Commit Graph

26226 Commits

Author SHA1 Message Date
Manuel Klimek 60b8016197 Allows retrieving all files in a CompilationDatabase.
Patch by Tobias Koenig, some test changes by myself.

llvm-svn: 160167
2012-07-13 12:31:45 +00:00
Richard Smith 57e7ff9c0c Provide a special-case diagnostic when two class member functions instantiate
to the same signature. Fix a bug in the type printer which would cause this
diagnostic to print wonderful types like 'const const int *'.

llvm-svn: 160161
2012-07-13 04:12:04 +00:00
Rafael Espindola eca5cd20a1 Fix a bug in my previous commit. The problem is not that we were not using the
canonical decl for the template, but that we were not merging attributes for
templates at all!

llvm-svn: 160157
2012-07-13 01:19:08 +00:00
Dmitri Gribenko e7bb944c85 Attaching comments to declarations during parsing: handle more Objective-C declarations.
llvm-svn: 160156
2012-07-13 01:06:46 +00:00
Dmitri Gribenko e00ffc7bb8 Comment parsing: repaint the bikesched: rename 'HTML open tags' to 'HTML start tags' and 'HTML close tags' to 'HTML end tags' according to HTML spec.
llvm-svn: 160153
2012-07-13 00:44:24 +00:00
Dmitri Gribenko 9460fbf219 Make concepts of optional and forbidden end tags separate. Thanks Jordan for pointing this!
llvm-svn: 160149
2012-07-12 23:37:09 +00:00
Rafael Espindola 6140847647 Use the canonical template decl when trying to find if it has a visibility
attribute.

llvm-svn: 160139
2012-07-12 20:05:04 +00:00
Timur Iskhodzhanov c5098ad371 [Windows] Use thiscall as the default calling convention for class methods. PR12785
llvm-svn: 160121
2012-07-12 09:50:54 +00:00
Benjamin Kramer a43b6999ff Add _rdrand{16,32,64}_step intrinsics to immintrin.h
llvm-svn: 160118
2012-07-12 09:33:03 +00:00
Benjamin Kramer 29f90e804d PR13312: Don't crash when printing a fixit that ends in a unicode character.
llvm-svn: 160112
2012-07-12 08:20:49 +00:00
Rafael Espindola 4dedd0ceb8 Process #pragma visibility early in the parsing of class definitions. Fixes
pr13338.

llvm-svn: 160105
2012-07-12 04:47:34 +00:00
Rafael Espindola 06e1b13209 Ignore visibility attributes after definitions. This matches newer (4.7) gcc's
behavior and is the first step in fixing pr13338.

llvm-svn: 160104
2012-07-12 04:32:30 +00:00
NAKAMURA Takumi e5494ffe4f Revert r160052, "Default to -std=c++11 on Windows.", for now.
Failing Tests (3):
    Clang :: Index/complete-cxx-inline-methods.cpp
    Clang :: Index/recursive-cxx-member-calls.cpp
    Clang :: SemaTemplate/inject-templated-friend-post.cpp

llvm-svn: 160103
2012-07-12 03:14:56 +00:00
John McCall 775086e67c Add the ObjFW runtime. Patch by Jonathan Schleifer!
llvm-svn: 160102
2012-07-12 02:07:58 +00:00
NAKAMURA Takumi 0e449c4e54 AST/CommentSema.cpp: Fix signess in abs() to appease msvc. It would not make sense to pass (unsigned)-(unsigned) to abs().
llvm-svn: 160097
2012-07-12 00:45:08 +00:00
NAKAMURA Takumi 7073024ffd clang/lib/Basic/CMakeLists.txt: Add missing dependency, ClangDiagnosticComment.
llvm-svn: 160096
2012-07-12 00:27:55 +00:00
Jordan Rose d66bee3f76 [analyzer] Don't inline virtual calls unless we can devirtualize properly.
Previously we were using the static type of the base object to inline
methods, whether virtual or non-virtual. Now, we try to see if the base
object has a known type, and if so ask for its implementation of the method.

llvm-svn: 160094
2012-07-12 00:16:25 +00:00
Tanya Lattner 7445ada9c8 Add OpenCL metadata for kernel arg names. This output is controlled via a flag as noted in the OpenCL Spec.
Includes a test case.

llvm-svn: 160092
2012-07-11 23:02:10 +00:00
Nico Weber 1325384b32 Don't try to do RVO on block variables that refer to an enclosing local.
Fixes PR13314, clang crashing on blocks refering to an enclosing local
when the enclosing function returns void.

llvm-svn: 160089
2012-07-11 22:50:15 +00:00
Richard Smith ded9c2ee92 Stop instantiating a class if we hit a static_assert failure. Also, if the
static_assert fails when parsing the template, don't diagnose it again on every
instantiation.

llvm-svn: 160088
2012-07-11 22:37:56 +00:00
Richard Smith 3d46052d90 Fix warning.
llvm-svn: 160087
2012-07-11 22:33:59 +00:00
Dmitri Gribenko f26054f0fb Enable comment parsing and semantic analysis to emit diagnostics. A few
diagnostics implemented -- see testcases.

I created a new TableGen file for comment diagnostics,
DiagnosticCommentKinds.td, because comment diagnostics don't logically
fit into AST diagnostics file.  But I don't feel strongly about it.

This also implements support for self-closing HTML tags in comment
lexer and parser (for example, <br />).

In order to issue precise diagnostics CommentSema needs to know the
declaration the comment is attached to.  There is no easy way to find a decl by 
comment, so we match comments and decls in lockstep: after parsing one
declgroup we check if we have any new, not yet attached comments.  If we do --
then we do the usual comment-finding process.

It is interesting that this automatically handles trailing comments.
We pick up not only comments that precede the declaration, but also
comments that *follow* the declaration -- thanks to the lookahead in
the lexer: after parsing the declgroup we've consumed the semicolon
and looked ahead through comments.

Added -Wdocumentation-html flag for semantic HTML errors to allow the user to 
disable only HTML warnings (but not HTML parse errors, which we emit as
warnings in -Wdocumentation).

llvm-svn: 160078
2012-07-11 21:38:39 +00:00
Argyrios Kyrtzidis 6d7833f1b6 Introduce a flag in SourceManager to treat non-system source files
as "volatile", meaning there's a high enough chance that they may
change while we are trying to use them.

This flag is only enabled by libclang.
Currently "volatile" source files will be stat'ed immediately
before opening them, because the file size stat info
may not be accurate since when we got it (e.g. from the PCH).
This avoids crashes when trying to reference mmap'ed memory
from a file whose size is not what we expect.

Note that there's still a window for a racing issue to occur
but the window for it should be way smaller than before.
We can consider later on to avoid mmap completely on such files.

rdar://11612916

llvm-svn: 160074
2012-07-11 20:59:04 +00:00
Jordan Rose b13eb8dca5 Allow -verify directives to be filtered by preprocessing.
This is accomplished by making VerifyDiagnosticsConsumer a CommentHandler,
which then only reads the -verify directives that are actually in live
blocks of code. It also makes it simpler to handle -verify directives that
appear in header files, though we still have to manually reparse some files
depending on how they are generated.

This requires some test changes. In particular, all PCH tests now have their
-verify directives outside the "header" portion of the file, using the @line
syntax added in r159978. Other tests have been modified mostly to make it
clear what is being tested, and to prevent polluting the expected output with
the directives themselves.

Patch by Andy Gibbs! (with slight modifications)

The new Frontend/verify-* tests exercise the functionality of this commit,
as well as r159978, r159979, and r160053 (Andy's other -verify enhancements).

llvm-svn: 160068
2012-07-11 19:58:23 +00:00
Daniel Jasper bb89ae9392 Add a hook to supply a custom CompilationDatabase. To add a custom CompilationDatabase, make it implement findCompilationDatabaseForDirectory in CustomCompilationDatabase.h and set USE_COSTUM_COMPILATION_DATABASE.
Differential Revision: http://llvm-reviews.chandlerc.com/D4

llvm-svn: 160061
2012-07-11 19:13:13 +00:00
Chad Rosier 4dce73af05 [driver] Fix so that clang can find correct path prefix for libc object files
from GNU binutils supporting multi-arch folder for ARM target.
Patch by Jiangning Liu <jiangning.liu@arm.com>.

llvm-svn: 160060
2012-07-11 19:08:21 +00:00
Rafael Espindola 1fa9fade38 Handle #pragma visibility in explicit specializations and enums.
llvm-svn: 160057
2012-07-11 18:14:09 +00:00
Jordan Rose 6f524ac9f0 Emit -verify diagnostics even when we have a fatal error.
Previously we'd halt at the fatal error as expected, but not actually emit
any -verify-related diagnostics. This lets us catch cases that emit a
/different/ fatal error from the one we expected.

This is implemented by adding a "force emit" mode to DiagnosticBuilder, which
will cause diagnostics to immediately be emitted regardless of current
suppression. Needless to say this should probably be used /very/ sparingly.

