John McCall [Wed, 20 Jun 2012 06:18:46 +0000 (06:18 +0000)]
Restructure how the driver communicates information about the
target Objective-C runtime down to the frontend: break this
down into a single target runtime kind and version, and compute
all the relevant information from that. This makes it
relatively painless to add support for new runtimes to the
compiler. Make the new -cc1 flag, -fobjc-runtime=blah-x.y.z,
available at the driver level as a better and more general
alternative to -fgnu-runtime and -fnext-runtime. This new
concept of an Objective-C runtime also encompasses what we
were previously separating out as the "Objective-C ABI", so
fragile vs. non-fragile runtimes are now really modelled as
different kinds of runtime, paving the way for better overall
differentiation.
As a sort of special case, continue to accept the -cc1 flag
-fobjc-runtime-has-weak, as a sop to PLCompatibilityWeak.
I won't go so far as to say "no functionality change", even
ignoring the new driver flag, but subtle changes in driver
semantics are almost certainly not intended.
Jordan Rose [Wed, 20 Jun 2012 05:34:32 +0000 (05:34 +0000)]
[analyzer] Move failing 'new' test cases back into new.cpp instead of XFAILing.
Per Anna's comment, this is a better way to handle "to-do list"-type failures.
This way we'll know if any of the features get fixed; in an XFAIL file, /all/
the cases have to be fixed before lit would tell us anything.
Jordan Rose [Wed, 20 Jun 2012 01:32:01 +0000 (01:32 +0000)]
[analyzer] Invalidate placement args; return the pointer given to placement new
The default global placement new just returns the pointer it is given.
Note that other custom 'new' implementations with placement args are not
guaranteed to do this.
In addition, we need to invalidate placement args, since they may be updated by
the allocator function. (Also, right now we don't properly handle the
constructor inside a CXXNewExpr, so we need to invalidate the placement args
just so that callers know something changed!)
This invalidation is not perfect because CallOrObjCMessage doesn't support
CXXNewExpr, and all of our invalidation callbacks expect that if there's no
CallOrObjCMessage, the invalidation is happening manually (e.g. by a direct
assignment) and shouldn't affect checker-specific metadata (like malloc state);
hence the malloc test case in new-fail.cpp. But region values are now
properly invalidated, at least.
The long-term solution to this problem is to rework CallOrObjCMessage into
something more general, rather than the morass of branches it is today.
James Dennett [Wed, 20 Jun 2012 00:56:32 +0000 (00:56 +0000)]
Documentation cleanup:
* Escaped # and < characters in Doxygen comments as needed;
* Removed a Doxygen comment in HeaderSearch.cpp that was redundant with
the corresponding comment in the header file.
James Dennett [Wed, 20 Jun 2012 00:50:53 +0000 (00:50 +0000)]
Documentation cleanup:
* Escape < characters in Doxygen comments as needed;
* Demote one Doxygen comment to a regular comment to fix a Doxygen warning.
Kaelyn Uhrain [Wed, 20 Jun 2012 00:36:03 +0000 (00:36 +0000)]
Add the PCH file name to the message about not being able to read the PCH.
Also add a couple of unit tests to check the invalid-PCH error messages
to satisfy PR4568 and for the assertion (introduced in r149918 and fixed
in r158769) that would cause clang to crash when given an empty PCH.
Dmitri Gribenko [Wed, 20 Jun 2012 00:34:58 +0000 (00:34 +0000)]
Structured comment parsing, first step.
* Retain comments in the AST
* Serialize/deserialize comments
* Find comments attached to a certain Decl
* Expose raw comment text and SourceRange via libclang
Sean Hunt [Tue, 19 Jun 2012 23:57:03 +0000 (23:57 +0000)]
Reapply r158700 and fixup patches, minus one hunk that slipped through and
caused a crash in an obscure case. On the plus side, it caused me to catch
another bug by inspection.
Tanya Lattner [Tue, 19 Jun 2012 23:09:52 +0000 (23:09 +0000)]
Extend the support for cl-std to include 1.2.
Add error checking for the static qualifier which is now allowed in certain situations for OpenCL 1.2. Use the CL version to turn on this feature.
Added test case for 1.2 static storage class feature.
objective-c: warn when autosynthesizing a property which has same
name as an existing ivar since this is common source of error
when people remove @synthesize to take advantage of autosynthesis.
// rdar://11671080
Richard Smith [Tue, 19 Jun 2012 21:28:35 +0000 (21:28 +0000)]
Fix -Wc++11-narrowing warnings for narrowing negative values to larger unsigned
types to actually includes the value, rather than saying <uninitialized>.
objc-arc: captured block variable accessed in its block literal
initializer need be null initialized before initializer takes
hold, just like any other initialized retainable object pointer.
