Chandler Carruth [Tue, 29 Mar 2011 08:08:18 +0000 (08:08 +0000)]
Fix a bug in how we were resolving the address of overloaded functions
when the resolution took place due to a single template specialization
being named with an explicit template argument list. In this case, the
"resolution" doesn't take into account the target type at all, and
therefore can take place for functions, static member functions, and
*non-static* member functions. The latter weren't being properly checked
and their proper form enforced in this scenario. We now do so.
The result of this last form slipping through was some confusing logic
in IsStandardConversion handling of these resolved address-of
expressions which eventually exploded in an assert. Simplify this logic
a bit and add some more aggressive asserts to catch improperly formed
expressions getting into this routine.
Finally add systematic testing of member functions, both static and
non-static, in the various forms they can take. One of these is
essentially PR9563, and this commit fixes the crash in that PR. However,
the diagnostics for this are still pretty terrible. We at least are now
accepting the correct constructs and rejecting the invalid ones rather
than accepting invalid or crashing as before.
Ted Kremenek [Tue, 29 Mar 2011 01:40:00 +0000 (01:40 +0000)]
Add workaround for Sema issue found in <rdar://problem/9188004>, which leads to an assertion failure in the uninitialized variables analysis. The problem is that Sema isn't properly registering a variable in a DeclContext (which -Wuninitialized relies on), but
my expertise on the template instantiation logic isn't good enough to fix this problem for real. This patch worksaround the
problem in -Wuninitialized, but we should fix it for real later.
Chandler Carruth [Sun, 27 Mar 2011 20:00:08 +0000 (20:00 +0000)]
Flip the default for showing include stacks on notes to false. This
required modifying a few tests that specifically use note include stacks
to check the source manager's view of include stacks. I've simply added
the flag to these tests for now, they may have to be more substantially
changed if we decide to remove support for note include stacks
altogether.
Also, add a test for include stacks on notes that was supposed to go in
with the previous commit.
Chandler Carruth [Sun, 27 Mar 2011 09:46:56 +0000 (09:46 +0000)]
Diagnose uninitialized uses of a variable within its own initializer.
This is basically the same idea as the warning on uninitialized uses of
fields within an initializer list. As such, it is on by default and
under -Wuninitialized.
Original patch by Richard Trieu, with some massaging from me on the
wording and grouping of the diagnostics.
Anton Yartsev [Sun, 27 Mar 2011 09:32:40 +0000 (09:32 +0000)]
supported: AltiVec vector initialization with a single literal according to PIM section 2.5.1 - after initialization all elements have the value specified by the literal
John McCall [Sun, 27 Mar 2011 09:00:25 +0000 (09:00 +0000)]
We were emitting construction v-tables with internal linkage all the time.
Emit them instead with the linkage of the VTT.
I'm actually really ambivalent about this; it's what GCC does, but outside
of improving code size (if the linkage is coalescing), I'm not sure it's
at all relevant. Construction vtables are naturally referenced only by the
VTT, which is itself only referenced by complete-object constructors and
destructors; giving the construction vtables possibly-external linkage is
important if you have an optimization that drills through the VTT to a
reference to a particular construction vtable which it cannot just emit
itself.
Chandler Carruth [Sun, 27 Mar 2011 01:50:55 +0000 (01:50 +0000)]
Add an option to suppress include stack printing on note diagnostics.
These stacks are often less important than those on primary diagnostics.
As the number of notes grows, this becomes increasingly important. The
include stack printing is clever and doesn't print stacks for adjacent
diagnostics from the same file, but when a note is in between a sequence
of errors in a header file, and the notes all refer to some other file,
we end up getting a worst-case ping-pong of include stacks that take up
a great deal of vertical space.
Still, for now, the default behavior isn't changed. We can evaluate user
feedback with the flag.
Patch by Richard Trieu, a couple of style tweaks from me.
Chandler Carruth [Sun, 27 Mar 2011 00:04:55 +0000 (00:04 +0000)]
Add -f[no-]strict-overflow to the Clang driver. Use it to set the
default for -fwrapv if that flag isn't specified explicitly. We always
prefer an explict setting of -fwrapv when present. Also adds support for
-fno-wrapv to allow disabling -fwrapv even when -fno-strict-overflow is
passed.