Patch by Andy Gibbs! Tests for all of Andy's -verify patches coming soon.

llvm-svn: 160053
2012-07-11 16:50:36 +00:00
Nico Weber cc19aec7ae Default to -std=c++11 on Windows.
llvm-svn: 160052
2012-07-11 16:46:17 +00:00
Justin Holewinski 5fafdd9d1d Fix handling of curly braces in NVPTX inline asm
Fixes bug 13322

Patch by Dmitry Mikushin

llvm-svn: 160050
2012-07-11 15:34:55 +00:00
Eric Christopher f8b9809fab Temporarily revert this to see if it brings the gdb bot back.
llvm-svn: 160049
2012-07-11 15:32:13 +00:00
Axel Naumann 23c1676dee LLVM_ON_WIN32 case: use the proper key in the UniqueFiles map.
llvm-svn: 160041
2012-07-11 09:41:34 +00:00
Rafael Espindola 5f0a1b89d7 Don't process #pragma visibility during instantiation. The visibility of the
instantiation depends on the template, its arguments and parameters, but not
where it is instantiated.

llvm-svn: 160034
2012-07-11 02:15:51 +00:00
Eric Christopher 2977378974 The end of a block doesn't necessarily need a line table entry unless
there's something going on there. Remove the unconditional line entry
and only add one if we're emitting cleanups (any other statements
would be handled normally).

Fixes rdar://9199234

llvm-svn: 160033
2012-07-11 01:49:26 +00:00
Jordan Rose 8889cf008d [analyzer] Add debug.DumpCalls, which prints out any CallEvents it sees.
This is probably not so useful yet because it is not path-sensitive, though
it does try to show inlining with indentation.

This also adds a dump() method to CallEvent, which should be useful for
debugging.

llvm-svn: 160030
2012-07-10 23:56:23 +00:00
Jordan Rose 6cd16c5152 [analyzer] Guard against C++ member functions that look like system functions.
C++ method calls and C function calls both appear as CallExprs in the AST.
This was causing crashes for an object that had a 'free' method.

<rdar://problem/11822244>

llvm-svn: 160029
2012-07-10 23:13:01 +00:00
Richard Smith 9fce7bc721 Fix crash when constant-evaluating a CXXConstructExpr representing
value-initialization for an array of class type with a trivial default
constructor.

llvm-svn: 160024
2012-07-10 22:12:55 +00:00
Jordan Rose eab627b951 [analyzer] Construct stack variables directly in their VarDecl.
Also contains a number of tweaks to inlining that are necessary
for constructors and destructors. (I have this enabled on a private
branch, but it is very much unstable.)

llvm-svn: 160023
2012-07-10 22:08:01 +00:00
Jordan Rose d1d54aa131 [analyzer] Use CallEvent for building inlined stack frames.
In order to accomplish this, we now build the callee's stack frame
as part of the CallEnter node, rather than the subsequent BlockEdge node.
This should not have any effect on perceived behavior or diagnostics.

This makes it safe to re-enable inlining of member overloaded operators.

llvm-svn: 160022
2012-07-10 22:07:57 +00:00
Jordan Rose fbe6dba15a [analyzer] Make CallEnter, CallExitBegin, and CallExitEnd not be StmtPoints
These ProgramPoints are used in inlining calls,
and not all calls have associated statements anymore.

llvm-svn: 160021
2012-07-10 22:07:52 +00:00
Jordan Rose 4ee71b8a18 [analyzer] Add a CXXDestructorCall CallEvent.
While this work is still fairly tentative (destructors are still left out of
the CFG by default), we now handle destructors in the same way as any other
calls, instead of just automatically trying to inline them.

llvm-svn: 160020
2012-07-10 22:07:47 +00:00
Jordan Rose 681cce9908 [analyzer] Add new PreImplicitCall and PostImplicitCall ProgramPoints.
These are currently unused, but are intended to be used in lieu of PreStmt
and PostStmt when the call is implicit (e.g. an automatic object destructor).

This also modifies the Data1 field of ProgramPoints to allow storing any
pointer-sized value, as opposed to only aligned pointers. This is necessary
to store SourceLocations.

There is currently no BugReporter support for these; they should be skipped
over in any diagnostic output.

This commit also tags checkers that currently rely on function calls only
occurring at StmtPoints.

llvm-svn: 160019
2012-07-10 22:07:42 +00:00
DeLesley Hutchins 868830f727 Thread safety analysis: impove handling of trylock expressions.
llvm-svn: 160018
2012-07-10 21:47:55 +00:00
Chad Rosier 67055f5ddc 80-column violations and whitespace.
llvm-svn: 160017
2012-07-10 21:35:27 +00:00
Daniel Jasper 1dad183b38 Add more matchers and do cleanups.
Reviewers: klimek

    Differential Revision: http://ec2-50-18-127-156.us-west-1.compute.amazonaws.com/D2

llvm-svn: 160013
2012-07-10 20:20:19 +00:00
Arnaud A. de Grandmaison 617f5269c3 Adds support for auto-detection of compilation databases, looking in a directory and all its parents.
llvm-svn: 159998
2012-07-10 16:56:35 +00:00
Axel Naumann b307400326 Improve r159256 following Chandler's comments:
Implement UniqueFileContainer::erase(), camelCase, add comment on future optimizations of the cache versus de-optimizations of invalidations.

llvm-svn: 159997
2012-07-10 16:50:27 +00:00
Anna Zaks 8698dd63d7 [analyzer] Remove redundant check (scalar type is a superset of integer)
PR13319 Reported by Jozsef Mihalicza.

llvm-svn: 159996
2012-07-10 16:27:55 +00:00
Axel Naumann b2f1a46402 Clear diagnostic counts after processing the diagnostics.
The consumer might see multiple input files (e.g. for cling) and since r159977 the count is maintained across input files.

llvm-svn: 159995
2012-07-10 16:24:07 +00:00
Manuel Klimek 65fd0e1fd3 Adds support for auto-detection of compilation databases
from a source file and changes clang-check to make use of this.

This makes clang-check just work on in-tree builds, and allows
easy setup via a symlink per source directory to make clang-check
work without any extra configuration.

llvm-svn: 159990
2012-07-10 13:10:51 +00:00
Jordan Rose b8b2ca6ffb Allow -verify directives to specify a min and max count, not just "+".
void f(); // expected-note 0+ {{previous declaration is here}}
  void g(); // expected-note 0-1 {{previous declaration is here}}

The old "+" syntax is still an alias for "1+", and single numbers still work.

Patch by Andy Gibbs!

llvm-svn: 159979
2012-07-10 02:57:26 +00:00
Jordan Rose e1572eb3e2 Allow line numbers on -verify directives.
// expected-warning@10 {{some text}}

The line number may be absolute (as above), or relative to the current
line by prefixing the number with either '+' or '-'.

Patch by Andy Gibbs!

llvm-svn: 159978
2012-07-10 02:57:03 +00:00
Jordan Rose 6dae761810 Clean up VerifyDiagnosticsConsumer in preparation for upcoming enhancements.
Patch by Andy Gibbs!

llvm-svn: 159977
2012-07-10 02:56:15 +00:00
Richard Trieu 50f5f46a93 Properly update the FormattedArgs vector when the template type diffing falls
back to regular type printing.

llvm-svn: 159976
2012-07-10 01:46:04 +00:00
Tanya Lattner bcffcdfd18 Patch by Anton Lokhmotov to add OpenCL work group size attributes.
llvm-svn: 159965
2012-07-09 22:06:01 +00:00
Dmitri Gribenko 17709ae8d9 Comment lexing: fix lexing to actually work in non-error cases.
llvm-svn: 159963
2012-07-09 21:32:40 +00:00
Abramo Bagnara cb43567374 The delete argument should not be converted to void*.
llvm-svn: 159961
2012-07-09 21:15:43 +00:00
Fariborz Jahanian 05d0d447c4 objective-c: provide fixit hint for @autoreleasepool
and similar other keywords. // rdar://10723084

llvm-svn: 159956
2012-07-09 20:00:35 +00:00
Matt Beaumont-Gay ad0bb8e361 Silence unused variable warning in -Asserts build
llvm-svn: 159954
2012-07-09 18:55:31 +00:00
Chad Rosier 6213549751 80-column and whitespace.
llvm-svn: 159942
2012-07-09 17:31:28 +00:00
Jordan Rose 12e730c6e3 Better parser recovery in Objective-C containers.
Previously it was possible to get an infinite-loop-on-invalid with a namespace
decl within @interface. Since 'namespace' is normally a safe place to retry
top-level parsing, we just didn't consume the token.

This adds a flag that tracks whether we have temporarily left Objective-C
scope to parse a C-like declaration, and uses that to better recover from
parse problems by stopping at possible method declarations and at @end. To
fix the original problem, we do /not/ stop at 'namespace' when in an
Objective-C @interface or @protocol context (but still do in @implementation).

llvm-svn: 159941
2012-07-09 16:54:53 +00:00
Jordan Rose e04395e2f4 [analyzer] When inlining, make sure we use the definition decl.
This was a regression introduced during the CallEvent changes; a call to
FunctionDecl::hasBody was also being used to replace the decl found by
lookup with the actual definition. To keep from making this mistake again
(particularly if/when we start inlining Objective-C methods), this commit
adds a "getDefinition()" method to CallEvent, which should do the right
thing under any circumstances.

llvm-svn: 159940
2012-07-09 16:54:49 +00:00
Jordan Rose f919803001 Downgrade the "direct comparison" error for ObjC literals to a warning.
Chris pointed out that while the comparison is certainly problematic
and does not have well-defined behavior, it isn't any worse than some
of the other abuses that we merely warn about and doesn't need to make
the compilation fail.

