Richard Smith [Wed, 14 Nov 2012 07:36:28 +0000 (07:36 +0000)]
PR14279: Work around this major miscompilation by treating move operations as
non-trivial if they would not call a move operation, even if they would in fact
call a trivial copy operation. A proper fix is to follow, but this small
directed fix is intended for porting to the 3.2 release branch.
Richard Smith [Wed, 14 Nov 2012 07:06:31 +0000 (07:06 +0000)]
When we have a MemberExpr referring to an overloaded static member function,
and we resolve it to a specific function based on the type which it's used as,
don't forget to mark it as referenced.
NAKAMURA Takumi [Wed, 14 Nov 2012 02:21:48 +0000 (02:21 +0000)]
SemaTemplateDeduction.cpp: Prune two obsolete descriptions. [-Wdocumentation]
* getMostSpecialized()
/// \param Index if non-NULL and the result of this function is non-nULL,
/// receives the index corresponding to the resulting function template
/// specialization.
* DeduceTemplateArguments()
/// \param Name the name of the function being called. This is only significant
/// when the function template is a conversion function template, in which
/// case this routine will also perform template argument deduction based on
/// the function to which
Eli Friedman [Wed, 14 Nov 2012 02:18:46 +0000 (02:18 +0000)]
Suppress elided variadic macro argument extension diagnostic for macros using
the related comma pasting extension.
In certain cases, we used to get two diagnostics for what is essentially one
extension. This change suppresses the first diagnostic in certain cases
where we know we're going to print the second diagnostic. The
diagnostic is redundant, and it can't be suppressed in the definition
of the macro because it points at the use of the macro, so we want to
avoid printing it if possible.
The implementation works by detecting constructs which look like comma
pasting at the time of the definition of the macro; this information
is then used when the macro is used. (We can't actually detect
whether we're using the comma pasting extension until the macro is
actually used, but we can detecting constructs which will be comma
pasting if the varargs argument is elided.)
David Blaikie [Wed, 14 Nov 2012 01:52:05 +0000 (01:52 +0000)]
Provide the correct mangling and linkage for certain unnamed nested classes.
This corrects the mangling and linkage of classes (& their member functions) in
cases like this:
struct foo {
struct {
void func() { ... }
} x;
};
we were accidentally giving this nested unnamed struct 'no' linkage where it
should've had the linkage of the outer class. The mangling was incorrecty too,
mangling as TU-wide unnamed type mangling of $_X rather than class-scoped
mangling of UtX_.
This also fixes -Wunused-member-function which would incorrectly diagnose
'func' as unused due to it having no linkage & thus appearing to be TU-local
when in fact it might be correctly used in another TU.
Similar mangling should be applied to function local classes in similar cases
but I've deferred that for a subsequent patch.
Review/discussion by Richard Smith, John McCall, & especially Eli Friedman.
Richard Smith [Wed, 14 Nov 2012 00:50:40 +0000 (00:50 +0000)]
Remove another questionable use of hasTrivial*. The relevant thing for this
test was whether the /selected/ operator= was trivial, not whether the class
had any trivial (or any non-trivial) operator=s.
John McCall [Wed, 14 Nov 2012 00:49:39 +0000 (00:49 +0000)]
Accept and pass arguments to __unknown_anytype in argument
positions of Objective-C methods.
It is possible to recover a lot of type information about
Objective-C methods from the reflective metadata for their
implementations. This information is not rich when it
comes to struct types, however, and it is not possible to
produce a type in the debugger's round-tripped AST which
will really do anything useful during type-checking.
Therefore we allow __unknown_anytype in these positions,
which essentially disables type-checking for that argument.
We infer the parameter type to be the unqualified type of
the argument expression unless that expression is an
explicit cast, in which case it becomes the type-as-written
of that cast.
Nico Weber [Tue, 13 Nov 2012 22:09:44 +0000 (22:09 +0000)]
[ms] Make mangleIntegerLiteral less aware of exact type of the literal.
Integer literal mangling does not actually depend on exact type of the literal.