Douglas Gregor [Sat, 26 Mar 2011 12:10:19 +0000 (12:10 +0000)]
On Mac OS X, the presence of an 'availability' attribute for that
platform implies default visibility. To achieve these, refactor our
lookup of explicit visibility so that we search for both an explicit
VisibilityAttr and an appropriate AvailabilityAttr, favoring the
VisibilityAttr if it is present.
David Chisnall [Sat, 26 Mar 2011 11:48:37 +0000 (11:48 +0000)]
Reformatted doc comments so that they are now difficult to edit in any editor that doesn't have explicit doxygen support, as per LLVM style guidelines.
Douglas Gregor [Sat, 26 Mar 2011 03:35:55 +0000 (03:35 +0000)]
Extend the new 'availability' attribute with support for an
'unavailable' argument, which specifies that the declaration to which
the attribute appertains is unavailable on that platform.
John McCall [Sat, 26 Mar 2011 01:53:26 +0000 (01:53 +0000)]
Fix the recovery from missing semis on @property declarations to not consume
the following '@'. Conceivably, we could skip tokens until something that
can validly start an @interface declaration here, but it's not clear that
it matters.
Devang Patel [Fri, 25 Mar 2011 21:26:13 +0000 (21:26 +0000)]
Provide blockDecl's startLoc to startFunction. This fixes hidden bug exposed by recent code gen changes. This is tested by global-blocks-lines.exp in gdb testsuite.
David Chisnall [Fri, 25 Mar 2011 11:57:33 +0000 (11:57 +0000)]
Continuing work on ObjC tidyup:
- Moved the CGObjCRuntime functions out of CGObjCMac.cpp into CGObjCRuntime.cpp
- Added generic functions in CGObjCRuntime for emitting @try and @synchronize
blocks, usable by any runtime that uses DWARF exceptions.
- Made the GNU runtimes use these functions.
It should now be possible to replace the equivalent functions in
CGObjCNonFragileABIMac with simple calls to these two functions, providing the
runtime functions as arguments. I'll post a diff to the list for review before
making any changes to the Mac runtime stuff.
Douglas Gregor [Thu, 24 Mar 2011 14:35:16 +0000 (14:35 +0000)]
Minor fix in the injection of labels, since we want to look at the redeclaration context of each declaration in the identifier chain. Should fix Linux self-host
John McCall [Thu, 24 Mar 2011 11:26:52 +0000 (11:26 +0000)]
Insomniac refactoring: change how the parser allocates attributes so that
AttributeLists do not accumulate over the lifetime of parsing, but are
instead reused. Also make the arguments array not require a separate
allocation, and make availability attributes store their stuff in
augmented memory, too.
Douglas Gregor [Thu, 24 Mar 2011 10:35:39 +0000 (10:35 +0000)]
Fix the insertion of label declarations into the identifier chain in
the case where we only have a single identifier with that name in the
chain. Fixes PR9463 for real this time.
David Chisnall [Wed, 23 Mar 2011 22:52:06 +0000 (22:52 +0000)]
Fixed type error in last commit (forgot that now that selectors are not
accessed via the indirect pointer, they don't need to be pointers to pointers).
Finished moving the message lookup code into separate subclasses for each
runtime. Also performed a few smallish related tidies.
We're now bitcasting the result of the message lookup functions, rather than
casting the lookup functions themselves, so the messages.m test needed updating
to reflect this.
Ted Kremenek [Wed, 23 Mar 2011 21:33:21 +0000 (21:33 +0000)]
Fix CFG-construction bug when run from AnalysisBasedWarnings::IssueWarnings() where block-level expressions that need
to be recorded in the Stmt*->CFGBlock* map were not always done so. Fixes <rdar://problem/9171946>.
David Chisnall [Wed, 23 Mar 2011 16:36:54 +0000 (16:36 +0000)]
Initial work on refactoring GNU runtime code (long overdue - it's quite obvious
that I hadn't used C++ for several years before writing most of this code).
Still lots more to do. This set of changes includes:
- Remove the distinction between typed and untyped selectors. More accurately
reflect what the runtime does, by using typed selectors everywhere, with an
empty type field if the types are unknown. Now we just store a small list of
types for each selector (in theory, this should always be exactly one, but
this constraint was not enforced back in 1986 when it should have been).
