Commit Graph

9463 Commits

Author SHA1 Message Date
John McCall 4cec5f806b Fix another case of giving the wrong value kind to a dependent cast to
a non-dependent type.

llvm-svn: 120384
2010-11-30 02:05:44 +00:00
Nick Lewycky 5421621d2b Replace \r\n with \n. No functional change.
llvm-svn: 120379
2010-11-30 01:47:38 +00:00
Argyrios Kyrtzidis 091d97c709 Revert r120331 since it causes spurious warnings and a possible assertion hit when self-host.
llvm-svn: 120351
2010-11-29 23:42:03 +00:00
Fariborz Jahanian e1802eefac Incomplete enum types not to be treated as integer type
when checking for integer signed/unsigned-ness. PR8694,
// rdar://8707031

llvm-svn: 120345
2010-11-29 23:18:09 +00:00
Argyrios Kyrtzidis 569cad9734 Emit warnings if we are returning a reference to a local temporary.
The issue was brought to our attention by Matthieu Monrocq.

llvm-svn: 120331
2010-11-29 22:32:08 +00:00
Nico Weber 44887f69d5 Always use a function's decl context when building default arguments. Fixes http://http://llvm.org/pr8479.
llvm-svn: 120299
2010-11-29 18:19:25 +00:00
John McCall 032092ffc2 Revert r120063, it was wrong.
llvm-svn: 120296
2010-11-29 18:01:58 +00:00
Douglas Gregor 6399366d17 Add a test for C++ [stmt.label]p1, from Sashan!
llvm-svn: 120291
2010-11-29 16:37:40 +00:00
NAKAMURA Takumi c66906a6cb test/CMakeLists.txt: Implement the target "check-all". For now, it has no dependencies.
llvm-svn: 120280
2010-11-29 01:18:56 +00:00
Nico Weber dcf6fc5dc4 Revert parts of r120266 that I did not mean to commit
llvm-svn: 120267
2010-11-28 22:58:38 +00:00
Nico Weber 20c9f1db48 Minor whitespace and comment fixes. No functionality change.
llvm-svn: 120266
2010-11-28 22:53:37 +00:00
Anders Carlsson 66bbf50877 Look through parentheses when deciding whether an expr is a temporary object. Fixes PR8683.
llvm-svn: 120247
2010-11-28 16:40:49 +00:00
Sebastian Redl 002a3c506a Forgot a file in r120182
llvm-svn: 120184
2010-11-26 18:37:14 +00:00
Sebastian Redl 346203105f Allow access to non-static members without an object in sizeof expressions, in C++0x. Patch by Jakub Wieczorek.
llvm-svn: 120182
2010-11-26 16:28:07 +00:00
John McCall 29ac8e2ecd For internal consistency's sake, compute the value kind of a dependent cast
based on the known properties of the casted-to type.  Fixes a crash on spirit.

llvm-svn: 120180
2010-11-26 10:57:22 +00:00
Zhongxing Xu ec0b8e32ee Regionstore: support derived-to-base cast by creating a CXXBaseObjectRegion.
llvm-svn: 120173
2010-11-26 08:21:53 +00:00
Zhongxing Xu 33f13595c9 handle CXXFunctionalCastExpr in visitLValue and Environment.
llvm-svn: 120143
2010-11-25 03:18:57 +00:00
Nick Lewycky ef4f456866 Tie DefineVTablesUsed() in with recursive function instantiation so that we emit
a useful template instantiation stack. Fixes PR8640.

This also causes a slight change to where the "instantianted from" note shows up
in truly esoteric cases (see the change to test/SemaCXX/destructor.cpp), but
that isn't directly the fault of this patch.

llvm-svn: 120135
2010-11-25 00:35:20 +00:00
Anders Carlsson 0c75b49dc1 Fix a (probably very old) regression where we weren't using the typedef name for anonymous tag types.
llvm-svn: 120113
2010-11-24 21:05:09 +00:00
Zhongxing Xu 8be65a792f When getting CXXThisRegion from CXXMethodDecl, use the qualifiers. This is
to be consistent with the type of 'this' expr in the method.
此行及以下内容将会被忽略--

M    test/Analysis/method-call.cpp
M    include/clang/Checker/PathSensitive/GRExprEngine.h
M    lib/Checker/GRCXXExprEngine.cpp

llvm-svn: 120094
2010-11-24 13:48:50 +00:00
John McCall 8a1013f8c3 Redeclarations of using declarations are not okay in function scopes.
Not sure what I was thinking before.

Fixes PR8668.

llvm-svn: 120063
2010-11-23 22:03:51 +00:00
John McCall 20c00e0e7b This test does not need to test the alignment of pointers. Make it pass
on 64-bit systems.

llvm-svn: 120014
2010-11-23 10:15:48 +00:00
Francois Pichet ce87a7e45c long => int to make the test pass on system where sizeof(long) = 64.
llvm-svn: 120005
2010-11-23 07:15:13 +00:00
Francois Pichet 0c71f6c5d3 Microsoft C anonymous struct implementation.
Documentation: http://msdn.microsoft.com/en-us/library/z2cx9y4f.aspx

llvm-svn: 120000
2010-11-23 06:07:27 +00:00
Nico Weber 13c13d4da8 Remove one I just added, add a more focused test for why the current code is correct.
llvm-svn: 119969
2010-11-22 13:12:28 +00:00
Nico Weber 774303425b Fix the source range of CXXNewExprs. Fixes http://llvm.org/pr8661.
llvm-svn: 119966
2010-11-22 10:30:56 +00:00
NAKAMURA Takumi 3c3b3a7d8b test/Sema/x86-builtin-palignr.c: Now we can remove XFAIL:win32, according to r119958.
llvm-svn: 119961
2010-11-22 09:29:49 +00:00
Zhanyong Wan 6dace61730 Fix PR8419. Reviewed by kremenek and xuzhongxing.
llvm-svn: 119960
2010-11-22 08:45:56 +00:00
Chandler Carruth 45c2fb1e69 Undo part of my previous commit to mm_malloc.h, going back to the use of
stdlib.h. There were numerous problems with forward declaring 'malloc' and
'free', but the most important is that these are reserved by POSIX and may be
implemented via a function-like macro.

As suggested by Dale Johannesen, I'm instead guarding the only include of this
in our builtin headers with __STDC_HOSTED__, and I've removed the include of
the header from the test suite. I'll discuss with folks whether we want to have
a hosted section of the test suite or not, and add it (and perhaps other tests)
back there if that's the direction.

llvm-svn: 119958
2010-11-22 08:06:31 +00:00
Anders Carlsson a369f8dc3a The 'X' printf type has a valid alternative form. Fixes PR8641.
llvm-svn: 119946
2010-11-21 18:34:21 +00:00
Zhongxing Xu 2c966716ff Handle CFGAutomaticObjDtor.
llvm-svn: 119897
2010-11-20 06:53:12 +00:00
Argyrios Kyrtzidis 90ee2a4ecf Don't warn for empty 'if' body if there is a macro that expands to nothing, e.g:
if (condition)
    CALL(0); // empty macro but don't warn for empty body.

Fixes rdar://8436021.

llvm-svn: 119838
2010-11-19 20:54:25 +00:00
Fariborz Jahanian e7d62020a7 objc_msgSend is not a builtin type in non-objc mode.
Fixes //rdar://8686888

llvm-svn: 119813
2010-11-19 18:16:46 +00:00
Douglas Gregor 3a001f48e4 When parsing something that looks like an ill-formed
protocol-qualifier list without a leading type (e.g., <#blah#>), don't
complain about it being an archaic protocol-qualifier list unless it
actually parses as one.

llvm-svn: 119805
2010-11-19 17:10:50 +00:00
Douglas Gregor a750e8e6d8 Extend the libclang diagnostic API to provide information about the
option name, category ID, and category name corresponding to a diagnostic.

llvm-svn: 119802
2010-11-19 16:18:16 +00:00
Fariborz Jahanian 148d113e55 Fix a bug where write-barriers for assignment through reference
types was not being generated for objc pointers.
// rdar://8681766.

llvm-svn: 119751
2010-11-18 22:39:16 +00:00
John McCall 4bc41aee2b Add an assertion, fix a whole bunch of bugs, comment the assertion
out because there are still bugs left.

llvm-svn: 119722
2010-11-18 19:01:18 +00:00
John McCall 7decc9e4ea Calculate the value kind of an expression when it's created and
store it on the expression node.  Also store an "object kind",
which distinguishes ordinary "addressed" l-values (like
variable references and pointer dereferences) and bitfield,
@property, and vector-component l-values.