This will simplify calling mangleIntegerLiteral when literal type is not known,
for example, when sizes or offsets are mangled as integer literals.
Also, call mangleNumber instead of directly printing mangled values of 0/1, to
avoid this knowledge from being in multiple places.
Kaelyn Uhrain [Tue, 13 Nov 2012 21:23:31 +0000 (21:23 +0000)]
For classes that have the warn_unused_result attribute, don't apply the
attribute to the class' methods even when they return an instance of the
class (e.g. assignment operators).
Benjamin Kramer [Tue, 13 Nov 2012 15:32:35 +0000 (15:32 +0000)]
This patch makes the behavior of clang consistent with the behavior of gcc 4.6 in cases where both -fPIC and -fPIE is used.
- Separately check if -fPIE was specified in the command line and define both __PIC__ and __PIE__ when -fPIE is used. We need to check this separately because -fPIE will infer -fPIC even if its not explicitly used.
- Fixed existing tests.
- Added new tests for cases where both -fPIC and -fPIE is used.
Author: Tareq A. Siraj <tareq.a.siraj@intel.com> Fixes: PR13221
Review: http://llvm-reviews.chandlerc.com/D94
Anna Zaks [Tue, 13 Nov 2012 03:18:01 +0000 (03:18 +0000)]
Fix a Malloc Checker FP by tracking return values from initWithCharacter
and other functions.
When these functions return null, the pointer is not freed by
them/ownership is not transfered. So we should allow the user to free
the pointer by calling another function when the return value is NULL.
Bill Wendling [Tue, 13 Nov 2012 02:31:58 +0000 (02:31 +0000)]
Use the 'count' attribute instead of the 'upper_bound' attribute.
If we have a type 'int a[1]' and a type 'int b[0]', the generated DWARF is the
same for both of them because we use the 'upper_bound' attribute. Instead use
the 'count' attrbute, which gives the correct number of elements in the array.
<rdar://problem/12566646>
Bill Wendling [Tue, 13 Nov 2012 00:54:24 +0000 (00:54 +0000)]
Disable accelerator tables when compiling with LTO.
LTO doesn't generate correct accelerator tables. This is due to the general lack
correct of debug info for LTO. Disable it when using LTO.
<rdar://problem/12401423>
Richard Smith [Tue, 13 Nov 2012 00:54:12 +0000 (00:54 +0000)]
Fix some wrong-code bugs in implicitly-defined assignment operators:
- In C++11, perform overload resolution over all assignment operators, rather than just looking for copy/move assignment operators.
- Clean up after temporaries produced by operator= immediately, rather than accumulating them until the end of the function.
Ted Kremenek [Tue, 13 Nov 2012 00:12:13 +0000 (00:12 +0000)]
Fix bad CFG construction bug when handling C++ 'try' statements.
This code assigned the last created CFGBlock* to the variable 'Block',
which is a scratch variable which is null'ed out after a block is
completed. By assigning the last created block to 'Block', we start
editing a completed block, inserting CFGStmts that should be in
another block. This was the case with 'try'. The test case that
showed this had a while loop inside a 'try', and the logic before
the while loop was being included as part of the "condition block"
for the loop. This showed up as a bogus dead store, but could
have lots of implications.
Turns out this bug was replicated a few times within CFG.cpp, so
I went and fixed up those as well.
Nick Lewycky [Tue, 13 Nov 2012 00:08:34 +0000 (00:08 +0000)]
When filtering the list of associated namespaces so that we don't suggest people
add functions to namespace 'std', also filter out namespaces with '__' anywhere
in the name.
Kaelyn Uhrain [Mon, 12 Nov 2012 23:48:05 +0000 (23:48 +0000)]
Enable C++11 attribute syntax for warn_unused_result and allow it to be
applied to CXXRecordDecls, where functions with that return type will
inherit the warn_unused_result attribute.