Revert the release notes change (r159766) now that this is just a new warning.

llvm-svn: 159939
2012-07-09 16:54:44 +00:00
NAKAMURA Takumi d7ba69de90 [CMake] Get rid of unconditional dependency to ClangDiagnosticCommon. Only clangBasic and clangASTMatchers need it.
llvm-svn: 159931
2012-07-09 14:12:20 +00:00
Alexander Kornienko 20f6fc6220 Inline storage of attributes in AttributedStmt.
llvm-svn: 159925
2012-07-09 10:04:07 +00:00
Richard Smith 34349003cf PR13136:
* When substituting a reference to a non-type template parameter pack where the
   corresponding argument is a pack expansion, transform into an expression
   which contains an unexpanded parameter pack rather than into an expression
   which contains a pack expansion. This causes the SubstNonTypeTemplateParmExpr
   to be inside the PackExpansionExpr, rather than outside, so the expression
   still looks like a pack expansion and can be deduced.

 * Teach MarkUsedTemplateParameters that we can deduce a reference to a template
   parameter if it's wrapped in a SubstNonTypeTemplateParmExpr (such nodes are
   added during alias template substitution).

llvm-svn: 159922
2012-07-09 03:07:20 +00:00
Simon Atanasyan ecedf3ddec MIPS: Range check __builtin_mips_wrdsp / __builtin_mips_rddsp arguments against the upper/lower values.
llvm-svn: 159911
2012-07-08 09:30:00 +00:00
Richard Smith 7ebb07c87c PR13243: When deducing a non-type template parameter which is specified as an
expression, skip over any SubstNonTypeTemplateParmExprs which alias templates
may have inserted before checking for a DeclRefExpr referring to a non-type
template parameter declaration.

llvm-svn: 159909
2012-07-08 04:37:51 +00:00
Richard Smith 74801c8182 PR13293: Defer deduction of an auto type with a dependent declarator, such as "auto (*f)(T t)".
llvm-svn: 159908
2012-07-08 04:13:07 +00:00
Richard Smith 8a874c93d9 PR9793: Treat substitution as an instantiation step for the purpose of the
-ftemplate-depth limit.  There are various ways to get an infinite (or merely
huge) stack of substitutions with no intervening instantiations. This is also
consistent with gcc's behavior.

llvm-svn: 159907
2012-07-08 02:38:24 +00:00
NAKAMURA Takumi 17a2ff2494 ASTMatchers/CMakeLists.txt: Add dependencies to generated headers, or "make clean; make ASTMatchers" would fail.
llvm-svn: 159906
2012-07-07 23:13:30 +00:00
Richard Smith cd3d5d9597 Reject 'int a[1][];' in Sema rather than crashing in IR generation. Found by a
misreduction of PR13290.

llvm-svn: 159905
2012-07-07 23:00:31 +00:00
Richard Smith 1b9f2eb76b PR13290: Constant-evaluation support for CXXConstructExprs which construct a
multidimensional array of class type. Also, preserve zero-initialization when
evaluating an initializer list for an array, in case the initializers refer to
later elements (which have preceding zero-initialization).

llvm-svn: 159904
2012-07-07 22:48:24 +00:00
Benjamin Kramer 1e250395fa Wire up -mrdrnd for X86.
For some reason GCC decided to call the feature rdrnd instead of rdrand,
which requires translating it for LLVM.

llvm-svn: 159897
2012-07-07 09:39:18 +00:00
Richard Smith e20c83d9ed PR12670: Support for initializing an array of non-aggregate class type from an
initializer list. Patch by Olivier Goffart, with extra testcases by Meador Inge
and Daniel Lunow.

llvm-svn: 159896
2012-07-07 08:35:56 +00:00
Richard Smith 4ff9ff974c When marking virtual functions as used for a class' vtable, mark all functions
which will appear in the vtable as used, not just those ones which were
declared within the class itself. Fixes an issue reported as comment#3 in
PR12763 -- we sometimes assert in codegen if we try to emit a reference to a
function declaration which we've not marked as referenced. This also matches
gcc's observed behavior.

llvm-svn: 159895
2012-07-07 06:59:51 +00:00
John McCall 8dda7b27ee Distinguish more carefully between free functions and C++ instance methods
in the ABI arrangement, and leave a hook behind so that we can easily
tweak CCs on platforms that use different CCs by default for C++
instance methods.

llvm-svn: 159894
2012-07-07 06:41:13 +00:00
Ted Kremenek b22ea2acea Re-apply r159875 with fixes.
- Split pedantic driver flag test into separate test file, and XFAIL on cygwin,mingw32
- Fix bug in tablegen logic where a missing '{' caused errors to be included in -Wpedantic.

llvm-svn: 159892
2012-07-07 05:53:30 +00:00
NAKAMURA Takumi eec53e4069 Revert rr159875, "Implement -Wpedantic and --no-pedantic to complement -Weverything." It broke several builds.
I suspect FileCheck might match assertion failure, even if clang/test/Misc/warning-flags.c passed the test.

> 0. Program arguments: bin/./clang -### -pedantic -Wpedantic clang/test/Driver/warning-options.cpp

llvm-svn: 159886
2012-07-07 02:48:02 +00:00
Dmitri Gribenko 44cd7e6746 Restrict the set of declaration kinds for which we allow trailing comments.
llvm-svn: 159878
2012-07-06 23:27:33 +00:00
Ted Kremenek 9877f689f2 Implement -Wpedantic and --no-pedantic to complement -Weverything.
This patch introduces some magic in tablegen to create a "Pedantic" diagnostic
group which automagically includes all warnings that are extensions.  This
allows a user to suppress specific warnings traditionally under -pedantic used
an ordinary warning flag.  This also allows users to use #pragma to silence
specific -pedantic warnings, or promote them to errors, within blocks of text
(just like any other warning).

-Wpedantic is NOT an alias for -pedantic.  Instead, it provides another way
to (a) activate -pedantic warnings and (b) disable them.  Where they differ
is that -pedantic changes the behavior of the preprocessor slightly, whereas
-Wpedantic does not (it just turns on the warnings).

The magic in the tablegen diagnostic emitter has to do with computing the minimal
set of diagnostic groups and diagnostics that should go into -Wpedantic, as those
diagnostics that already members of groups that themselves are (transitively) members
of -Wpedantic do not need to be included in the Pedantic group directly.  I went
back and forth on whether or not to magically generate this group, and the invariant
was that we always wanted extension warnings to be included in -Wpedantic "some how",
but the bookkeeping would be very onerous to manage by hand.

-no-pedantic (and --no-pedantic) is included for completeness, and matches many of the
same kind of flags the compiler already supports.  It does what it says: cancels out
-pedantic.  One discrepancy is that if one specifies --no-pedantic and -Weverything or
-Wpedantic the pedantic warnings are still enabled (essentially the -W flags win).  We
can debate the correct behavior here.

Along the way, this patch nukes some code in TextDiagnosticPrinter.cpp and CXStoredDiagnostic.cpp
that determine whether to include the "-pedantic" flag in the warning output.  This is
no longer needed, as all extensions now have a -W flag.

This patch also significantly reduces the number of warnings not under flags from 229
to 158 (all extension warnings).  That's a 31% reduction.

llvm-svn: 159875
2012-07-06 23:07:31 +00:00
Jordan Rose 3f89e0ec93 [analyzer] Be careful about LazyCompoundVals, which may be for the first field.
We use LazyCompoundVals to avoid copying the contents of structs and arrays
around in the store, and when we need to pass a struct around that already
has a LazyCompoundVal we just use the original one. However, it's possible
that the first field of a struct may have a LazyCompoundVal of its own, and
we currently can't distinguish a LazyCompoundVal for the first element of a
struct from a LazyCompoundVal for the entire struct. In this case we should
just drop the optimization and make a new LazyCompoundVal that encompasses
the old one.