Currently we're not using these for much, but I aim to switch
pretty much everything calculating l-valueness over to them.
For now they shouldn't necessarily be trusted.

llvm-svn: 119685
2010-11-18 06:31:45 +00:00
Anton Yartsev 3f8f2886c1 comparison of AltiVec vectors now gives bool result (fix for 7533)
llvm-svn: 119678
2010-11-18 03:19:30 +00:00
Argyrios Kyrtzidis ef6c8da50f -Rename -Wargument-larger-than -> -Wlarge-by-value-copy
-Improve the diagnostic message
-Add some comments

Suggestions by Chris.

llvm-svn: 119594
2010-11-18 00:20:36 +00:00
Argyrios Kyrtzidis af84ec0a96 Introduce option -Wargument-larger-than[=N] which warns about function definitions if they take by-value
or return by-value any POD that is larger than some threshold (default is 64 bytes).

Implements rdar://8548050.

llvm-svn: 119583
2010-11-17 23:11:54 +00:00
Douglas Gregor b4a7c0385d Do not produce "purely informative" code completion results for
Objective-C message sends, which have only whitespace in their
TypedText chunk. Such results have no purpose.

llvm-svn: 119569
2010-11-17 21:36:08 +00:00
Fariborz Jahanian 8046af7d50 Warn if direct accessing synthesized ivar backing the property in
nonofragile-abi2. Fixes //rdar://8673791

llvm-svn: 119543
2010-11-17 19:41:23 +00:00
Argyrios Kyrtzidis 56e879d1da Don't emit warn_logical_and_in_logical_or for cases like "a && b || 0".
llvm-svn: 119540
2010-11-17 19:18:19 +00:00
Argyrios Kyrtzidis 14a96625b8 Don't warn for parentheses for the '&&' inside '||' for cases like:
assert(a || b && "bad");

since this is safe. This way we avoid a big source of such warnings which in this case are practically useless.

Note that we don't handle *all* cases where precedence wouldn't matter because of constants since
this is a bit costly to check, and IMO clarifying precedence with parentheses is good for
readability in general.

llvm-svn: 119533
2010-11-17 18:26:36 +00:00
Douglas Gregor 30313cbb80 When libclang is walking a member access expression, don't walk into
an implicit "this"; it causes clang_getCursor() to find the implicit
"this" expression (which isn't written in the source!) rather than the
actual member.

llvm-svn: 119516
2010-11-17 17:15:08 +00:00
Douglas Gregor 29ee422861 When comparing the source range of a declaration against the region of
interest (e.g., as used by clang_getCursor()), count the
decl-specifier-seq as part of the source range, as we do for
clang_annotateTokens(). Makes clang_getCursor() work properly for the
result types of functions, for example.

llvm-svn: 119514
2010-11-17 17:14:07 +00:00
Douglas Gregor b1b71e50a0 For an Objective-C @synthesize statement, e.g.,
@synthesize foo = _foo;

keep track of the location of the ivar ("_foo"). Teach libclang to
visit the ivar as a member reference.

llvm-svn: 119447
2010-11-17 01:03:52 +00:00
Bob Wilson db04141f9a Re-enable test with modifications for Neon vector changes in clang.
I mistakenly thought that this was checking for vector name mangling, but
it is not.  Since we're no longer wrapping Neon vectors in structs, this
test can just return a vector directly.  There are already other tests for
that, so just to make this interesting, change the test to return a struct
of two vectors.

llvm-svn: 119434
2010-11-17 00:44:53 +00:00
Douglas Gregor 68dbaead7b Fix source-range information for Objective-C properties. Previously,
we were just getting a range covering only the property name, which is
certainly not correct (and broke token annotation, among other
things). 

Also, teach libclang about the relationship between
@synthesize/@dynamic and @property, so we get property name and
cursor-reference information for @synthesize and @dynamic.

llvm-svn: 119409
2010-11-17 00:13:31 +00:00
John McCall 5b26f65b3d Reset the lifetime-managed flag between emission of the agg conditional
branches.  Fixes PR8623.

llvm-svn: 119408
2010-11-17 00:07:33 +00:00
John McCall 4f29b49de1 Support compound complex operations as l-values in C++. Add a test
case based on CodeGen/volatile-1.c which tests the current C++
semantics, and note the many, many places we fall short of them.

llvm-svn: 119402
2010-11-16 23:07:28 +00:00
Argyrios Kyrtzidis f89a56c74b Warn about arg1 && arg2 || arg3, as GCC 4.3+ does. Fixes rdar://8659922
llvm-svn: 119381
2010-11-16 21:00:12 +00:00
Bob Wilson 849857445f Temporarily disable this check for Neon vector type mangling.
I'll reenable it soon when I'm done reworking <arm_neon.h>.

llvm-svn: 119366
2010-11-16 19:08:19 +00:00
Ted Kremenek 0c27bcfd05 Static analyzer: Catch calls to malloc() with
allocation sizes of 0 bytes.

Fixes PR 2899.

llvm-svn: 119364
2010-11-16 18:47:04 +00:00
Argyrios Kyrtzidis 20ee5ae871 Emit a specific diagnostic when typedefing C++ bool, mirroring gcc.
Fixes rdar://8365458

llvm-svn: 119359
2010-11-16 18:18:13 +00:00
Chandler Carruth b19289331d This really seems like a boring set of fixes to our tests to make them more
independent of the underlying system. Let me know if any of these are too
aggressive.

llvm-svn: 119345
2010-11-16 10:26:08 +00:00
John McCall 07bb19667a Simplify some complex emission and implement correct semantics for
assignment to volatiles in C.  This in effect reverts some of mjs's
work in and around r72572.  Basically, the C++ standard is quite
clear, except that it lies about volatile behavior approximating
C's, whereas the C standard is almost actively misleading.

llvm-svn: 119344
2010-11-16 10:08:07 +00:00
Chandler Carruth 1c8383dccd Fix PR8625 and correctly interpret member-calls to static members when
producing warnings.

This feels really fragile, and I've not audited all other argument index-based
warnings. I suspect we'll grow this bug on another warning eventually. It might
be nice to adjust the argument indices when building up the attribute AST node,
as we already have to remember about the 'this' argument within that code to
produce correct errors.

llvm-svn: 119340
2010-11-16 08:49:43 +00:00
Chandler Carruth 743682bb9f Re-work the handling of implicit 'this' arguments and silly GCC-style attribute
argument indexes. This handles the offsets in a consistent manner for all of
the attributes which I saw working with these concepts. I've also added tests
for the attribute that motivated this: nonnull.

I consolidated the tests for format attributes into one file, and fleshed them
out a bit to trigger more of the warning cases. Also improved the quality of
some of the diagnostics that occur with invalid argument indices.