Also includes a tiny fix (with no discernable behavior change for
existing code) to re-sync AttributeDeclKind enum and
err_attribute_wrong_decl_type with warn_attribute_wrong_decl_type since
the enum is used with both diagnostic messages to chose the correct
description.
Anna Zaks [Mon, 12 Nov 2012 23:40:29 +0000 (23:40 +0000)]
[analyzer] Follow up to r167762 - precisely determine the adjustment
conditions.
The adjustment is needed only in case of dynamic dispatch performed by
the analyzer - when the runtime declaration is different from the static
one.
Document this explicitly in the code (by adding a helper). Also, use
canonical Decls to avoid matching against the case where the definition
is different from found declaration.
This fix suppresses the testcase I added in r167762, so add another
testcase to make sure we do test commit r167762.
Richard Smith [Mon, 12 Nov 2012 23:33:00 +0000 (23:33 +0000)]
Factor duplicated implicit memcpy call generation code out of copy/move
assignment generation. This incidentally avoids reusing the same Expr* across
multiple statements in the same object; that was generating slightly broken
ASTs, but I couldn't trigger any observable bad behavior, so no test.
Anna Zaks [Mon, 12 Nov 2012 22:06:24 +0000 (22:06 +0000)]
[analyzer] Fix a regression (from r 165079): compare canonical types.
Suppresses a leak false positive (radar://12663777).
In addition, we'll need to rewrite the adjustReturnValue() method not to
return UnknownVal by default, but rather assert in cases we cannot
handle. To make it possible, we need to correctly handle some of the
edge cases we already know about.
Richard Smith [Mon, 12 Nov 2012 21:38:00 +0000 (21:38 +0000)]
Remove abuse of hasTrivial*, and fix miscompile wherein global arrays with
internal linkage, no uses, trivial construction, and nontrivial destruction
were not emitted.
Chad Rosier [Mon, 12 Nov 2012 21:32:24 +0000 (21:32 +0000)]
If we encounter a fatal error, exit with status 70. For BSD systems this is
defined as an internal software error. This notifies the driver to report
diagnostics information.
rdar://11951540
Ted Kremenek [Mon, 12 Nov 2012 21:20:48 +0000 (21:20 +0000)]
Per discussion on cfe-dev, re-enable suppression of -Wimplicit-fallthrough on C, but also include dialects of C++ earlier than C++11.
There was enough consensus that we *can* get a good language solution
to have an annotation outside of C++11, and without this annotation
this warning doesn't quite mean's completeness criteria for this
kind of warning. For now, restrict this warning to C++11 (where an
annotation exists), and make this the behavior for the LLVM 3.2 release.
Afterwards, we will hammer out a language solution that we are all
happy with.
Bill Wendling [Mon, 12 Nov 2012 06:42:51 +0000 (06:42 +0000)]
Check that the input size is correct for the given constraint.
The 'a', 'c', and 'd' constraints on i386 mean a 32-bit register. We cannot
place a 64-bit value into the 32-bit register. Error out instead of causing the
compiler to spew general badness.
<rdar://problem/12415959>
Rafael Espindola [Mon, 12 Nov 2012 04:10:23 +0000 (04:10 +0000)]
In Sema::MergeVarDecl we handle merging of storage classes and visibility
attributes. In cases where the merged declaration is fully equivalent to the
two original ones, some of the code was getLVForDecl was duplicated.
Cases that are still handled in getLVForDecl are things like
__private_extern__ int N;
int N;
For which we cannot produce a single merged decl with all the information.
Daniel Jasper [Sun, 11 Nov 2012 22:14:55 +0000 (22:14 +0000)]
Fix binding of nodes in case of forEach..() matchers.
When recursively visiting the generated matches, the aggregated bindings need
to be copied during the recursion. Otherwise, we they might not be properly
overwritten (which is shown by the test), or there might be bound nodes present
that were bound on a different matching branch.
Douglas Gregor [Sat, 10 Nov 2012 07:24:09 +0000 (07:24 +0000)]
Rework my implementation of circular-reference finding to not use
CXXRecordDecl::forallBases, which does *not* do what I need. Fixes the
failure introduced in r167651.