PR13264 / <rdar://problem/11802440>

llvm-svn: 159866
2012-07-06 21:59:56 +00:00
Matt Beaumont-Gay 4106ea3b4e Sprinkle llvm_unreachable around to placate GCC's -Wreturn-type.
llvm-svn: 159860
2012-07-06 21:13:09 +00:00
Fariborz Jahanian dabd133186 objc-arc: warn when assigning retained object to
a 'weak' property just as we do the same for
'weak' variables. // rdar://11814185

llvm-svn: 159859
2012-07-06 21:09:27 +00:00
Fariborz Jahanian 0021347926 objective-c++11: extend c++11 range-based loop to iterate
over objective-c container collection. // rdar://9293227

llvm-svn: 159847
2012-07-06 19:04:04 +00:00
Dmitri Gribenko 7dd29d4d3d Don't store pointers into a std::vector (RawCommentList::Comments). Although
currently we take address of std::vector's contents only after we finished
adding all comments (so no reallocation can happen), this will change in
future.

llvm-svn: 159845
2012-07-06 18:19:34 +00:00
Dmitri Gribenko bacb9f65a7 Stop using new[] on llvm::BumpPtrAllocator.
llvm-svn: 159833
2012-07-06 16:41:59 +00:00
Dmitri Gribenko 08630059c5 Use C++ cast.
llvm-svn: 159830
2012-07-06 15:40:08 +00:00
Benjamin Kramer b2c42807a4 Remove unreachable default case to pacify clang's -Wcovered-switch-default.
llvm-svn: 159829
2012-07-06 15:27:25 +00:00
NAKAMURA Takumi 388817236f ASTContext.cpp: Appease msvc to cast NULL to typed pointer, or msvc mistakes NULL as int.
llvm-svn: 159818
2012-07-06 11:51:12 +00:00
Manuel Klimek cb93f785de Build-fix: Remove non-existent directories from Makefiles.
llvm-svn: 159807
2012-07-06 06:00:30 +00:00
Manuel Klimek 04616e4776 Adds the AST Matcher library, which provides a in-C++ DSL to express
matches on interesting parts of the AST, and callback mechanisms to
act on them.

llvm-svn: 159805
2012-07-06 05:48:52 +00:00
Dmitri Gribenko 017b677737 CommentParser.cpp: remove unused variable.
llvm-svn: 159797
2012-07-06 01:14:58 +00:00
Lang Hames aa53b936ec Add -ffp-contract = { fast | on | off } command line option support.
This flag sets the 'fp-contract' mode, which controls the formation of fused
floating point operations. Available modes are:

- Fast: Form fused operations anywhere. 
- On: Form fused operations where allowed by FP_CONTRACT. This is the default
      mode.
- Off: Don't form fused operations (in future this may be relaxed to forming
       fused operations where it can be proved that the result won't be
       affected).

Currently clang doesn't support the FP_CONTRACT pragma, so the 'On' and 'Off'
modes are equivalent.

llvm-svn: 159794
2012-07-06 00:59:19 +00:00
Fariborz Jahanian 97920394c0 Added a new memberfor Parser, to be used soon
for doing delayed parsing of c++ method defined in
objc class implementations.

llvm-svn: 159792
2012-07-06 00:42:20 +00:00
Dmitri Gribenko ec92531c29 Implement AST classes for comments, a real parser for Doxygen comments and a
very simple semantic analysis that just builds the AST; minor changes for lexer
to pick up source locations I didn't think about before.

Comments AST is modelled along the ideas of HTML AST: block and inline content.

* Block content is a paragraph or a command that has a paragraph as an argument
  or verbatim command.
* Inline content is placed within some block.  Inline content includes plain
  text, inline commands and HTML as tag soup.

llvm-svn: 159790
2012-07-06 00:28:32 +00:00
Argyrios Kyrtzidis 89b928eb6a [objcmt] Check for classes that accept 'objectForKey:' (or the other selectors
that the migrator handles) but return their instances as 'id', resulting
in the compiler resolving 'objectForKey:' as the method from NSDictionary.

When checking if we can convert to subscripting syntax, check whether
the receiver is a result of a class method from a hardcoded list of
such classes. In such a case return the specific class as the interface
of the receiver.

llvm-svn: 159788
2012-07-06 00:07:09 +00:00
Fariborz Jahanian 31a2568096 objective-c: don't involve properties when checking
under -Wsuper-class-method-mismatch for method
mismatches in current and suprt class.
// rdar://11793793

llvm-svn: 159784
2012-07-05 22:26:07 +00:00
Argyrios Kyrtzidis 13b92929dc [objcmt] Allow migrating to subscripting syntax for other classes
(apart from NSDictionary/NSArray) that implement objectForKey:/objectAtIndex/etc.
and the subscripting methods as well.

Part of rdar://11734969

llvm-svn: 159783
2012-07-05 21:49:51 +00:00
DeLesley Hutchins 09bcefcb6d Thread-safety analysis: eliminate false positives in case where the definition
duplicates attributes on the declaration.  Also eliminates a false negative in
ReleasableMutexLock.  Fixing this bug required some refactoring.

llvm-svn: 159780
2012-07-05 21:16:29 +00:00
Anna Zaks 40b87fc1f9 [analyzer] Add a statistic for maximum CFG size and a script to summarize analyzer stats from scan-build output.
llvm-svn: 159776
2012-07-05 20:44:02 +00:00
Simon Atanasyan 9780e4a2bb MIPS: Define __mips_dsp_rev / __mips_dspr2 / __mips_dsp macros
if -mdsp or -mdspr2 options are provided.

llvm-svn: 159774
2012-07-05 20:16:22 +00:00
Fariborz Jahanian 8de7955f35 objective-c++ parsing. Turn off delayed parsing
of out-of-line c++ method definition which happens
to be inside an objc class implementation
until I can figure out how to do it. This is to fix 
a broken project.

llvm-svn: 159772
2012-07-05 19:34:20 +00:00
Simon Atanasyan 9c6f1f7f23 MIPS: Add -mdsp/-mno-dsp and -mdspr2/-mno-dspr2 command line options support.
llvm-svn: 159769
2012-07-05 19:23:00 +00:00
Simon Atanasyan 9b1932d4e4 MIPS: Factor out the code converting command line options to target features.
llvm-svn: 159767
2012-07-05 18:51:43 +00:00
Manuel Klimek 5da9dcb275 Adapts the FrontendAction convenience functions so that it can be
used with classes that generate ASTConsumers; this allows decoupling
the ASTConsumer generation from the Frontend library (like, for example,
the MatchFinder in the upcoming ASTMatcher patch).

llvm-svn: 159760
2012-07-05 18:13:01 +00:00
Simon Atanasyan 72244b6e4f MIPS: Define __mips16 macro if -mips16 option is provided.
llvm-svn: 159753
2012-07-05 16:06:06 +00:00
Simon Atanasyan 9f444d5a1b MIPS: Replace the pair of boolean flags by enumeration to hold selected float ABI.
llvm-svn: 159752
2012-07-05 15:32:46 +00:00
Simon Atanasyan 6f23fa0f18 MIPS: Add -mips16 / -mno-mips16 command line support.
llvm-svn: 159747
2012-07-05 14:19:39 +00:00
Richard Smith d86812d95c PR13273: When performing list-initialization with an empty initializer list,
actually perform value initialization rather than trying to fake it with a call
to the default constructor. Fixes various bugs related to the previously-missing
zero-initialization in this case.

I've also moved this and the other list initialization 'special case' from
TryConstructorInitialization into TryListInitialization where they belong.

llvm-svn: 159733
2012-07-05 08:39:21 +00:00
Benjamin Kramer 4ab984e74b Drop the ASTContext.h include from DeclFriend.h and DeclTemplate.h.
llvm-svn: 159723
2012-07-04 20:19:54 +00:00
Benjamin Kramer 2ef3031496 Remove get(V)BaseClassOffsetInBits, the CharUnit functions should be used instead.
No functionality change.

llvm-svn: 159719
2012-07-04 18:45:14 +00:00
Benjamin Kramer 1ea8e092be Drop the ASTContext.h include from Stmt.h and fix up transitive users.
This required moving the ctors for IntegerLiteral and FloatingLiteral out of
line which shouldn't change anything as they are usually called through Create
methods that are already out of line.

ASTContext::Deallocate has been a nop for a long time, drop it from ASTVector
and make it independent from ASTContext.h

Pass the StorageAllocator directly to AccessedEntity so it doesn't need to
have a definition of ASTContext around.

llvm-svn: 159718
2012-07-04 17:04:04 +00:00
Benjamin Kramer e2a929df73 Split out the "empty" case for compound statement into a separate ctor.
Move the ASTContext-dependent version out of line.

llvm-svn: 159717
2012-07-04 17:03:41 +00:00
Benjamin Kramer 300c063db3 CXXRecordDecl: Split getBases/getVBases into a slow and a fast path.
This avoids costly computation of getASTContext() and drops the header
dependency from DeclCXX.h to ASTContext.h.

llvm-svn: 159716
2012-07-04 17:03:33 +00:00
David Chisnall da22535573 Hoist the logic for selecting the Objective-C dispatch method into the runtime
class, from the target.  No functionality change, just less duplicated logic.

llvm-svn: 159710
2012-07-04 11:52:24 +00:00
David Chisnall 314896c9fc A few more cleanups for the GNU family of ObjC runtimes.
llvm-svn: 159708
2012-07-04 10:37:03 +00:00
Abramo Bagnara e06a8887d8 Renamed RawComment kinds to avoid name clash.
llvm-svn: 159706
2012-07-04 07:30:26 +00:00
Bob Wilson ef285b4c22 Move a comment from the commit message into the code.
llvm-svn: 159696
2012-07-04 00:18:41 +00:00
Fariborz Jahanian 8be1ecd615 Obj-C++11 parser: handle a fall out of delayed
c-function parsing when a declaration with
C++0x braced-init-list is inside an @implementation.

llvm-svn: 159693
2012-07-03 23:22:13 +00:00
Jordan Rose 017591ab45 [analyzer] For now, don't inline non-static member overloaded operators.
Our current inlining support (specifically RegionStore::enterStackFrame)
doesn't know that calls to overloaded operators may be calls to non-static
member functions, and that in these cases the first argument should be
treated as 'this'. This caused incorrect results and sometimes crashes.