The only really questionable change here is supporting the implicit this
argument for the ownership attribute. I'm not sure it's really a sensible
concept there, but implemented the logic for consistency.

llvm-svn: 119339
2010-11-16 08:35:43 +00:00
Argyrios Kyrtzidis 9813d3221d Improve diagnostic for calling non-const method on const object. Fixes rdar://7743000
llvm-svn: 119336
2010-11-16 08:04:45 +00:00
Zhongxing Xu 1ade3267d7 Handle member initializer in C++ ctor.
- Add a new Kind of ProgramPoint: PostInitializer.
 - Still use GRStmtNodeBuilder. But special handling PostInitializer in 
   GRStmtNodeBuilder::GenerateAutoTransition().
 - Someday we should clean up the interface of GRStmtNodeBuilder.

llvm-svn: 119335
2010-11-16 07:52:17 +00:00
Bob Wilson bafdb73e79 Add a test for Neon vector mangling
llvm-svn: 119304
2010-11-16 00:32:31 +00:00
Bob Wilson 1d54ee75da Add tests for new Neon vector type attributes.
llvm-svn: 119303
2010-11-16 00:32:28 +00:00
John McCall 36459fdbad Add another case to the whitelist of cast kinds that can convert to bool.
Fixes PR8608.

llvm-svn: 119293
2010-11-16 00:12:50 +00:00
Ted Kremenek 18fb1665bf Remove invalid assertion from CFG builder. When building the CFG pieces for a ternary '?' expression,
it is possible for the confluence block to only have a single predecessor due to calls to 'noreturn'
functions.  Fixes assertion failure reported in PR 8619.

llvm-svn: 119284
2010-11-15 22:59:22 +00:00
Ted Kremenek db73d599b7 Relax assertion in SValuator so that we don't crash when analyzing a call via a function pointer that
casts the return value to something completely different.  While we need better reasoning here,
we should definately not crash.

llvm-svn: 119177
2010-11-15 20:09:42 +00:00
Fariborz Jahanian c961119dcd Restore these tests. I think I fixed the problem.
We shall see.

llvm-svn: 119158
2010-11-15 17:06:17 +00:00
Douglas Gregor ee430a39c3 Implement C++0x [temp.func.order]p3 (aka DR532) properly. In
particular, we only add the implement object parameter type if only
one of the function templates is a non-static member function
template.

Moreover, since this DR differs from existing practice in C++98/03,
this commit implements the existing practice (which ignores the
first parameter of the function template that is not the non-static
member function template) in C++98/03 mode.

llvm-svn: 119145
2010-11-15 15:41:16 +00:00
Zhongxing Xu fee455fcf5 Revert r118991.
Elidable CXXConstructExpr should inhibit calling destructor for temporary 
that is copied, not the one created. This is because eliding copy constructor 
means that the object that was to be copied will be constructed directly in 
memory the copy would be constructed in.

llvm-svn: 119044
2010-11-14 15:23:50 +00:00
John McCall 83fe49d1e1 Always emit full loads from volatile complex l-values.
Return the result of a complex assignment with the original values,
not by performing a load from the l-value;  this is the correct
semantics in C, although not in C++.

llvm-svn: 119037
2010-11-14 09:40:28 +00:00
NAKAMURA Takumi e82858299c test/CodeGenCXX/dyncast.cpp: Remove XFAIL:win32 with tweaking llvm/utils/FileCheck in r119023, due to DOSish crlf issue.
llvm-svn: 119024
2010-11-14 03:29:22 +00:00
Fariborz Jahanian 9b4dc09746 Removing these test till redo them so they won't depend
on the built compiler.

llvm-svn: 119013
2010-11-13 23:04:46 +00:00
Fariborz Jahanian e988bdac44 Block API patch to do copy ctor of copied-in cxx objects in
copy helper function and dtor of copied cxx objects
in dispose helper functions. __block variables
TBD next.

llvm-svn: 119011
2010-11-13 21:53:34 +00:00
Douglas Gregor 052caec1f3 When complaining about ambiguous overload resolution for a unary or
binary operator, provide the types.

llvm-svn: 119008
2010-11-13 20:06:38 +00:00
Douglas Gregor 72ebdabc5d When we're type-checking the result of calling a conversion function
(while computing user conversion sequences), make sure that a result
of class type is a complete class type. Had we gone through
ActOnCallExpr, this would have happened when we built the CallExpr.

Fixes PR8425.

llvm-svn: 119005
2010-11-13 19:36:57 +00:00
Chandler Carruth 456daba135 Switch to using %clang_cc1 properly rather than manually passing the -cc1 flag.
llvm-svn: 118996
2010-11-13 10:19:35 +00:00
Zhongxing Xu 1b038fa00f Do not add implicit dtors for CXXBindTemporaryExpr with elidable
CXXConstructExpr.

llvm-svn: 118991
2010-11-13 07:30:59 +00:00
Douglas Gregor 52773dcef6 Implement C++ [over.match.funcs]p4 as it concerns partial ordering of
function templates. Fixes PR8130.

llvm-svn: 118944
2010-11-12 23:44:13 +00:00
John McCall 31f82720d0 Replace one hack with a different hack: strip out the ObjectType
parameters to the Transform*Type functions and instead call out
the specific cases where an object type and the unqualified lookup
results are important.  Fixes an assert and failed compile on
a testcase from PR7248.

llvm-svn: 118887
2010-11-12 08:19:04 +00:00
Ted Kremenek bb437a6999 Fix use of an uninitialized SourceLocation because DeclarationNameLoc failed to completely zero-initialize itself.
Now we explicitly memset all of its values.

This bug was uncovered by the 'Index/recursive-cxx-member-calls.cpp', which exhibited an assertion
on an i386 darwin build of clang.  Adding this test case back since the assertion is now resolved.

llvm-svn: 118881
2010-11-12 04:25:07 +00:00
Douglas Gregor cbd0710a12 When performing initialization of a copy of a temporary object, use
direct-initialization (rather than copy-initialization) to initialize
the temporary, allowing explicit constructors. Fixes PR8342.

llvm-svn: 118880
2010-11-12 03:34:06 +00:00
Ted Kremenek 843c53828f RegionStore/BasicStore: do not return UndefinedVal for accesses to concrete addresses; instead return UnknownVal. This
leads it up to checkers (e.g., DereferenceChecker) to guard against illegal accesses (e.g., null dereferences).

Fixes PR 5272 and <rdar://problem/6839683>.

llvm-svn: 118852
2010-11-11 23:10:10 +00:00
Ted Kremenek b76f2402b0 Temporarily remove this test. It is causing an assertion failure in the builtbot.
llvm-svn: 118785
2010-11-11 08:17:57 +00:00
Ted Kremenek c7a5bae597 Annotate tokens in a separate thread to avoid blowing out stack space. While the CursorVisitor
is gradually becoming more data recursive, AnnotateTokensVisitor does its own recursive call
within the visitor that can still blow out the stack.  This can potentially be reworked to avoid this,
but for now just do token annotation on a separate thread.

llvm-svn: 118783
2010-11-11 08:05:23 +00:00
Ted Kremenek 92209a45b9 Generalize data-recursive visitation in CursorVisitor to also handle MemberExprs
and CXXCallMemberExprs.  This scheme is hopefully general enough to extend to the
rest of the visitor if necessary.

llvm-svn: 118782
2010-11-11 08:05:18 +00:00
John McCall deebbcf20d Undo a refactor-o and base the bitfield-truncation warning on the
uncoerced value.  Also, whitelist bool bitfields, which aren't
really a truncation.

llvm-svn: 118778
2010-11-11 05:33:51 +00:00
NAKAMURA Takumi 78e3fdb8a7 test/CMakeLists.txt: Use ${LLVM_LIT_ARGS} and remove redundant arguments "-sv" and "--no-progress-bar".
llvm-svn: 118777
2010-11-11 04:09:51 +00:00
John McCall 1f425648ca Extend the bitfield-truncation warning to initializations.
rdar://problem/8652606

llvm-svn: 118773
2010-11-11 03:21:53 +00:00
Douglas Gregor 44c6ee7729 Improve ASTUnit's capture of diagnostics so that the
diagnostic-capturing client lives as long as the ASTUnit itself
does. Otherwise, we can end up with crashes when we get a diagnostic
outside of parsing/code completion. The circumstances under which this
happen are really hard to reproduce, because a file needs to change
from under us.

llvm-svn: 118751
2010-11-11 00:39:14 +00:00
Fariborz Jahanian 2a5deb56a4 Adding couple of Block API, a bug fix and
a test change, all for blocks. wip.

llvm-svn: 118745
2010-11-11 00:11:38 +00:00
John McCall 817d4af557 When -Wconversion computes the range of a type, it uses the (bit-)range
of the enumerators rather than the actual expressible range.  This is 
great when dealing with opaque *values* of that type, but when computing
the range of the type for purposes of converting *into* it, it produces
warnings in cases we don't care about (e.g. enum_t x = 500;).  Divide 
the logic into these two cases and use the more conservative range for
targets.    

llvm-svn: 118735
2010-11-10 23:38:19 +00:00
Devang Patel 82abab0812 test case for r118726.
llvm-svn: 118727
2010-11-10 22:19:57 +00:00
Douglas Gregor 869853eea1 Instantiate class member template partial specialization declarations
in the order they occur within the class template, delaying
out-of-line member template partial specializations until after the
class has been fully instantiated. This fixes a regression introduced
by r118454 (itself a fix for PR8001).

llvm-svn: 118704
2010-11-10 19:44:59 +00:00
Fariborz Jahanian a0a9d85a5d Check for duplicate declaration of a property in current and
other class extensions. // rdar://7629420

llvm-svn: 118689
2010-11-10 18:01:36 +00:00
John McCall ad31b5f2cb Propagate the deprecated and unavailable attributes from a
@property declaration to the autogenerated methods.  I'm uncertain
whether this should apply to attributes in general, but these are
a reasonable core.

