UndefValue
when asking for a ConstantStruct with all undef elements.

llvm-svn: 149025
2012-01-26 02:33:22 +00:00
Rafael Espindola 273fd772ab Fix our handling of #pragma GCC visibility.
Now the lexer just produces a token and the parser is the one responsible for
activating it.
This fixes problem like the one pr11797 where the lexer and the parser were not
in sync. This also let us be more strict on where in the file we accept
these pragmas.

llvm-svn: 149014
2012-01-26 02:02:57 +00:00
Eli Friedman 9562f39e2f Don't stack-allocate an IntegerLiteral which can be referred to after the current method returns. PR11744, part 2.
llvm-svn: 148995
2012-01-25 23:20:27 +00:00
Eli Friedman 5995489d27 Fix a crash for an edge case of the GNU ?: extension.
llvm-svn: 148923
2012-01-25 05:04:17 +00:00
Eric Christopher f3ed1bc8ac Testcase for PR11345.
llvm-svn: 148585
2012-01-20 22:10:18 +00:00
Sebastian Pop 422377cfd3 rename -ccc-host-triple into -target
llvm-svn: 148582
2012-01-20 22:01:23 +00:00
Richard Smith f8379a0fc3 constexpr: converted constant expression handling for enumerator values, case
values and non-type template arguments of integral and enumeration types.

This change causes some legal C++98 code to no longer compile in C++11 mode, by
enforcing the C++11 rule that narrowing integral conversions are not permitted
in the final implicit conversion sequence for the above cases.

llvm-svn: 148439
2012-01-18 23:55:52 +00:00
Fariborz Jahanian 07b5f4ca03 output body of folded case again.
llvm-svn: 148361
2012-01-17 23:55:19 +00:00
Fariborz Jahanian 9ce39e299d Folding away unreachable case statement.
patch (slightly revised) by Aaron Ballman.

llvm-svn: 148359
2012-01-17 23:39:50 +00:00
Fariborz Jahanian 1352eea8c9 Another test for 148243
non-constant-folded-switch containing a constant-folded switch.

llvm-svn: 148247
2012-01-16 18:22:26 +00:00
Fariborz Jahanian 7c4dcf4b93 A little more elaborate test for r148243
llvm-svn: 148245
2012-01-16 18:07:45 +00:00
Fariborz Jahanian 60bb82a584 Patch for r148243 which was left behind.
llvm-svn: 148244
2012-01-16 17:52:22 +00:00
Richard Smith dafff94759 constexpr irgen: Add irgen support for APValue::Struct, APValue::Union,
APValue::Array and APValue::MemberPointer. All APValue values can now be emitted
as constants.

Add new CGCXXABI entry point for emitting an APValue MemberPointer. The other
entrypoints dealing with constant member pointers are no longer necessary and
will be removed in a later change.

Switch codegen from using EvaluateAsRValue/EvaluateAsLValue to
VarDecl::evaluateValue. This performs caching and deals with the nasty cases in
C++11 where a non-const object's initializer can refer indirectly to
previously-initialized fields within the same object.

Building the intermediate APValue object incurs a measurable performance hit on
pathological testcases with huge initializer lists, so we continue to build IR
directly from the Expr nodes for array and record types outside of C++11.

llvm-svn: 148178
2012-01-14 04:30:29 +00:00
Rafael Espindola 2f869a3f4f Remember if a type has its visibility set explicitly or implicitly.
With that, centralize the way we merge visibility, always preferring explicit over
implicit and then picking the most restrictive one.
Fixes pr10113 and pr11690.

llvm-svn: 148163
2012-01-14 00:30:36 +00:00
Eli Friedman d749c6bf2e Revert r148138; it's causing test failures.
llvm-svn: 148141
2012-01-13 21:33:06 +00:00
Sebastian Pop 9a8d528ddf rename -ccc-host-triple into -target
llvm-svn: 148138
2012-01-13 20:37:02 +00:00
Eli Friedman 09d272d025 A base subobject type doesn't make sense for unions; don't try to compute it. Based on patch by Yin Ma. Fixes PR11751.
llvm-svn: 148093
2012-01-13 03:58:31 +00:00
Richard Smith 02ab9c2940 Allow constant-folding of references which were formed in a manner not permitted
in a constant expression, for compatibility with g++.

llvm-svn: 148020
2012-01-12 06:08:57 +00:00
John McCall 8ba2f20bc0 Fix the mangling of class template arguments in a particular
dependent case.  Thanks to Jason Merrill for pointing this out.

llvm-svn: 147653
2012-01-06 05:06:35 +00:00
Eli Friedman b1bc368ca7 Address Richard's review comments on r147561 (Evaluate support for address-of-label differences).
llvm-svn: 147631
2012-01-05 23:59:40 +00:00
Eli Friedman 50c9d08fa5 Fix test on Release builds.
llvm-svn: 147565
2012-01-04 23:41:09 +00:00
Eli Friedman fd5e54da2d Add an APValue representation for the difference between two address-of-label expressions. Add support to Evaluate and CGExprConstant for generating/handling them. Remove the special-case for such differences in Expr::isConstantInitializer.
With that done, remove a bunch of buggy code from CGExprConstant for handling scalar expressions which is no longer necessary.