The long-term fix will be to rewrite RegionStore::enterStackFrame to use
CallEvent and its subclasses, but for now we can just disable these
problematic calls by classifying them under a new CallEvent,
CXXMemberOperatorCall.

llvm-svn: 159692
2012-07-03 22:55:57 +00:00
Fariborz Jahanian 8a369a82d5 Obj-C++11 parser: fix broken parsing of c-function
defined in class implementations.

llvm-svn: 159691
2012-07-03 22:54:28 +00:00
Fariborz Jahanian f30f6ac7b9 Obj-C++11 parser: turn off buffering of
c-function defined in objc class
implementation for now.

llvm-svn: 159690
2012-07-03 22:29:23 +00:00
Fariborz Jahanian 450bb6e8ff objective-c: Refactor parse/sema portion of
objective-c's fast enumeration statement,
for more work to come.

llvm-svn: 159689
2012-07-03 22:00:52 +00:00
David Chisnall b601c96892 Rename the GCC Objective-C runtime to gcc from gnu-fragile and the GNUstep
runtime to gnustep from gnu.  Fix EH for the GCC runtime.

llvm-svn: 159684
2012-07-03 20:49:52 +00:00
Bob Wilson 1e148fe064 When using -pg targeting OS X 10.8, pass -no_new_main to the linker.
By default on OS X 10.8, we don't link with a crt1.o file and the linker
knows to use _main as the entry point.  But, when compiling with -pg, we
need to link with the gcrt1.o file, and the linker needs to be told to use
the "start" symbol as the entry point.  The -no_new_main linker option does
that last part.  <rdar://problem/11491405>

llvm-svn: 159683
2012-07-03 20:42:10 +00:00
DeLesley Hutchins 3a8d6cff14 Thread safety analysis: improve handling of smart pointers.
llvm-svn: 159679
2012-07-03 19:47:18 +00:00
Akira Hatanaka e1e3ad3d11 Make the following changes in the way Mips handles vector arguments and return
values:

- Return integer vectors in integer registers.
- Pass vector arguments in integer registers.
- Set an upper bound for argument alignment. The largest alignment is 8-byte
  for O32 and 16-byte for N32/64.

llvm-svn: 159676
2012-07-03 19:24:06 +00:00
DeLesley Hutchins 0c1da20bf8 Thread Safety Analysis: handle expressions involving temporaries,
e.g. ExprWithCleanups.

llvm-svn: 159674
2012-07-03 18:25:56 +00:00
Dmitri Gribenko 767ea0ff4a CommentBriefParser: remove dead store. Found by Clang Analyzer.
llvm-svn: 159673
2012-07-03 18:10:20 +00:00
Hal Finkel 6b984f084c Add additional architecture defines for PPC targets.
Patch by Andy Gibbs.

llvm-svn: 159665
2012-07-03 16:51:04 +00:00
Argyrios Kyrtzidis 4cf2ffe28b [libclang] Protect against a race condition where a thread
may be destroying an ASTUnit while cleanupOnDiskMapAtExit is
getting called.

rdar://11781241

llvm-svn: 159664
2012-07-03 16:30:52 +00:00
Matt Beaumont-Gay 36af16affd Silence warning in -Asserts build
llvm-svn: 159635
2012-07-03 03:55:58 +00:00
Nico Weber 4b18c3ff40 Share ConvertUTF8toWide() between Lex and CodeGen.
llvm-svn: 159634
2012-07-03 02:24:52 +00:00
Nico Weber b8124d1af1 Rename -Wself-assign-memvar to -Wself-assign-field to improve local consistency a bit.
(cf -Wunused-private-field and several other existing -field diagnostics.)

llvm-svn: 159633
2012-07-03 02:03:06 +00:00
Chandler Carruth c4702dac1c Remove a redundant assignment to the FDecl variable from *inside* it's
initializer.

I really feel like Clang should warn about this, but I can't describe
a good reason. GCC will warn on this in some cases under
-Wsequence-point, but it actually seems like a false positive for that
warning....

llvm-svn: 159631
2012-07-03 00:15:11 +00:00
John McCall 4e8ca4fa14 Significantly simplify CGExprAgg's logic about ignored results:
if we want to ignore a result, the Dest will be null.  Otherwise,
we must copy into it.  This means we need to ensure a slot when
loading from a volatile l-value.

With all that in place, fix a bug with chained assignments into
__block variables of aggregate type where we were losing insight into
the actual source of the value during the second assignment.

llvm-svn: 159630
2012-07-02 23:58:38 +00:00
James Dennett 0bb4164ce8 Documentation cleanup: Delete a duplicated/malformed doc comment.
llvm-svn: 159628
2012-07-02 23:56:17 +00:00
Fariborz Jahanian 577574ab88 objective-c: just as we have done for method definitions,
c-functions declared in implementation should have their 
parsing delayed until the end so, they can access forward
declared private methods. // rdar://10387088

llvm-svn: 159626
2012-07-02 23:37:09 +00:00
Richard Smith b721e301df -Wuninitialized: assume that an __attribute__((returns_twice)) function might
initialize any variable. This is extremely conservative, but is sufficient for
now.

llvm-svn: 159620
2012-07-02 23:23:04 +00:00
DeLesley Hutchins ab0d4e6cd8 Thread safety analysis: fixed bug that occurs when very silly people
use scoped_lockable without putting unlock_function on the
destructor.

llvm-svn: 159609
2012-07-02 22:26:29 +00:00
Jordan Rose a4ee064cf3 [analyzer] Introduce CXXAllocatorCall to handle placement arg invalidation.
This is NOT full-blown support for operator new, but removes some nasty
duplicated code introduced in r158784.

llvm-svn: 159608
2012-07-02 22:21:47 +00:00
DeLesley Hutchins 6e6dbb7618 Thread safety analysis: fixed incorrect error message at the end of a locks_required function.
llvm-svn: 159607
2012-07-02 22:16:54 +00:00
DeLesley Hutchins 2a15baf968 Thread safety analysis: don't warn in case of duplicate annotation.
llvm-svn: 159606
2012-07-02 22:12:12 +00:00
DeLesley Hutchins c4a6e51596 Thread Safety Analysis: turn off checking within trylock functions.
llvm-svn: 159601
2012-07-02 21:59:24 +00:00
Jordan Rose e8d5567426 [analyzer] If 'super' is known to be nil, we can still mark its range.
llvm-svn: 159596
2012-07-02 21:41:56 +00:00
Jordan Rose 62146f67b7 Revert "Remove unused member (& consequently unused parameter) in SA's Call code."
...and instead add an accessor. We're not using this today, but it's something
that should probably stay in the source for potential clients, and it doesn't
cost a lot. (ObjCPropertyAccess is only created on the stack, and right now
there's only ever one alive at a time.)

This reverts r159581 / commit 8e674e1da34a131faa7d43dc3fcbd6e49120edbe.

llvm-svn: 159595
2012-07-02 21:41:53 +00:00
David Blaikie df9582cca5 Include -D files in crash report repro scripts. (PR13255)
Now that we're only using -frewrite-includes rather than full preprocessing
when producing repro source files, we should also include command line macro
definitions in the repro script.

I don't have a test case for this because I'm not sure if/how I can open the
crash report file when the name is only known by scraping the crash report
output. Suggestions welcome if anyone thinks it'd be helpful.

llvm-svn: 159592
2012-07-02 21:28:00 +00:00
Jordan Rose d39e5f1463 In blocks, only pretend that enum constants have enum type if necessary.
In C, enum constants have the type of the enum's underlying integer type,
rather than the type of the enum. (This is not true in C++.) Thus, when a
block's return type is inferred from an enum constant, it is incompatible
with expressions that return the enum type.

In r158899, I told block returns to pretend that enum constants have enum
type, like in C++. Doug Gregor pointed out that this can break existing code.

Now, we don't check the types of return statements until the end of the block.
This lets us go back and add implicit casts in blocks with mixed enum
constants and enum-typed expressions.