Implements rdar://problem/8617301

llvm-svn: 118676
2010-11-10 07:01:40 +00:00
Argyrios Kyrtzidis 2703bebc58 Replace UsingDecl's SmallPtrSet of UsingShadowDecls with a linked list to avoid leaking memory.
Fixes rdar://8649963.

llvm-svn: 118674
2010-11-10 05:40:41 +00:00
John McCall a8987a294d Friend function declarations can overload with tag declarations.
Fixes PR7915.

llvm-svn: 118670
2010-11-10 03:01:53 +00:00
John McCall 9b72f89f0f Diagnose attempst to template using declarations and using directives.
Recover from the latter and fail early for the former.  Fixes PR8022.

llvm-svn: 118669
2010-11-10 02:40:36 +00:00
John McCall fd81c52947 Tweak to bitfield-overflow warning: don't warn about storing
a positive value into a signed bitfield of the exact width of
the value.

llvm-svn: 118657
2010-11-10 00:26:50 +00:00
John McCall 0e6cfaf6c0 Typo.
llvm-svn: 118649
2010-11-09 23:36:43 +00:00
John McCall d2a5312e14 Add a warning for implicit truncation of constant values due to
bitfield assignment.

Implements rdar://problem/7809123

llvm-svn: 118647
2010-11-09 23:24:47 +00:00
John McCall 18a2c2c0b9 Split out -Wconversion warnings about constant precision into their
own subcategory, -Wconstant-conversion, which is on by default.

Tweak the constant folder to give better results in the invalid
case of a negative shift amount.

Implements rdar://problem/6792488

llvm-svn: 118636
2010-11-09 22:22:12 +00:00
Fariborz Jahanian b332ba3a51 Test moved to its correct place.
llvm-svn: 118635
2010-11-09 21:40:30 +00:00
Fariborz Jahanian d11da7e52a Restore patch reversed in r118475. Fixes
// rdar://8632525

llvm-svn: 118634
2010-11-09 21:38:20 +00:00
Douglas Gregor 0124e9ba9a Attempt to resolve overloaded functions in comma expressions and
conditional operators. Fixes PR7863.

llvm-svn: 118631
2010-11-09 21:07:58 +00:00
Fariborz Jahanian dbee986290 Issues good diagnostic when @end is missing.
// rdar://8283484

llvm-svn: 118629
2010-11-09 20:38:00 +00:00
Chris Lattner 0786544ac5 fix PR8380, a crash on invalid due to an illogical DeclSpec SourceRange being constructed.
llvm-svn: 118625
2010-11-09 20:14:26 +00:00
Douglas Gregor 928479e727 Revert the fix for PR8013.
That bug concerned the well-formedness of code such as (&ovl)(a, b,
c). GCC rejects the code, while EDG accepts it. On further study of the
standard, I see no support for EDG's position: in particular, C++
[over.over] does not list this as a context where we can take the
address of an overloaded function, C++ [over.call.func] does not
reference the address-of operator at any point, and C++ [expr.call]
claims that the function argument in a call is either a function
lvalue or a pointer-to-function; (&ovl) is neither.

llvm-svn: 118620
2010-11-09 20:03:54 +00:00
Charles Davis 99202b358f Use the right calling convention when mangling names in the Microsoft C++
mangler. Now member functions and pointers thereof have their calling
convention mangled as __thiscall if they have the default CC (even though,
they technically still have the __cdecl CC).

llvm-svn: 118598
2010-11-09 18:04:24 +00:00
Douglas Gregor 8f225bb508 Handle overload resolution when calling an overloaded function set
with, e.g., (&f)(a, b, c). Fixes PR8013.

llvm-svn: 118508
2010-11-09 16:13:15 +00:00
Douglas Gregor ba57154614 Make #pragma unused work for static local variables.
llvm-svn: 118500
2010-11-09 14:57:47 +00:00
Douglas Gregor 229bebdeb5 ntroduce clang_getSpellingLocation() into libclang, to provide the
location where we're spelling a token even within a
macro. clang_getInstantiationLocation() tells where we instantiated
the macro.

I'm still not thrilled with the CXSourceLocation/CXSourceRange APIs,
since they gloss over macro-instantiation information.

Take 2: this time, adjusted tests appropriately and used a "simple"
approach to the spelling location.

llvm-svn: 118495
2010-11-09 06:24:54 +00:00
Douglas Gregor 12a610d1fa Revert r118492, which didn't update all of its tests accordingly
llvm-svn: 118494
2010-11-09 05:52:02 +00:00
Douglas Gregor d1ea3f0d29 Introduce clang_getSpellingLocation() into libclang, to provide the
location where we're spelling a token even within a
macro. clang_getInstantiationLocation() tells where we instantiated
the macro.

I'm still not thrilled with the CXSourceLocation/CXSourceRange APIs,
since they gloss over macro-instantiation information.

llvm-svn: 118492
2010-11-09 05:28:47 +00:00
Douglas Gregor 2927c0c817 Teach code completion not to include out-of-line declarations and
definitions in its results. The original declarations will be visible
wherever they are declared.

llvm-svn: 118484
2010-11-09 03:59:40 +00:00
Argyrios Kyrtzidis 8b4cf00cdc Fix the test for Release.
llvm-svn: 118483
2010-11-09 03:43:49 +00:00
Douglas Gregor 4aff944f1a Add __nullptr as a C++0x nullptr literal, available in C++98
llvm-svn: 118482
2010-11-09 03:43:04 +00:00
Douglas Gregor 23ab745e79 Fix source locations in unnamed bitfield diagnostic, from Jakub
Wieczorek! Fixes PR8025.

llvm-svn: 118481
2010-11-09 03:31:16 +00:00
Fariborz Jahanian 1e2d01e77d Remove this test for now.
llvm-svn: 118476
2010-11-09 02:24:38 +00:00
Fariborz Jahanian 3655e59688 Reverse patch for // rdar://8632525. It might
has broken a build.

llvm-svn: 118475
2010-11-09 02:24:08 +00:00
Fariborz Jahanian 115393ce66 Remove space from rdar URIs.
llvm-svn: 118474
2010-11-09 02:16:57 +00:00
Ted Kremenek dcf85a8d18 Teach AttrNonNullChecker about transparent unions. Fixes crash reported in <rdar://problem/8642434>.
llvm-svn: 118473
2010-11-09 02:11:43 +00:00
Argyrios Kyrtzidis 8909686b22 Fix miscompilation regarding VLAs; subscription of VLA pointers was incorrect.
Fixes rdar://8644873 & http://llvm.org/PR8567.

llvm-svn: 118468
2010-11-09 01:30:48 +00:00
John McCall 478382521e When re-using a vtable slot for the nearest overridden method, just because
there's no return adjustment from the overridden to the overrider doesn't
mean there isn't a return adjustment from the overrider to the final
overrider.  This matters if we're emitting a virtual this-adjustment thunk
because the overrider virtually inherits from the class providing the
nearest overridden method.  Do the appropriate return adjustment in this case.