Michael Liao [Sat, 10 Nov 2012 05:17:46 +0000 (05:17 +0000)]
Add clang support of RTM from TSX
- New options '-mrtm'/'-mno-rtm' are added to enable/disable RTM feature
- Builtin macro '__RTM__' is defined if RTM feature is enabled
- RTM intrinsic header is added and introduces 3 new intrinsics, namely
'_xbegin', '_xend', and '_xabort'.
- 3 new builtins are added to keep compatible with gcc, namely
'__builtin_ia32_xbegin', '__builtin_ia32_xend', and '__builtin_ia32_xabort'.
- Test cases for pre-defined macro and new intrinsic codegen are added.
Ted Kremenek [Sat, 10 Nov 2012 01:47:40 +0000 (01:47 +0000)]
Disable -Wimplicit-fallthrough when not using C++.
The rationale is that there is no good workflow to silence the warning
for specific cases, other than using pragmas. This is because the
attribute to decorate an explicit fall through is only available
in C++11.
By that argument, this should probably also be disabled unless one
is using C++11, but apparently there is an explicit test case for
this warning when using C++98. This will require further discussion
on cfe-commits.
Jordan Rose [Sat, 10 Nov 2012 01:40:08 +0000 (01:40 +0000)]
[analyzer] When invalidating symbolic offset regions, take fields into account.
Previously, RegionStore was being VERY conservative in saying that because
p[i].x and p[i].y have a concrete base region of 'p', they might overlap.
Now, we check the chain of fields back up to the base object and check if
they match.
This only kicks in when dealing with symbolic offset regions because
RegionStore's "base+offset" representation of concrete offset regions loses
all information about fields. In cases where all offsets are concrete
(s.x and s.y), RegionStore will already do the right thing, but mixing
concrete and symbolic offsets can cause bindings to be invalidated that
are known to not overlap (e.g. p[0].x and p[i].y).
This additional refinement is tracked by <rdar://problem/12676180>.
Douglas Gregor [Sat, 10 Nov 2012 01:18:17 +0000 (01:18 +0000)]
Diagnostic circular inheritance involving dependent base classes. We
would have diagnosed this at instantiation time anyway, if only we
didn't hang on all of these test cases. Fixes <rdar://problem/12629723>
David Blaikie [Sat, 10 Nov 2012 01:04:23 +0000 (01:04 +0000)]
PR14296: function parameter name collisions in function try/catch
C++11 3.3.3/2 "A parameter name shall not be redeclared in the outermost block
of the function definition nor in the outermost block of any handler associated
with a function-try-block."
It's not totally clear to me whether the "FIXME" case is covered by this, but
Richard Smith thinks it probably should be. It's just a bit more involved to
fix that case.
Richard Smith [Fri, 9 Nov 2012 23:03:14 +0000 (23:03 +0000)]
PR13788: Don't perform checks on the initializer of a dependently-typed
variable. Previously we didn't notice the type was dependent if the only
dependence came from an array bound.
Richard Smith [Fri, 9 Nov 2012 22:36:44 +0000 (22:36 +0000)]
PR14303: Add a NoDriverOption flag to those options which are not accepted by
the driver (the options defined in CC1Options.td) and exclude their help from
"clang --help".
Add a FileCharacteristic parameter to SourceManager::createFileIDForMemBuffer
for completeness and use it in CompilerInstance::InitializeSourceManager if
the input is a memory buffer.
Andy Gibbs [Fri, 9 Nov 2012 13:24:30 +0000 (13:24 +0000)]
Improved support for removing the comma preceding __VA_ARGS__ where __VA_ARGS__
is empty in a variadic macro expansion. This fixes a divergence in support for
the ", ## __VA_ARGS__" GCC extension which differed in behaviour when in strict
C99 mode (note: there is no change in behaviour has been made in the gnu99 mode
that clang uses by default). In addition, there is improved support for the
Microsoft alternative extension ", __VA_ARGS__".
projects / clang / log