<rdar://problem/11662489> (again)

llvm-svn: 159591
2012-07-02 21:19:23 +00:00
Eli Friedman c9f439461b When we're looking for redeclarations which might provide a definition in CodeGen, make sure we examine all the redeclarations. PR13252.
llvm-svn: 159586
2012-07-02 21:05:30 +00:00
Douglas Gregor b9b8b81ef2 Be more eager about setting the 'Invalid' bit on an invalid class
template instantiation. I wasn't able to reproduce this down to
anything small enough to put in our test suite, but it's "obviously"
okay to set the invalid bit earlier and precludes a
known-broken-but-not-marked-broken class from being used elsewhere.

llvm-svn: 159584
2012-07-02 21:00:41 +00:00
David Blaikie 977a35644e Remove unused member (& consequently unused parameter) in SA's Call code.
This member became unused in r159559.

llvm-svn: 159581
2012-07-02 20:44:57 +00:00
Abramo Bagnara b1cdde7d21 Reintroduced FieldDecl public methods setBitWidth and removeBitWidth.
llvm-svn: 159579
2012-07-02 20:35:48 +00:00
Ted Kremenek de21a1c964 Bail out the LiveVariables analysis when the CFG is very large, as
we are encountering some scalability issues with memory usage.   The
appropriate long term fix is to make the analysis more scalable, but
this will at least prevent the analyzer swapping when
analyzing very large functions.

llvm-svn: 159578
2012-07-02 20:21:52 +00:00
Ted Kremenek 146ef384da Fix subtle bug in AnalysisConsumer where we would not analyze functions whose parent
in the call graph had been inlined but for whatever reason we did not inline some
of its callees.

Also, fix a related traversal bug where we meant to do a BFS of the callgraph but
instead were doing a DFS.

llvm-svn: 159577
2012-07-02 20:21:48 +00:00
Douglas Gregor 271205cedb Reduce default template instantiation depth to 512; we're blowing out
the stack too often with 1024. Fixes <rdar://problem/11678534>.

llvm-svn: 159573
2012-07-02 19:56:23 +00:00
Jordan Rose 682b31621f [analyzer] Convert existing checkers to use check::preCall and check::postCall.
llvm-svn: 159563
2012-07-02 19:28:21 +00:00
Jordan Rose afe7c2c2bf [analyzer] Add generic preCall and postCall checks.
llvm-svn: 159562
2012-07-02 19:28:16 +00:00
Jordan Rose f3c12ac3b1 [analyzer] Convert CXXConstructExpr over to use CallEvent for evaluation.
llvm-svn: 159561
2012-07-02 19:28:12 +00:00
Jordan Rose cbeef55886 [analyzer] Use CallEvent for inlining and call default-evaluation.
llvm-svn: 159560
2012-07-02 19:28:09 +00:00
Jordan Rose 547060b30b [analyzer] Finish replacing ObjCMessage with ObjCMethodDecl and friends.
The preObjCMessage and postObjCMessage callbacks now take an ObjCMethodCall
argument, which can represent an explicit message send (ObjCMessageSend) or an
implicit message generated by a property access (ObjCPropertyAccess).

llvm-svn: 159559
2012-07-02 19:28:04 +00:00
Jordan Rose 6bad4905d7 [analyzer] Begin replacing ObjCMessage with ObjCMethodCall and friends.
Previously, the CallEvent subclass ObjCMessageInvocation was just a wrapper
around the existing ObjCMessage abstraction (over message sends and property
accesses). Now, we have abstract CallEvent ObjCMethodCall with subclasses
ObjCMessageSend and ObjCPropertyAccess.

In addition to removing yet another wrapper object, this should make it easy
to add a ObjCSubscriptAccess call event soon.

llvm-svn: 159558
2012-07-02 19:27:56 +00:00
Jordan Rose 7ab0182e33 [analyzer] Move the last bits of CallOrObjCMessage over to CallEvent.
This involved refactoring some common pointer-escapes code onto CallEvent,
then having MallocChecker use those callbacks for whether or not to consider
a pointer's /ownership/ as escaping. This still needs to be pinned down, and
probably we want to make the new argumentsMayEscape() function a little more
discerning (content invalidation vs. ownership/metadata invalidation), but
this is a good improvement.

As a bonus, also remove CallOrObjCMessage from the source completely.

llvm-svn: 159557
2012-07-02 19:27:51 +00:00
Jordan Rose 2995349f3e [analyzer] Convert CallAndMessageChecker and ObjCSelfInitChecker to CallEvent.
Both of these got uglier rather than cleaner because we don't have preCall and
postCall yet; properly wrapping a CallExpr in a CallEvent requires doing a bit
of deconstruction on the callee. Even when we have preCall and postCall we may
want to expose the current CallEvent to pre/postStmt<CallExpr>.

llvm-svn: 159556
2012-07-02 19:27:46 +00:00
Jordan Rose eec1539a31 [analyzer] Convert RetainCountChecker to use CallEvent as much as possible.
This ended allowing quite a bit of cleanup, and some minor changes.

- CallEvent makes it easy to use hasNonZeroCallbackArg more aggressively, which
  we check in order to avoid false positives with callbacks that might release
  the object.
- In order to support this for functions which consume their arguments, there
  are two new ArgEffects: DecRefAndStopTracking and DecRefMsgAndStopTracking.
  These act just like StopTracking, except that if the object only had a
  return count of +1 it's now considered released instead (so we still get
  use-after-free messages).
- On the plus side, we no longer have to special-case
  +[NSObject performSelector:withObject:afterDelay:] and friends.
- The use of IdentifierInfos in the method summary cache is now hidden; only
  the ObjCInterfaceDecl gets passed around most of the time.
- Since we cache all "simple" summaries and check every function call, there is
  no real benefit to having NULL stand in for default summaries anymore.
- Whitespace, unused methods, etc.

Even more simplification to come when we get check::postCall and can unify all
these other post* checks.

llvm-svn: 159555
2012-07-02 19:27:43 +00:00
Jordan Rose 742920c8e7 [analyzer] Add a new abstraction over all types of calls: CallEvent
This is intended to replace CallOrObjCMessage, and is eventually intended to be
used for anything that cares more about /what/ is being called than /how/ it's
being called. For example, inlining destructors should be the same as inlining
blocks, and checking __attribute__((nonnull)) should apply to the allocator
calls generated by operator new.

llvm-svn: 159554
2012-07-02 19:27:35 +00:00
Argyrios Kyrtzidis 442dd80715 [PCH] Make sure that all newly introduced visible decls in a DeclContext
coming from an AST file are registered for serialization.

A static data member instantiation of in a chained PCH could be missed
when serializing decls; the result was that when emitting the visible decls
map of its DeclContext, we would use a DeclID that was not actually emitted,
leading to crashes or hangs.

Fix this by making sure such decls are always registered for serialization.
Also introduce extra sanity checks to make sure we don't register new
declarations or types after we have serialized the types/decls block.

rdar://11728990

llvm-svn: 159550
2012-07-02 19:19:01 +00:00
Richard Smith 200f47c65d A ':' after an enum-specifier at class scope is a bitfield, not a typo for a ';'.
llvm-svn: 159549
2012-07-02 19:14:01 +00:00
Dmitri Gribenko 3292d06a1b Add a new libclang completion API to get brief documentation comment that is
attached to a declaration in the completion string.

Since extracting comments isn't free, a new code completion option is
introduced.

A new code completion option that enables including brief comments
into CodeCompletionString should be a, err, code completion option.
But because ASTUnit caches global declarations during parsing before
even completion consumer is created, the option is duplicated as a
translation unit option (in both libclang and ASTUnit, like the option
to cache code completion results).

llvm-svn: 159539
2012-07-02 17:35:10 +00:00
Craig Topper 6490bdcf72 Rename tzcnt intrinsics to match gcc.
llvm-svn: 159515
2012-07-02 06:52:51 +00:00
Meador Inge cfb6090912 PR13189: va_list broken with precompiled headers
For some targets a structure named __va_list_tag is built to help define
the __builtin_va_list type.  However, __va_list_tag was not being treated as a
predefined type thus causing problems when serializing the AST.  This commit
fixes that oversight by adding the necessary support to treat __va_list_tag
as a predefined type.

llvm-svn: 159508
2012-07-01 15:57:25 +00:00
Jordan Rose 58d547200e Add support for the C11 _Alignof keyword.
This behaves like the existing GNU __alignof and C++11 alignof keywords;
most of the patch is simply adding the third token spelling to various places.

llvm-svn: 159494
2012-06-30 21:33:57 +00:00
Fariborz Jahanian d4879414f4 blocks: fixes a crash when encoding block type
with argument type of size 0. // rdar://11777609
PR13229.

llvm-svn: 159477
2012-06-30 00:48:59 +00:00
Fariborz Jahanian 271b8d43ce objective-c IRGen: fixes a crash when method type is being mangled
when an argument type size is 0. // rdar://11777609, PR13229

llvm-svn: 159472
2012-06-29 22:54:56 +00:00
David Blaikie 5d577a225e Use -frewrite-includes for crash reports.
In future changes we should:
* use __builtin_trap rather than derefing 'random' volatile pointers.
* avoid dumping temporary files into /tmp when running tests, instead
  preferring a location that is properly cleaned up by lit.

Review by Chandler Carruth.

llvm-svn: 159469
2012-06-29 22:03:56 +00:00
Kaelyn Uhrain 7da8c7d8d4 In Sema::ClassifyName, try to avoid nonsensical corrections to
keywords when doing type correction.

llvm-svn: 159464
2012-06-29 21:30:39 +00:00
Richard Trieu c605844510 Update the %diff modifer to have an alternate string print when a template tree
is selected.  This will allow more flexibility when converting diagnostics to
use template type diffing.