Fixes PR7611.

llvm-svn: 118466
2010-11-09 01:18:05 +00:00
Nick Lewycky 6147891648 Don't lose track of previous-declarations when instantiating a class template.
Fixes PR8001.

llvm-svn: 118454
2010-11-08 23:29:42 +00:00
Fariborz Jahanian 94ad274c24 Provide a precise builtin declaration for objc_msgSend
to avoid a bogus warning. Fixes //rdar: //8632525

llvm-svn: 118451
2010-11-08 22:53:18 +00:00
Ted Kremenek 15c1e368bb Add CursorVisitor::VisitBinaryOperator() to explicitly handle the case where we can blow out the stack due
to deeply nested BinaryOperators.  This is done by turning the explicit recursion into being data recursive.

Fixes: <rdar://problem/8289205>
llvm-svn: 118444
2010-11-08 21:14:15 +00:00
Douglas Gregor dc520b0f08 If an instance method messages is sending to an expression of type Class,
and we statically can compute a bound on the actual type (e.g.,
because it's a send to the the magic "class" instance method), code
complete as if we were performing a class message send to that class.

llvm-svn: 118443
2010-11-08 21:12:30 +00:00
John McCall f71b45367f Document Clang's support for attributes on individual enumerators and
tweak the documentation for deprecation-with-message.  Provide __has_feature
tests for both.  rdar://problem/8605692

llvm-svn: 118435
2010-11-08 19:48:17 +00:00
Argyrios Kyrtzidis 8566356acb When building a compound literal, check that the base element of the array is complete.
Fixes rdar://8620582 & http://llvm.org/PR7905

llvm-svn: 118428
2010-11-08 19:14:19 +00:00
Fariborz Jahanian 086a24a2be Fixes a rewrite bug, rewriting nested property usage
inside blocks. Fixes //rdar: //8608293.

llvm-svn: 118425
2010-11-08 18:37:50 +00:00
Douglas Gregor bd6b17f4ef Improve our handling of C++ [class.copy]p3, which specifies that a
constructor template will not be used to copy a class object to a
value of its own type. We were eliminating all constructor templates
whose specializations look like a copy constructor, which eliminated
important candidates. Fixes PR8182.

llvm-svn: 118418
2010-11-08 17:16:59 +00:00
Douglas Gregor bcd625305b When attempting reference binding to an overloaded function, also
consider that we might be trying to bind a reference to a class type,
which involves a constructor call. Fixes PR7425.

llvm-svn: 118407
2010-11-08 15:20:28 +00:00
Douglas Gregor b0a746e391 Add missing test case for PR8230
llvm-svn: 118402
2010-11-08 03:58:01 +00:00
Douglas Gregor e81f58e180 Properly diagnose invalid casts to function references. Patch by
Faisal Vali, tweaked by me. Fixes PR8230.

llvm-svn: 118400
2010-11-08 03:40:48 +00:00
NAKAMURA Takumi 2fbf57a949 test/CodeGenCXX/rtti-linkage.cpp: Get rid of sort(1) to remove XFAIL: win32.
On the certain system, bogus SORT.EXE is picked up. Its sort-order is incompatible to POSIX.

llvm-svn: 118391
2010-11-08 01:05:10 +00:00
Douglas Gregor 1ccc8416a0 Remove broken support for variadic templates, along with the various
abstractions (e.g., TemplateArgumentListBuilder) that were designed to
support variadic templates. Only a few remnants of variadic templates
remain, in the parser (parsing template type parameter packs), AST
(template type parameter pack bits and TemplateArgument::Pack), and
Sema; these are expected to be used in a future implementation of
variadic templates.

But don't get too excited about that happening now.

llvm-svn: 118385
2010-11-07 23:05:16 +00:00
Rafael Espindola c8f008f649 Use ld directly on linux. Changes from the previous try:
*) Try to detect as much as possible from the system itself, not the distro.
   This should make it easier to port to a new distro and more likely to
   work on a unknown one.
*) The distro enum now doesn't include the arch. Just use the existing
   host detection support in LLVM.
*) Correctly handle --sysroot.

A small regression is that now clang will pass bitcode file to the linker.
This is necessary for the gold plugin support to work.

It might be better to detect this at configure/cmake time, but doing it in
c++ first is a lot easier.

llvm-svn: 118382
2010-11-07 20:14:31 +00:00
Anders Carlsson 2ceb347e8b A union cannot contain static data members or data members of reference type.
llvm-svn: 118381
2010-11-07 19:13:55 +00:00
John McCall 2faab302e0 ARM EH uses a different personality function in C.
llvm-svn: 118366
2010-11-07 02:35:25 +00:00
Anders Carlsson 3fd50319d3 Don't warn when matching %p to nullptr.
llvm-svn: 118344
2010-11-06 14:58:53 +00:00
John McCall cdf7ef5437 Simplify the logic for emitting guard variables for template static
data members by delaying the emission of the initializer until after
linkage and visibility have been set on the global.  Also, don't
emit a guard unless the variable actually ends up with vague linkage,
and don't use thread-safe statics in any case.

llvm-svn: 118336
2010-11-06 09:44:32 +00:00
Argyrios Kyrtzidis 7d391557cc When determining which template partial specialization is more specialized,
make sure to setup the instantiation stack. Fixes rdar://8620775 & http://llvm.org/PR8234

llvm-svn: 118314
2010-11-05 23:25:18 +00:00
Douglas Gregor 25edf4302f When searching for an instantiated declaration requires instantiation
of its parent context, be sure to update the parent-context pointer
after instantiation. Fixes two anonymous-union instantiation issues in
<rdar://problem/8635664>.

llvm-svn: 118313
2010-11-05 23:22:45 +00:00
Argyrios Kyrtzidis 452707c125 Read/write from/to PCH the diagnostic mappings that the user set so that e.g. #pragma clang diagnostic can be used in a PCH.
Fixes rdar://8435969.

llvm-svn: 118303
2010-11-05 22:10:18 +00:00
Douglas Gregor 16443fdd31 Teach clang_getCursorReferenced() that a
CXXConstructorExpr/CXXTemporaryObjectExpr references the constructor
it calls. Then, tweak clang_getCursor() to prefer such a call over a
type reference to the type being called.

llvm-svn: 118297
2010-11-05 21:11:19 +00:00
Fariborz Jahanian 90d2e57ff7 Proper rewriting of block envokation with
qualified ObjC pointer types in its argument list.
// rdar: //8608902

llvm-svn: 118286
2010-11-05 18:34:46 +00:00
Anders Carlsson e981a8c231 Expressions of type std::nullptr_t can be used as sentinels.
llvm-svn: 118276
2010-11-05 15:21:33 +00:00
Anders Carlsson 7da7cc568f Implement [over.ics.rank]p4: A conversion that does not convert an std::nullptr_t to bool is better than one than does.
llvm-svn: 118269
2010-11-05 00:12:09 +00:00
Anders Carlsson 898b1fb1e2 Pass a -I flag when compiling nullptr.cpp so the typeinfo header can be found. This should fix the errors seen on the bot.
llvm-svn: 118246
2010-11-04 12:18:32 +00:00
Argyrios Kyrtzidis 4b52007c35 Don't be so eager to replace UsingDecls in a DeclContext's lookup table;
check that the TargetNestedNameDecl is the same first.

llvm-svn: 118239
2010-11-04 08:48:52 +00:00
Anders Carlsson ba37e1eb49 std::nullptr_t is a fundamental type for RTTI purposes.
llvm-svn: 118238
2010-11-04 05:28:09 +00:00
Anders Carlsson 2683fd6f8c Mangle std::nullptr_t as specified by the Itanium C++ ABI.
llvm-svn: 118236
2010-11-04 04:31:32 +00:00
Argyrios Kyrtzidis 9a94d9b876 Don't instantiate members not belonging in the semantic context of the template.
e.g. for:

    template <int i> class A {
      class B *g;
    };

'class B' has the template as lexical context but semantically it is
introduced in namespace scope.