// rdar://11016025
Meador Inge [Tue, 19 Jun 2012 18:17:30 +0000 (18:17 +0000)]
Revert predefined decl tracking.
r158085 added some logic to track predefined declarations. The main reason we
had predefined declarations in the input was because the __builtin_va_list
declarations were injected into the preprocessor input. As of r158592 we
explicitly build the __builtin_va_list declarations. Therefore the predefined
decl tracking is no longer needed.
Chad Rosier [Tue, 19 Jun 2012 17:48:02 +0000 (17:48 +0000)]
[driver] Make the crash diagnostic message more visable. Bug reports are being
filed, but still missing the preprocessed source and associated run script.
rdar://11684107
Sean Hunt [Tue, 19 Jun 2012 03:39:03 +0000 (03:39 +0000)]
Improve the specification of spellings in Attr.td.
Note that this is mostly a structural patch that handles the change from the old
spelling style to the new one. One consequence of this is that all AT_foo_bar
enum values have changed to not be based off of the first spelling, but rather
off of the class name, so they are now AT_FooBar and the like (a straw poll on
IRC showed support for this). Apologies for code churn.
Most attributes have GNU spellings as a temporary solution until everything else
is sorted out (such as a Keyword spelling, which I intend to add if someone else
doesn't beat me to it). This is definitely a WIP.
I've also killed BaseCheckAttr since it was unused, and I had to go through
every attribute anyway.
[objcmt] When checking whether the subscripting methods are declared use
ObjCInterfaceDec::lookupInstanceMethod to make sure we check categories as well
and update related tests.
Rafael Espindola [Tue, 19 Jun 2012 01:26:10 +0000 (01:26 +0000)]
Add a -fuse-init-array option to cc1 and map to the UseInitArray target
option. On the driver, check if we are using libraries from gcc 4.7 or newer
and if so pass -fuse-init-array to the frontend.
The crtbegin*.o files in gcc 4.7 no longer call the constructors listed in
.ctors, so we have to use .init_array.
Kaelyn Uhrain [Tue, 19 Jun 2012 00:37:47 +0000 (00:37 +0000)]
Improve the error message when a function overload candidate is rejected
because it expects a reference and receives a non-l-value.
For example, given:
int foo(int &);
template<int x> void b() { foo(x); }
clang will now print "expects an l-value for 1st argument" instead of
"no known conversion from 'int' to 'int &' for 1st argument". The change
in wording (and associated code to detect the case) was prompted by
comment #5 in PR3104, and should be the last bit of work needed for the
bug.
Ted Kremenek [Tue, 19 Jun 2012 00:37:39 +0000 (00:37 +0000)]
Sink definition of IBOutlet, IBOutletCollection, and IBAction into
the compiler predefines buffer. These are essentially part of
the Objective-C language.
Jordan Rose [Mon, 18 Jun 2012 23:58:49 +0000 (23:58 +0000)]
Change -Winternal-linkage-in-inline from ExtWarn to Warning in C++.
Per post-commit review, it's not appropriate to use ExtWarn in C++, because
we can't prove that the inline function will actually be defined in more than
one place (and thus we can't prove that this violates the ODR).
This removes the warning entirely from uses in the main source file in C++.
Jordan Rose [Mon, 18 Jun 2012 22:09:19 +0000 (22:09 +0000)]
Support -Winternal-linkage-in-inline in C++ code.
This includes treating anonymous namespaces like internal linkage, and allowing
const variables to be used even if internal. The whole thing's been broken out
into a separate function to avoid nested ifs.
Jordan Rose [Mon, 18 Jun 2012 21:31:37 +0000 (21:31 +0000)]
Add new tool 'diag-build' for showing enabled warnings in a project.
diag-build acts as a wrapper for 'diagtool show-enabled', in the same way
that scan-build acts as a wrapper for the static analyzer. The common case is
simple: use 'diag-build make' or 'diag-build xcodebuild' to list the warnings
enabled for the first compilation command we see. Other build systems require
you to manually specify "dry-run" and "use $CC and $CXX"; if there is a build
system you are interested in, please add it to the switch statement.
diag-build is fairly stupid right now, but it serves its basic purpose.
Hopefully it can grow to meet any additional requirements.
Move a few static functions from DiagnosticRenderer.cpp into SourceManager.
This simplifies the code a little bit, since these functions all took a
SourceManager parameter and called a bunch of methods on it, and makes
the functions available to other users.