Also updated the internal manual and test cases for correctly keeping the bold
attribute and for tree printing.

llvm-svn: 159463
2012-06-29 21:12:16 +00:00
Ted Kremenek afddb9c81c Revert "Tweak insecureAPI analyzer checks to have the ability to be individually disabled."
Jordan Rose corrected me that this actually isn't needed.

llvm-svn: 159462
2012-06-29 21:01:35 +00:00
Ted Kremenek a33b078e73 Tweak insecureAPI analyzer checks to have the ability to be individually disabled.
The solution is a bit inefficient: it creates N checkers, one for each check, and
each check does a dispatch on the function name.  This is redundant, but we can fix
this once we have the proper ability to enable/disable subchecks.

Fixes <rdar://problem/11780180>.

llvm-svn: 159459
2012-06-29 20:44:58 +00:00
Richard Trieu cf86744ffe Add a fix-it hint note to -Wunique-enum to suggest that the last element gets
initialized with the next to last element to silence the warning.

llvm-svn: 159458
2012-06-29 20:36:14 +00:00
Fariborz Jahanian 2794ad59aa objective-c modern translator: synthesize argument type
correctly for blocks and function pointer arguments
in the written constructor.  // rdar://11359268

llvm-svn: 159456
2012-06-29 19:55:46 +00:00
Fariborz Jahanian 0ad4d60b8c modern objc translator: string.h is unavailable on some non-OSX
platforms declare memset in rewritten code instead.

llvm-svn: 159454
2012-06-29 19:33:05 +00:00
Fariborz Jahanian 9699c1ef38 fix a typo in my last commit. I thought I recompiled,
but apparently I did not.

llvm-svn: 159452
2012-06-29 19:05:11 +00:00
Fariborz Jahanian 1db30fc071 objc diagnostic: pass IdentifierInfo* to the diagnostic system
to produce quotes instead of adding qoute to the test.

llvm-svn: 159450
2012-06-29 18:43:30 +00:00
Douglas Gregor 158dec5d20 std::nullptr_t support in MS headers, from João Matos.
llvm-svn: 159448
2012-06-29 18:28:41 +00:00
Douglas Gregor 3f568c100e Patch for handling C99 veriadic macros when using precompiled headers,
from Filipe Cabecinhas!

llvm-svn: 159446
2012-06-29 18:27:59 +00:00
Fariborz Jahanian 4a031bdc8f objective-c modern translator: Translation into objc_msgSend_stret
entry point which requires nil check before calling objc_msgSend_stret.
// rdar://11359268 - wip.

llvm-svn: 159445
2012-06-29 18:27:08 +00:00
Dmitri Gribenko 025d518e3e Factor out a check for block commands (that implicitly start a new paragraph) into a separate function.
llvm-svn: 159444
2012-06-29 18:19:20 +00:00
David Blaikie af031a9af7 Avoid redundant error when redefining a function as deleted.
Reviewed by Doug Gregor.

llvm-svn: 159442
2012-06-29 18:00:25 +00:00
Nico Weber 3052abd9c2 Change condition to be the same as in SemaTemplateInstantiate.
No functionality change.

llvm-svn: 159436
2012-06-29 16:39:58 +00:00
Chandler Carruth 6b22da9e89 Update Clang to reflect the new home of IRBuilder.h as of r159421.
llvm-svn: 159422
2012-06-29 12:39:08 +00:00
Bill Wendling 128f8bdb78 Update based on move of DIBuilder.h to include/llvm.
llvm-svn: 159415
2012-06-29 08:32:29 +00:00
Axel Naumann 693344b505 From Philippe Canal:
Update the two function overloads
    void TemplateSpecializationType::PrintTemplateArgumentList(raw_ostream &OS,....
to behave like
    std::string TemplateSpecializationType::PrintTemplateArgumentList(const TemplateArgument *Args,...
hence making sure that clang consistently adds a space between two '>' at the end of nested template arguments.

llvm-svn: 159412
2012-06-29 07:30:33 +00:00
Manman Ren f865ba0c0e X86: add more GATHER intrinsics in Clang
Support the following intrinsics:
  _mm_i32gather_pd, _mm256_i32gather_pd,
  _mm_i64gather_pd, _mm256_i64gather_pd,
  _mm_i32gather_ps, _mm256_i32gather_ps,
  _mm_i64gather_ps, _mm256_i64gather_ps,
  _mm_i32gather_epi64, _mm256_i32gather_epi64,
  _mm_i64gather_epi64, _mm256_i64gather_epi64,
  _mm_i32gather_epi32, _mm256_i32gather_epi32,
  _mm_i64gather_epi32, _mm256_i64gather_epi32

llvm-svn: 159410
2012-06-29 05:19:13 +00:00
Timur Iskhodzhanov 099ec01569 Fix PR13234 - crash when trying to report an unsupported ABI feature
llvm-svn: 159405
2012-06-29 01:14:21 +00:00
Douglas Gregor 98c3cfc3b9 When a builtin that requires a constant is given a type- or
value-dependent expression, don't complain that it wasn't the constant
we wanted. Fixes <rdar://problem/11688587> and PR11074.

llvm-svn: 159404
2012-06-29 01:05:22 +00:00
Manman Ren 86c3250b82 X86: add more GATHER intrinsics in Clang
Corrected type for index of _mm256_mask_i32gather_pd
  from 256-bit to 128-bit
Corrected types for src|dst|mask of _mm256_mask_i64gather_ps
  from 256-bit to 128-bit

Support the following intrinsics:
  _mm_mask_i32gather_epi64, _mm256_mask_i32gather_epi64,
  _mm_mask_i64gather_epi64, _mm256_mask_i64gather_epi64,
  _mm_mask_i32gather_epi32, _mm256_mask_i32gather_epi32,
  _mm_mask_i64gather_epi32, _mm256_mask_i64gather_epi32

llvm-svn: 159403
2012-06-29 00:54:35 +00:00
Douglas Gregor 6bd56ca836 Teach the __is_trivially_assignable and __is_trivially_constructible
type traits that assignment to/construction of a lifetime-qualified
object under ARC is *not* trivial. Fixes <rdar://problem/11738725>.

llvm-svn: 159401
2012-06-29 00:49:17 +00:00
Jordan Rose 42ee04d00a [analyzer] Add a test that we are, in fact, doing a DFS on the ExplodedGraph.
Previously:
...the comment said DFS...
...the WorkList being instantiated said BFS...
...and the implementation was actually DFS...
...due to an unintentional change in 2010...
...and everything kept working anyway.

This fixes our std::deque implementation of BFS, but switches back to a
SmallVector-based implementation of DFS.

We should probably still investigate the ramifications of DFS vs. BFS,
especially for large functions (and especially when we hit our block path
limit), since this might completely change our memory use. It can also mask
some bugs and reveal others depending on when we halt analysis. But at least
we will not have this kind of little mistake creep in again.

llvm-svn: 159397
2012-06-29 00:33:10 +00:00
Nico Weber 33fd523df1 Warn on self-assignment to member variables. PR13104.
llvm-svn: 159394
2012-06-28 23:53:12 +00:00
DeLesley Hutchins d162c91b1a Thread safety analysis: support release() function on scoped
lockable objects.

llvm-svn: 159387
2012-06-28 22:42:48 +00:00
Richard Trieu a71f0de454 When applying a template diff highlighting to a diagnostic message, remember
to reapply the bold formatting when needed.

llvm-svn: 159386
2012-06-28 22:39:03 +00:00
David Blaikie 9e55d74074 Fix template type diffing coloring (r159216) when forcing color output to a file (not a terminal)
Reviewed (over the shoulder) by Richard Trieu.

llvm-svn: 159381
2012-06-28 21:46:07 +00:00
Douglas Gregor 1edf57639a Support the use of "=delete" and "=default" with delayed template
parsing. Fixes <rdar://problem/11700604>.

llvm-svn: 159380
2012-06-28 21:43:01 +00:00
Fariborz Jahanian d9c8aace7b objective-c rewriter: refactoring of rewriting of
objc_msgSend_stret() API. No functionality change.

llvm-svn: 159379
2012-06-28 21:20:35 +00:00
Benjamin Kramer 46a72fb741 Dead code eliminate the massive hexagon builtin intrinsic supporting code.
The tablegen'd code does the same thing without this egregious duplication.
In my limited testing everything seems to work, however there can be
differences if the clang and llvm builtin definitions don't match.

llvm-svn: 159371
2012-06-28 20:08:55 +00:00
Benjamin Kramer 8652ca8a6a Now that we use the GCC builtin <-> llvm intrinsic, dead code eliminate the handwritten emitter.
The generated code uncovered an invalid prototype for __builtin_mips_shilo, fix it along the way.

llvm-svn: 159368
2012-06-28 19:10:01 +00:00
Simon Atanasyan 07ce7d8fb5 Support MIPS DSP Rev1 intrinsics.
This patch was reviewed in the llvm-commits list by Jim Grosbach.

llvm-svn: 159366
2012-06-28 18:23:16 +00:00
Rafael Espindola b27564af76 Compare the canonical types and document why we give up on the covariant case.
llvm-svn: 159360
2012-06-28 17:57:36 +00:00
Jordan Rose ee3d99f6be Warn on weak properties declared in protocols as well.
Previously this caused a crash, since protocols are not interfaces.

llvm-svn: 159357
2012-06-28 16:39:28 +00:00
Dmitri Gribenko 5c8897d37d Remove redundant check.
llvm-svn: 159355
2012-06-28 16:25:36 +00:00
Dmitri Gribenko df17d64881 Attaching comments to declarations: don't attach comments to implicit declarations.
llvm-svn: 159354
2012-06-28 16:19:39 +00:00
Rafael Espindola debc71cea3 Disable devirtualization when we have covariant returns. I will open a bug
for tracking this.

llvm-svn: 159351
2012-06-28 15:11:39 +00:00
Rafael Espindola 3b33c4ecf0 Don't devirtualize calls when we don't have the correct type of the this pointer
handy. It can be done, but we would have to build a derived-to-base cast
during codegen to compute the correct this pointer.