- Add some consistency to how runtime functions are created. These are all
generated via the LazyRuntimeFunction class (which might be useful outside
CGObjCGNU - feel free to move it into a header if it is). This function
stores the types of a function, looks it up the first time it's used, and
caches the result. This means that we're now not wasting time constructing
the llvm::FunctionType every time some of the functions are looked up, but
also not inserting references to runtime functions into the module if they're
not actually used.
- Started separating out the fragile and non-fragile ABI behaviours into two
subclasses of CGObjCGNU: CGObjCGCC for the legacy GCC runtime ABI and
CGObjCGNUstep for the new GNUstep ABI. Not all of the differences in
behaviour are factored out yet, but they will be in future commits.
- Removed all of the CodeGen:: things: we've been using namespace CodeGen in
this file for ages, so having explicit namespace specifiers is just a bit
confusing.
- Added a few more comments.
- Used llvm::StringRef instead of std::string in a few places.
- Finally got around to storing the module path in the module structure. The
ABI says that the compiler should do this, although it's not used in the
runtime or exposed outside the runtime, so it's pretty useless.
Still to do:
- We currently have two code paths for generating try blocks, one for ObjC and
one for ObjC++. Not only are these substantially similar, they are also very
similar to the CGObjCMac version. These need factoring out into a single
parameterised implementation, either in CGObjCRuntime or CodeGenFunction.
The EmitObjCXXTryStmt() function was added so that the changes to fix a bug
in time for the 2.9 release would be self-contained and reduce the chances of
breaking anything else, but these should be done properly as soon as
possible.
- Split up some large functions (e.g. GenerateClass()) into smaller functions
for generating the various data structures.
- The method lookup code into the two subclasses, removing the conditionals in
the message send functions.
- Add doxygen comments on the remaining undocumented functions.
- We seem to be generating global pointer variables for selectors, then storing
a pointer to the selector, then generating a load of this pointer (and then a
load of the real selector later) every time a static selector is used. I can
only assume I was asleep or drunk when I did this - we should just be
referencing the selectors directly in the selector array.
Douglas Gregor [Wed, 23 Mar 2011 15:13:44 +0000 (15:13 +0000)]
Teach DelayedDiagnostic to copy its string, rather than hope that the
string itself lives longer than the DelayedDiagnostic. Fixes a recent
use-after-free regression due to my availability attribute work.
Ted Kremenek [Wed, 23 Mar 2011 02:16:44 +0000 (02:16 +0000)]
Fix crash in clang_getInstantiationLoc() when SourceManager::getInstantiationLoc() can return a SourceLocatin with an invalid
FileID on invalid code. Fixes <rdar://problem/9164623>.
says that the function "foo" was introduced in 10.2, deprecated in
10.4, and completely obsoleted in 10.6. This attribute ties in with
the deployment targets (e.g., -mmacosx-version-min=10.1 specifies that
we want to deploy back to Mac OS X 10.1). There are several concrete
behaviors that this attribute enables, as illustrated with the
function foo() above:
- If we choose a deployment target >= Mac OS X 10.4, uses of "foo"
will result in a deprecation warning, as if we had placed
attribute((deprecated)) on it (but with a better diagnostic)
- If we choose a deployment target >= Mac OS X 10.6, uses of "foo"
will result in an "unavailable" warning (in C)/error (in C++), as
if we had placed attribute((unavailable)) on it
- If we choose a deployment target prior to 10.2, foo() is
weak-imported (if it is a kind of entity that can be weak
imported), as if we had placed the weak_import attribute on it.
Naturally, there can be multiple availability attributes on a
declaration, for different platforms; only the current platform
matters when checking availability attributes.
The only platforms this attribute currently works for are "ios" and
"macosx", since we already have -mxxxx-version-min flags for them and we
have experience there with macro tricks translating down to the
deprecated/unavailable/weak_import attributes. The end goal is to open
this up to other platforms, and even extension to other "platforms"
that are really libraries (say, through a #pragma clang
define_system), but that hasn't yet been designed and we may want to
shake out more issues with this narrower problem first.
projects / clang / log