Fixes PR11705.

llvm-svn: 147561
2012-01-04 23:13:47 +00:00
Rafael Espindola 2aa7acfadb Process attributes in explicit function template instantiations. Fixes part of
PR11690.

llvm-svn: 147523
2012-01-04 05:40:59 +00:00
Rafael Espindola 0b0620776a Process attributes in explicit template instatiation definitions. Fixes PR11690.
llvm-svn: 147441
2012-01-03 06:04:21 +00:00
Rafael Espindola eeb9d9fef6 Small cosmetic cleanups in code I will change anyway.
llvm-svn: 147424
2012-01-02 06:26:22 +00:00
Rafael Espindola 0f960a0a5a Consider visibility attributes in namespaces as being explicit. I.e., they
take precedence over command line options. Fixes PR10113.

llvm-svn: 147405
2012-01-01 18:06:40 +00:00
Richard Smith fddd384b73 Unrevert r147271, reverted in r147361.
Also temporarily remove the assumption from IR gen that we can emit IR for every
constant we can fold, since it isn't currently true in C++11, to fix PR11676.

Original comment from r147271:

constexpr: perform zero-initialization prior to / instead of performing a
constructor call when appropriate. Thanks to Eli for spotting this.

llvm-svn: 147384
2011-12-30 21:15:51 +00:00
Rafael Espindola f49e9ce849 Add test for PRr11676.
llvm-svn: 147363
2011-12-30 03:13:44 +00:00
Rafael Espindola fafe4b7572 Revert r147271. This fixes PR11676.
llvm-svn: 147362
2011-12-30 03:11:50 +00:00
Rafael Espindola 0b79046c5c Fix the visibility of methods of explicit template instantiation definition
when using -fvisibility-inlines-hidden. This matches gcc's behavior and
documentation.

Fixes PR11642.

llvm-svn: 147295
2011-12-27 21:15:28 +00:00
Richard Smith 6e52514c5d constexpr: support for evaluation and codegen of typeid constants.
llvm-svn: 147290
2011-12-27 12:18:28 +00:00
Richard Smith 771c4a1775 constexpr: perform zero-initialization prior to / instead of performing a
constructor call when appropriate. Thanks to Eli for spotting this.

llvm-svn: 147271
2011-12-25 20:00:17 +00:00
Richard Smith 966c1fb0c8 Always implicitly declare move assignment operations for dynamic classes, in
case they override virtual functions from a base class. Also fix -print-stats
counting of move assignment/construction.

llvm-svn: 147258
2011-12-24 21:56:24 +00:00
Richard Smith 0f8ede1654 Revert r146766, and add a testcase for which it introduced a wrong-code bug.
llvm-svn: 146961
2011-12-20 04:00:21 +00:00
Richard Smith 764d2fe666 Unlike in C++03, a constant-expression is not an unevaluated operand in C++11.
Split out a new ExpressionEvaluationContext flag for this case, and don't treat
it as unevaluated in C++11. This fixes some crash-on-invalids where we would
allow references to class members in potentially-evaluated constant expressions
in static member functions, and also fixes half of PR10177.

The fix to PR10177 exposed a case where template instantiation failed to provide
a source location for a diagnostic, so TreeTransform has been tweaked to supply
source locations when transforming a type. The source location is still not very
good, but MarkDeclarationsReferencedInType would need to operate on a TypeLoc to
improve it further.

Also fix MarkDeclarationReferenced in C++98 mode to trigger instantiation for
static data members of class templates which are used in constant expressions.
This fixes a link-time problem, but we still incorrectly treat the member as
non-constant. The rest of the fix for that issue is blocked on PCH support for
early-instantiated static data members, which will be added in a subsequent
patch.

llvm-svn: 146955
2011-12-20 02:08:33 +00:00
Eli Friedman e321795c60 Remove a non-gcc-compatible extension that would apply attributes on declarations without a declarator to structs. Add a warning for ignored attributes. Patch by Michael Han.
llvm-svn: 146796
2011-12-17 00:36:09 +00:00
Eric Christopher 0f79160bb7 Add a triple to this test to hopefully get it passing on i386-freebsd.
llvm-svn: 146795
2011-12-17 00:35:01 +00:00
Eric Christopher cd88813dbf When in a limited debug information situation if we can find the original
declaration for the type then go ahead and use that, it's still smaller
than creating an all new derived type.

Part of rdar://10335756 and others.

llvm-svn: 146779
2011-12-16 23:40:18 +00:00
Eric Christopher ea225fb412 Start down the path of getting clang to internally agree on structs versus
classes.

Part of rdar://10520586 and a couple others.

llvm-svn: 146778
2011-12-16 23:40:14 +00:00
Nick Lewycky f43e681ed7 Testcase for LLVM commit r146475.
llvm-svn: 146476
2011-12-13 05:09:22 +00:00
Nick Lewycky 304584f5a5 s/%clang-cc1/%clang_cc1/ for conformity.
llvm-svn: 146454
2011-12-13 00:32:15 +00:00
Eli Friedman 3c840aad54 Make CGRecordLayoutBuilder correctly switch over to a packed class when a class has a base whose alignment will break the class layout. <rdar://problem/10551376>.
llvm-svn: 146443
2011-12-12 23:13:20 +00:00
Douglas Gregor c7f46f20d1 When we manage to re-use an expression during tree transformation (=
template instantiation), and that expression might produce a
temporary, invoke MaybeBindToTemporary. Otherwise, we forget to
destroy objects, release objects, etc. Fixes <rdar://problem/10531073>.

llvm-svn: 146301
2011-12-10 00:23:21 +00:00
Eli Friedman de6a86b4f6 Make array new on a pointer to data member type work correctly. PR11523.
llvm-svn: 146291
2011-12-09 23:05:37 +00:00