I will handle covariant returns next.

llvm-svn: 159350
2012-06-28 14:28:57 +00:00
Hans Wennborg f60f6af9e8 Add -ftls-model command-line flag.
This allows for setting the default TLS model. (PR9788)

llvm-svn: 159336
2012-06-28 08:01:44 +00:00
Richard Smith 608c0b65d7 Add warning flag -Winvalid-pp-token for preprocessing-tokens which have
undefined behaviour, and move the diagnostic for '' from an Error into
an ExtWarn in this group. This is important for some users of the preprocessor,
and is necessary for gcc compatibility.

llvm-svn: 159335
2012-06-28 07:51:56 +00:00
Rafael Espindola ecbe2e9795 Fix another issue with devirtualizing calls to final methods by passing them
the correct this pointer. There is some potential for sharing a bit more
code with canDevirtualizeMemberFunctionCalls, but that can be done in an
independent patch.

llvm-svn: 159326
2012-06-28 01:56:38 +00:00
Dmitri Gribenko 0743f94671 Cleanup \brief comment. Since it is a single paragraph, no need to save newlines there.
llvm-svn: 159325
2012-06-28 01:38:21 +00:00
Eric Christopher c939e50846 Fix grammar.
llvm-svn: 159321
2012-06-28 01:20:05 +00:00
Bill Wendling 4adcf78275 Update the #include to find the DebugInfo.h in the correct place
llvm-svn: 159315
2012-06-28 00:30:19 +00:00
Dmitri Gribenko a1e9c8e783 Teach \brief parser about commands that start a new paragraph implicitly
llvm-svn: 159309
2012-06-28 00:01:41 +00:00
Argyrios Kyrtzidis 92aae0f7a4 Fix uninitialized variable use bug found by the clairvoyant static analyzer.
Commit::canReplaceText would not initialize its out 'Len' parameter before
returning true and it would be used uninitialized in Commit::replaceText.

llvm-svn: 159306
2012-06-27 23:45:44 +00:00
Dmitri Gribenko 7e8729b904 Attaching documentation comments to declarations: don't attach a comment to a declaration if there is a preprocessor directive between them.
llvm-svn: 159305
2012-06-27 23:43:37 +00:00
Dmitri Gribenko 632d58afab Fix an infinite loop in comment lexer: we were not advancing in the input character stream when we saw a '<' that is not a start of an HTML tag.
llvm-svn: 159303
2012-06-27 23:28:29 +00:00
Eli Friedman c24e2fb1fb Propagate lvalue alignment into bitfields. Per report on cfe-dev.
llvm-svn: 159295
2012-06-27 21:19:48 +00:00
Fariborz Jahanian bbd0d6992b objective-c mrc: Issue warning for mrc, as is done for arc, when
property retains a block object as it could be on 
the stack. // rdar://11761511

llvm-svn: 159293
2012-06-27 20:52:46 +00:00
Richard Smith 7659b12b02 Refactoring after r159290: don't hold onto and check a misleading QualType.
llvm-svn: 159292
2012-06-27 20:29:39 +00:00
Richard Smith 56471fdba8 Check for non-POD vararg argument type after default argument promotion, not
before, so we don't incorrectly think arguments of function type are non-POD.

llvm-svn: 159290
2012-06-27 20:23:58 +00:00
Fariborz Jahanian 91fc39e313 patch to suggest 'static' function should be 'static inline'
when it appears to be unused and occurs in a header.
// rdar://11202617

llvm-svn: 159282
2012-06-27 19:43:29 +00:00
Rafael Espindola b7f5a9c5cd Implement John McCall's review of r159212 other than the this pointer not
being updated. Will fix that in a second.

llvm-svn: 159280
2012-06-27 18:18:05 +00:00
Rafael Espindola a245edc82f Fix a crash I introduced in r159212.
llvm-svn: 159279
2012-06-27 17:44:39 +00:00
Dmitri Gribenko 1669f70273 Remove unsigned and a pointer from a comment token (so that each token can have only one semantic string value attached to it), at a cost of adding an additional token.
llvm-svn: 159270
2012-06-27 16:53:58 +00:00
Dmitri Gribenko 60ddd8a1b1 Comment lexer: counting backwards from token end is thought to be confusing. We already have a pointer to the beginning of the token, so use it to extract the text instead.
llvm-svn: 159269
2012-06-27 16:30:35 +00:00
Axel Naumann b15b37f73a Fix for r159256 on Windows.
llvm-svn: 159262
2012-06-27 11:42:56 +00:00
Axel Naumann 38179d9693 From Vassil Vassilev:
add interface for removing a FileEntry from the cache.
Forces a re-read the contents from disk, e.g. because a tool (like cling) wants to pick up a modified file.

llvm-svn: 159256
2012-06-27 09:17:42 +00:00
Dmitri Gribenko 60c7ec699b Initialize RawComment::BriefTextValid in other constructor, too.
llvm-svn: 159253
2012-06-27 05:48:36 +00:00
Daniel Jasper 37184c2e66 Introduce __has_feature(attribute_unused_on_fields) to determine whether
the current version of clang understands __attribute__((unused)) on
fields.

llvm-svn: 159252
2012-06-27 05:12:34 +00:00
Dmitri Gribenko 99e0942cd9 Simplify logic in BriefParser::Parse(), per Jordan's comment.
llvm-svn: 159247
2012-06-27 01:17:34 +00:00
Jordan Rose 89b917e399 [analyzer] RetainCountChecker: remove unused SelfOwn ArgEffect kind.
llvm-svn: 159245
2012-06-27 00:51:18 +00:00
Jordan Rose 88be136cff [analyzer] Remove unneeded helper function (it's in ASTContext.h)
llvm-svn: 159244
2012-06-27 00:51:16 +00:00
Chad Rosier c118395cff Whitespace.
llvm-svn: 159235
2012-06-26 22:30:43 +00:00
Timur Iskhodzhanov 1b1d630e0c [Windows] Improve mangling of templates when back references are present
llvm-svn: 159234
2012-06-26 22:29:50 +00:00
Benjamin Kramer 09811c7dd5 Enable -mcpu=native and -march=native for arm targets.
This is only implemented on linux at the moment.

llvm-svn: 159232
2012-06-26 22:20:06 +00:00
Chad Rosier 97e27446e0 Whitespace.
llvm-svn: 159229
2012-06-26 21:41:40 +00:00
Fariborz Jahanian 0638c15a52 preprocessing: gcc supports #line 0. So, treat this
as a gcc supported extension with usual treatment
with -pedantic (warn) and -pedantic-errors (error).
// rdar://11550996

llvm-svn: 159226
2012-06-26 21:19:20 +00:00
Richard Smith b063d93cfb Remove typedef which is unused after r159189.
llvm-svn: 159224
2012-06-26 20:54:26 +00:00
Dmitri Gribenko 5188c4b9cc Implement a lexer for structured comments.
llvm-svn: 159223
2012-06-26 20:39:18 +00:00
Manman Ren add5e9e289 X86: add GATHER intrinsics (AVX2) in Clang
Support the following intrinsics:
  _mm_mask_i32gather_pd, _mm256_mask_i32gather_pd, _mm_mask_i64gather_pd
  _mm256_mask_i64gather_pd, _mm_mask_i32gather_ps, _mm256_mask_i32gather_ps
  _mm_mask_i64gather_ps, _mm256_mask_i64gather_ps

llvm-svn: 159222
2012-06-26 19:55:09 +00:00
Rafael Espindola 727a771a5f Fix a bug in my previous patch: If we are not doing a virtual call because
the member expression is qualified, call the method specified in the code,
not the most derived one we can find.

llvm-svn: 159219
2012-06-26 19:18:25 +00:00
David Blaikie 47e4518117 Remove spurious semicolons committed in r159216.
llvm-svn: 159218
2012-06-26 18:52:09 +00:00