Fixes rdar://8611125 & http://llvm.org/PR8505

llvm-svn: 118235
2010-11-04 03:18:57 +00:00
Anders Carlsson a95069c52f It's OK to use nullptr in relational operators if the other side is a null pointer constant.
llvm-svn: 118234
2010-11-04 03:17:43 +00:00
Douglas Gregor 3ade5704fe When canonicalizing nested-name-specifiers involving dependent names
or dependent specializations, rip apart the dependent name/dependent
specialization to recanonicalize its pieces, because
nested-name-specifiers store "dependent-type::identifier" differently
than types do. Fixes PR7419.

llvm-svn: 118211
2010-11-04 00:09:33 +00:00
Fariborz Jahanian 733dde6035 Fixes a rewriting bug when type in a block argument type
is a objc qualified class type. // rdar: //8608902

llvm-svn: 118208
2010-11-03 23:50:34 +00:00
Fariborz Jahanian 147e1cbb49 Patch to rewrite objc qualified types which occur in
block pointer type arguments. Partial fix for 
// rdar: //8608902

llvm-svn: 118205
2010-11-03 23:29:24 +00:00
Argyrios Kyrtzidis 71731d6b05 Implement -working-directory.
When -working-directory is passed in command line, file paths are resolved relative to the specified directory.
This helps both when using libclang (where we can't require the user to actually change the working directory)
and to help reproduce test cases when the reproduction work comes along.

--FileSystemOptions is introduced which controls how file system operations are performed (currently it just contains
 the working directory value if set).
--FileSystemOptions are passed around to various interfaces that perform file operations.
--Opening & reading the content of files should be done only through FileManager. This is useful in general since
 file operations will be abstracted in the future for the reproduction mechanism.

FileSystemOptions is independent of FileManager so that we can have multiple translation units sharing the same
FileManager but with different FileSystemOptions.

Addresses rdar://8583824.

llvm-svn: 118203
2010-11-03 22:45:23 +00:00
Nick Lewycky 07e97c594d Make this error less specific but also less likely to cause confusion. Fixes
PR7702.

llvm-svn: 118181
2010-11-03 17:52:57 +00:00
Douglas Gregor b37c9af75d When producing overload candidates for binary built-in operators, keep
the sets of available conversions for the first and second arguments
separate. This is apparently the indent of C++ [over.built], and
reduces the number of overload candidates generated, eliminating some
ambiguities. Fixes PR8477.

llvm-svn: 118178
2010-11-03 17:00:07 +00:00
Zhongxing Xu 0806da8f28 fix test case.
llvm-svn: 118166
2010-11-03 11:24:56 +00:00
Marcin Swiderski 3ab17ad8ec Added generating destructors for temporary objects. Two cases I know of, that are not handled properly:
1. For statement: const C& c = C(0) ?: C(1) destructors generated for condition will not differ from those generated for case without prolonged lifetime of temporary,
2. There will be no destructor for constant reference member bound to temporary at the exit from constructor.

llvm-svn: 118158
2010-11-03 06:19:35 +00:00
Anders Carlsson 2d5f8b4201 When setting a tied check if it's already tied. If it's tied to another constraint it's invalid. Fixes PR3905.
llvm-svn: 118146
2010-11-03 02:54:51 +00:00
Anders Carlsson da1f5fc881 A number in an input constraint can't point to a read-write ('+') constraint. Fixes PR3904.
llvm-svn: 118145
2010-11-03 02:22:29 +00:00
Alexis Hunt a56cbccfc4 Provide an error when a non-identifier name (such as an operator) is used as a
parameter name.

Fixes PR8012.

llvm-svn: 118138
2010-11-03 01:07:06 +00:00
Douglas Gregor 15417cfa56 Improve source-location information for CXXConstructExpr nodes, by
ensuring that they cover all of their child nodes. There's still a
clang_getCursor()-related issue with CXXFunctionalCastExprs with
CXXConstructExprs as children (see FIXME in the test case); I'll look
at that separately.

llvm-svn: 118132
2010-11-03 00:35:38 +00:00
Ted Kremenek d40a439067 Hack to workaround deficiency in ObjC ASTs. Functions and variables may be declared
within an @implementation, but we have no way to record that information in the AST.
This may cause CursorVisitor to miss these Decls when doing a AST walk.

Fixes <rdar://problem/8595462>.

llvm-svn: 118109
2010-11-02 23:10:24 +00:00
John Thompson 6c2afc8d5d Tests for inline asm constraints, generic and x86.
llvm-svn: 118108
2010-11-02 23:03:52 +00:00
Fariborz Jahanian dba351df64 Fixes an assertion violation when bind to temporary
expression is a dependent expression. 
// rdar: // 8620524 and PR7851

llvm-svn: 118066
2010-11-02 21:05:53 +00:00
John McCall 8e7cb6dcfa Ensure that static local variables in function templates inherit the
visibility of their function.

llvm-svn: 118065
2010-11-02 21:04:24 +00:00
Douglas Gregor 05fcf84d14 Teach code completion to provide property results when the property
can be used to automatically synthesize an ivar.

llvm-svn: 118052
2010-11-02 20:36:02 +00:00
John McCall 23032653d7 Unbreak private_extern, which apparently we had zero tests for.
llvm-svn: 118034
2010-11-02 18:38:13 +00:00
Fariborz Jahanian dc9bc5aeb8 Fix a crash mangling decayed val argument-typed function.
// rdar: //8620510 and PR7666

llvm-svn: 118019
2010-11-02 16:54:00 +00:00
Argyrios Kyrtzidis 3050d9bdb8 Properly handle temporaries that are created in a AsmStmt.
Previously the temporaries would get destroyed before the asm call.

llvm-svn: 118001
2010-11-02 02:33:08 +00:00
John McCall 07072664c4 Ignore attributes on classes when calculating visibility for members
with their own explicit visibility attributes.  Basically we only want to
apply a single visibility attribute from any particular ancestry.

llvm-svn: 117998
2010-11-02 01:45:15 +00:00
Fariborz Jahanian d4c5348a55 Diagnose a coherant message when @interface
does not terminate with @end.
// rdar: //7824372

llvm-svn: 117991
2010-11-02 00:44:43 +00:00
Douglas Gregor e0f7a8ace9 When performing template argument deduction against a template-id,
only keep deduction results for successful deductions, so that they
can be compared against each other. Fixes PR8462, from Richard Smith!

llvm-svn: 117983
2010-11-02 00:02:34 +00:00
Douglas Gregor e5a6fed7dc Make my test case test what it meant to
llvm-svn: 117974
2010-11-01 23:33:11 +00:00
Ted Kremenek 818e5c1839 Encapsulate within CXCursor the notion of whether a VarDecl* is the first Decl in a DeclGroup. We use this
to recover some context that is currently not modeled directly in the AST.  Currently VarDecl's cannot
properly determine their source range because they have no context on whether or not they appear in a DeclGroup.
For the meantime, this bandaid suffices in libclang since that is where the correct SourceRange is directly needed.

Fixes <rdar://problem/8595749>.

llvm-svn: 117973
2010-11-01 23:26:51 +00:00
Douglas Gregor 4b8eca88b0 Teach the CStringChecker and PthreadLockChecker about non-identifier
declaration names, from Jim Goodnow II!

llvm-svn: 117970
2010-11-01 23:16:05 +00:00
Argyrios Kyrtzidis 375584c7eb Clean up temporaries created by an asm statement. Fixes rdar://8540491
llvm-svn: 117961
2010-11-01 21:51:42 +00:00
Fariborz Jahanian eb7714c290 Restore patch for PR8453 which was undone in r117829
(was project bugs).

llvm-svn: 117951
2010-11-01 20:47:16 +00:00
Argyrios Kyrtzidis ca76629618 Emit error when using a bound member function for something other than calling it.
Also avoids IRGen crashes due to accepting invalid code.

llvm-svn: 117943
2010-11-01 18:49:26 +00:00
Douglas Gregor b524d906fe Require that the types of the parameters of a block literal are complete.
llvm-svn: 117942
2010-11-01 18:37:59 +00:00
Zhongxing Xu 627a1868d2 Now initializer of C++ record type is visited as block-level expr.
Let the destination of AggExprVisitor be an explicit MemRegion. 
Reenable the test case.

llvm-svn: 117908
2010-11-01 09:09:44 +00:00
Zhongxing Xu 0b51d4db8c Make all CXXConstructExpr's block-level expressions. This is required by
method inlining. Temporarily fail a test case.

llvm-svn: 117907
2010-11-01 06:46:05 +00:00
John McCall e6e622e789 Only apply -fvisibility-inlines-hidden to definitions. Apparently
isInlined() just gives meaningless results for non-definitions.