James Dennett [Mon, 18 Jun 2012 18:42:20 +0000 (18:42 +0000)]
Documentation cleanup:
* Escaped "::" where needed to prevent Doxygen trying to make links;
* Updated one mention of C++0x to refer to C++11;
* Fixed a \brief summary to make it somewhat concise.
Jordan Rose [Mon, 18 Jun 2012 17:49:58 +0000 (17:49 +0000)]
Allow internal decls in inline functions if the function is in the main file.
This handles the very common case of people writing inline functions in their
main source files and not tagging them as inline. These cases should still
behave as the user intended. (The diagnostic is still emitted as an extension.)
I'm reworking this code anyway to account for C++'s equivalent restriction in
[basic.def.odr]p6, but this should get some bots back to green.
Sean Hunt [Mon, 18 Jun 2012 16:13:52 +0000 (16:13 +0000)]
Handle C++11 attribute namespaces automatically.
Now, as long as the 'Namespaces' variable is correct inside Attr.td, the
generated code will correctly admit a C++11 attribute only when it has the
appropriate namespace(s).
Richard Smith [Mon, 18 Jun 2012 06:11:04 +0000 (06:11 +0000)]
Extend the error recovery for a template-argument-list terminated by '>>' to
also deal with '>>>' (in CUDA), '>=', and '>>='. Fix the FixItHints logic to
deal with cases where the token is followed by an adjacent '=', '==', '>=',
'>>=', or '>>>' token, where a naive fix-it would result in a differing token
stream on a re-lex.
Chandler Carruth [Sun, 17 Jun 2012 11:47:03 +0000 (11:47 +0000)]
Switch users of SmallMap to use SmallDenseMap instead.
The most important change here is that the destructor and copy
constructor for StoredDeclsList will now reliably be run. Previously,
the destructors at least were missed in some cases. See the LLVM commits
discussions for why SmallMap is broken and going away.
James Dennett [Sun, 17 Jun 2012 06:45:58 +0000 (06:45 +0000)]
Documentation cleanup:
* Escaped # characters in Doxygen comments as needed;
* Added/reformatted \brief docs;
* Used a \file comment to document the file (MultipleIncludeOpt.h).
James Dennett [Sun, 17 Jun 2012 04:36:28 +0000 (04:36 +0000)]
Documentation cleanup:
* Added \file, \brief and \verbatim...\endverbatim markup, particularly around
documentation of subset of the grammars that are being parsed.
James Dennett [Sun, 17 Jun 2012 03:40:43 +0000 (03:40 +0000)]
Documentation cleanup:
* Removed docs for Lexer::makeFileCharRange from Lexer.cpp, as they're in
the header file;
* Reworked the documentation for SkipBlockComment so that it doesn't confuse
Doxygen's comment parsing;
* Added another summary with \brief markup.
Richard Smith [Sat, 16 Jun 2012 23:34:14 +0000 (23:34 +0000)]
-Wuninitialized bugfix: when entering the scope of a variable with no
initializer, it is uninitialized, even if we may be coming from somewhere where
it was initialized.
Eli Friedman [Sat, 16 Jun 2012 02:19:17 +0000 (02:19 +0000)]
Fix Sema and IRGen for atomic compound assignment so it has the right semantics when promotions are involved.
(As far as I can tell, this only affects some edge cases.)
Jordan Rose [Sat, 16 Jun 2012 01:28:00 +0000 (01:28 +0000)]
[analyzer] Return an UnknownVal when we try to get the binding for a VLA.
This happens in C++ mode right at the declaration of a struct VLA;
MallocChecker sees a bind and tries to get see if it's an escaping bind.
It's likely that our handling of this is still incomplete, but it fixes a
crash on valid without disturbing anything else for now.
[AST/libclang] Fix the selector locations that are reported for a
method definition that has its '{' attached to the method name without
a space.
With a method like:
-(id)meth{
.....
}
the logic in ObjCMethodDecl that determined the selector locations got
confused because it was initialized based on an end location for '{' but
that end location changed to '}' after the method was finished.
Fix this by having an immutable end location for the declarator and
for getLocEnd() get the end location from the body itself.
Jordan Rose [Sat, 16 Jun 2012 00:09:20 +0000 (00:09 +0000)]
[analyzer] Buffers passed to CGBitmapContextCreate can escape.
Specifically, although the bitmap context does not take ownership of the
buffer (unlike CGBitmapContextCreateWithData), the data buffer can be extracted
out of the created CGContextRef. Thus the buffer is not leaked even if its
original pointer goes out of scope, as long as
- the context escapes, or
- it is retrieved via CGBitmapContextGetData and freed.
Actually implementing that logic is beyond the current scope of MallocChecker,
so for now CGBitmapContextCreate goes on our system function exception list.