Fixes rdar://problem/8614470

llvm-svn: 117887
2010-11-01 01:29:57 +00:00
Anders Carlsson 1827509176 Teach the constant expr evaluator about derived-to-base casts when no virtual bases are involved. Fixes PR5974.
llvm-svn: 117868
2010-10-31 20:41:46 +00:00
Zhanyong Wan e4f7df694d Make Clang static analyzer skip function template definitions. This fixes Clang PR 8426, 8427, & 8433. Reviewed by Ted Kremenek and Doug Gregor.
llvm-svn: 117853
2010-10-31 04:22:34 +00:00
Anders Carlsson a5df61a341 Don't try to evaluate the LHS or RHS of a member pointer binary operation. Fixes PR8507.
llvm-svn: 117850
2010-10-31 01:21:47 +00:00
Argyrios Kyrtzidis d3f0054395 Emit an error when trying to form a pointer-to-member to a bitfield.
As a bonus, avoids a crash on the IRGen side due to accepting invalid code.

llvm-svn: 117842
2010-10-30 19:52:22 +00:00
Argyrios Kyrtzidis 52355c8020 renamed: test/SemaCXX/ptrtomember-badcall.cpp -> test/SemaCXX/ptrtomember.cpp
llvm-svn: 117841
2010-10-30 19:52:19 +00:00
Daniel Dunbar cda8eb7a21 Revert r117678, "Qualified 'id' should implement all of static class type's", it breaks things.
llvm-svn: 117829
2010-10-30 19:22:48 +00:00
John McCall c273f24196 Better solution: calculate the visibility of functions and variables
independently of whether they're definitions, then teach IR generation to
ignore non-explicit visibility when emitting declarations.  Use this to
make sure that RTTI, vtables, and VTTs get the right visibility.

More of rdar://problem/8613093

llvm-svn: 117781
2010-10-30 11:50:40 +00:00
John McCall 36cd5cc150 GCC faithfully calculates visibility for variables independently of
whether it's a declaration or not, then ignores that information for
declarations unless it was explicitly given.  It's not totally clear
how that should be mapped into a sane system, but make an effort.

llvm-svn: 117780
2010-10-30 09:18:49 +00:00
Nick Lewycky b9b3927356 Preserve the template type parameter name when instantiating a templace.
Fixes PR8489.

llvm-svn: 117776
2010-10-30 06:48:20 +00:00
Argyrios Kyrtzidis 93513155e7 Don't remove the init expression from the initializer list if it had a semantic error.
We already flag the error with InitListChecker's hadError and we mess up the AST unnecessarily.
Fixes rdar://8605381.

llvm-svn: 117760
2010-10-30 01:06:26 +00:00
Argyrios Kyrtzidis b5488f45db Add test/SemaCXX/crash-PR7625.cpp into test/SemaCXX/crashes.cpp
llvm-svn: 117759
2010-10-30 01:06:23 +00:00
Argyrios Kyrtzidis 0f24185c76 test/SemaCXX/crash-8124080.cpp -> test/SemaCXX/crashes.cpp
llvm-svn: 117758
2010-10-30 01:06:21 +00:00
Ted Kremenek c00d5b9706 Add test case for <rdar://problem/8610363> (a bogus report of using an uninitialized field). This was
already fixed, but this serves for detecting regressions.

llvm-svn: 117754
2010-10-30 00:43:15 +00:00
John McCall 6a16b2f10c When list-initializing a vector, try to copy-initialize from vectors instead
of descending into the subelements.

rdar://problem/8345836

llvm-svn: 117749
2010-10-30 00:11:39 +00:00
Fariborz Jahanian e5c1bda4d1 Improve diagnostics reporting of un-implemented
methods in protocols when protocols are in system
headers and thus ignored. //rdar: //8227199

llvm-svn: 117739
2010-10-29 23:20:05 +00:00
Douglas Gregor d4c5ed038c Make the deserialization of C++ base class specifiers lazy, improving
the performance of C++ PCH and reducing stack depth in the reader.

llvm-svn: 117732
2010-10-29 22:39:52 +00:00
John McCall 37bb6c9832 Restore r117644, this time properly ignoring -fvisibility and type visibility
for namespace-scope variable declarations.

Apply visibility in IR gen to variables that are merely declared
and never defined.  We were previously emitting these with default
visibility unless they were declared with private_extern.

Ignore global visibility settings when computing visibility for
a declaration's context, and key several conditions on whether a
visibility attribute exists anywhere in the hierarchy as opposed
to whether it exists at the current level.

llvm-svn: 117729
2010-10-29 22:22:43 +00:00
Rafael Espindola 5275559c2e Update tests to not search of as.
llvm-svn: 117711
2010-10-29 21:18:11 +00:00
Fariborz Jahanian 531a3707a9 Qualified 'id' should implement all of static class type's
protocols, including those added to class, super class
and categories; otherewise issue a warning. This fixes
pr8453.

llvm-svn: 117678
2010-10-29 18:26:21 +00:00
Argyrios Kyrtzidis 706bbf84d8 When doing name lookup for members don't look into global/namespace scope.
Better performance and fixes rdar://8603569.

llvm-svn: 117656
2010-10-29 16:12:50 +00:00
Daniel Dunbar df8792128f Revert r117644, "Apply visibility in IR gen to variables that are merely
declared", it breaks things.

llvm-svn: 117653
2010-10-29 15:19:36 +00:00
John McCall 077dc60540 Don't assert on attempts to throw 'bool'. I wonder if in the history of C++
anyone has ever intentionally done this outside of a compiler test case.

llvm-svn: 117645
2010-10-29 08:14:02 +00:00
John McCall 4af6bf1fdc Apply visibility in IR gen to variables that are merely declared
and never defined.  We were previously emitting these with default
visibility unless they were declared with private_extern.

Ignore global visibility settings when computing visibility for
a declaration's context, and key several conditions on whether a
visibility attribute exists anywhere in the hierarchy as opposed
to whether it exists at the current level.

llvm-svn: 117644
2010-10-29 07:49:41 +00:00
Ted Kremenek 310c5a8e31 Don't flag idempotent '+' or '-' warnings for pointer arithmetic (typically false positives).
Fixes <rdar://problem/8601243>.

llvm-svn: 117635
2010-10-29 01:06:54 +00:00
John McCall 033caa5895 When computing the visibility of a class member, calculate the visibility
of its context without considering global settings like -fvisibility=hidden.
Fixes PR8492.

llvm-svn: 117628
2010-10-29 00:29:13 +00:00
John McCall 565141612f When emitting l-values for bool non-__block decl references, make a pointer
using the memory type;  fixes an assert.

Fixes rdar://problem/8605032

llvm-svn: 117610
2010-10-28 21:37:57 +00:00
John McCall 62b6862c5e Don't apply -fvisibility-inlines-hidden to extern templates.
Part 2 of rdar://problem/8595231

llvm-svn: 117567
2010-10-28 18:10:36 +00:00
John McCall 9de9160d55 Implement an indirect-goto optimization for goto *&&lbl and respect this
in the scope checker.  With that done, turn an indirect goto into a
protected scope into a hard error;  otherwise IR generation has to start
worrying about declarations not dominating their scopes, as exemplified
in PR8473.

If this really affects anyone, I can probably adjust this to only hard-error
on possible indirect gotos into VLA scopes rather than arbitrary scopes.
But we'll see how people cope with the aggressive change on the marginal
feature.

llvm-svn: 117539
2010-10-28 08:53:48 +00:00
John McCall 2efaf11695 Abandon the type-visibility optimization for functions. GCC doesn't do it,
and it's too much trouble to push for.  Fixes PR8478.

llvm-svn: 117532
2010-10-28 07:07:52 +00:00
John McCall b7139c43ae When computing visibility, use the latest declaration's explicit visibility
attribute.

Part of rdar://problem/8595231

llvm-svn: 117526
2010-10-28 04:18:25 +00:00
Ted Kremenek eaae1adb94 Add test case for <rdar://problem/8356342>.
llvm-svn: 117525
2010-10-28 02:50:32 +00:00
John McCall 071df46743 Implement the newest status quo for method override checking. The idea now
is that we need more information to decide the exact conditions for whether
one ObjCObjectPointer is an acceptable return/parameter override for another,
so we're going to disable that entire class of warning for now.  The
"forward developement" warning category, -Wmethod-signatures, can receive
unrestricted feature work, and when we're happy with how it acts, we'll
turn it on by default.

This is a pretty conservative change, and nobody's totally content with it.

llvm-svn: 117524
2010-10-28 02:34:38 +00:00
Ted Kremenek 0abd85c7ae Don't warn about unamed bitfield ivars in the ObjCUnusedIvarsChecker. Fixes <rdar://problem/8481311>.
llvm-svn: 117521
2010-10-28 02:16:22 +00:00
Stuart Hastings c6d136cba1 Test passes on Darwin; try to XFAIL on freebsd, linux, xp/msvc9.
llvm-svn: 117515
2010-10-28 00:47:58 +00:00
Devang Patel 1bee63f325 Fix context info for enums.
Radar 8595129

llvm-svn: 117507
2010-10-27 23:23:58 +00:00
John McCall 3a7f6926d1 Restore r117403 (fixing IR gen for bool atomics), this time being less
aggressive about the form we expect bools to be in.  I don't really have
time to fix all the sources right now.

llvm-svn: 117486
2010-10-27 20:58:56 +00:00
Fariborz Jahanian 5701585780 patch to do array-to-pointer conversion in a
statement-expression. // rdar: //8600553

llvm-svn: 117479
2010-10-27 20:10:28 +00:00
Rafael Espindola 9d798a07e4 Revert r117403 as it caused PR8480.
llvm-svn: 117456
2010-10-27 17:13:49 +00:00
Anders Carlsson b00c2144b3 Also devirtualize calls to a member functions where the containing class has been marked final.
llvm-svn: 117445
2010-10-27 13:34:43 +00:00
Anders Carlsson a7911fa3d7 If a virtual member function has the 'final' attribute, we can devirtualize calls to it.
llvm-svn: 117444
2010-10-27 13:28:46 +00:00
Zhongxing Xu d95ccd58a9 If visiting RHS causes us to finish 'Block', e.g. the RHS is a StmtExpr
containing a DoStmt, and the LHS doesn't create a new block, then we should
return RBlock.  Otherwise we'll incorrectly return NULL.

Also relax an assertion in VisitWhileStmt(). Reset 'Block' when it is finished.

llvm-svn: 117436
2010-10-27 03:23:10 +00:00
Douglas Gregor 19ead876d2 Disable this test until we figure out what madness it causes
llvm-svn: 117416
2010-10-26 23:34:31 +00:00
Dan Gohman 2f76cd756a Add support for code completion on stdin.
llvm-svn: 117414
2010-10-26 23:21:25 +00:00
Fariborz Jahanian 67ca8c4c7b Patch to provide guard when initializing instances
of static data member of a class template.
Fixes //rdar :// 8562966 and pr8409.

llvm-svn: 117410
2010-10-26 22:47:47 +00:00
John McCall 6bde954f47 Extract procedures to do scalar-to-memory and memory-to-scalar conversions
in IR gen, and use those to fix a correctness issue with bool atomic
intrinsics.  rdar://problem/8461234

llvm-svn: 117403
2010-10-26 22:09:15 +00:00
Douglas Gregor bb55f6bd23 Sign
llvm-svn: 117400
2010-10-26 21:25:18 +00:00
Douglas Gregor 9c0672f40a I hate this test
llvm-svn: 117390
2010-10-26 19:36:20 +00:00
Douglas Gregor fc0862da95 Something is seriously wonky with this test
llvm-svn: 117381
2010-10-26 18:25:19 +00:00
Douglas Gregor 415ce06ebb Fix silly typo
llvm-svn: 117373
2010-10-26 17:32:59 +00:00
Douglas Gregor 2a1d72d078 Teach typo correction not to return the same keyword that matches a
typo. This can happen with context-sensitive keywords like "super",
when typo correction didn't know that "super" wasn't permitted in this
context.

llvm-svn: 117372
2010-10-26 17:18:00 +00:00
John McCall edba521a03 Cookie crumbs.
llvm-svn: 117356
2010-10-26 06:41:10 +00:00
John McCall 75851b1a7f Actually, that doesn't really work, and anyway we should choose
conversion to id over conversion to void*.

llvm-svn: 117355
2010-10-26 06:40:27 +00:00
John McCall f28e83a73a Consider conversions of Objective-C pointers to 'id' to be basically of
the same rank as conversions of normal pointers to 'void*'.

Also, resurrect a test case.

Fixes rdar://problem/8592139

llvm-svn: 117354
2010-10-26 06:23:29 +00:00
John McCall 5fe84129e2 A couple of tweaks to the visibility rules:
- tags with C linkage should ignore visibility=hidden
  - functions and variables with explicit visibility attributes should
    ignore the linkage of their types
Either of these should be sufficient to fix PR8457.

Also, FileCheck-ize a test case.

llvm-svn: 117351
2010-10-26 04:59:26 +00:00
John McCall 710d2a831c Pending further discussion, re-enable warnings for Objective C
covariant/contravariant overrides and implementations, but do so under
control of a new flag (-Wno-objc-covariant-overrides, which yes does cover
contravariance too).

*At least* the covariance cases will probably be enabled by default shortly,
but that's not totally uncontroversial.

llvm-svn: 117346
2010-10-26 00:53:53 +00:00
Douglas Gregor f86c939bff When de-serializing a type that is supposed to be canonical, call
getCanonicalType() to make sure that the type we got back is actually
canonical. This is the case for most types, which always build a
canonical type when given canonical components. However, some types that
involve expressions in their canonicalization (e.g., array types with
dependent sizes) don't always build canonical types from canonical
components, because there is no such thing as a "canonical"
expression. Therefore, we do this extra mapping to ensure that the
canonical types we store are actually canonical.

llvm-svn: 117344
2010-10-26 00:51:02 +00:00
Ted Kremenek 6348dac7a0 Fix lazy symbolication bug in RegionStore involving fields of global variables. When invalidated, the entire
globals memory space gets assigned a symbolic value, but that value was not being used for lazy symbolication
of fields of globals.  This could result in cases where bogus null dereferences were being reported.

Fixes PR 8440.

llvm-svn: 117336
2010-10-26 00:06:17 +00:00
Ted Kremenek 731310e13e Tweak null dereference checker to give better diagnostics for null dereferences resulting from array accesses.
llvm-svn: 117334
2010-10-26 00:06:13 +00:00
Fariborz Jahanian 56143ae715 Patch for mis-compile of statement expressions with
non-trivial copy constructors. // rdar: //8540501.
A test will be added to llvm nightly tests.

llvm-svn: 117324
2010-10-25 23:27:26 +00:00
Douglas Gregor 59f16ed082 Look through the address-of operator to find the function being
called. Fixes another aspect of PR8314.

llvm-svn: 117308
2010-10-25 20:48:33 +00:00
Douglas Gregor 8e09a72f13 When we're calling a function that we know based on its K&R-style
function definition, we should still use a prototype to type-check and
convert the function arguments, if such a prototype exists. Fixes
PR8314.

llvm-svn: 117305
2010-10-25 20:39:23 +00:00