John McCall [Tue, 17 Nov 2009 02:14:36 +0000 (02:14 +0000)]
Carry lookup configuration throughout lookup on the LookupResult. Give
LookupResult RAII powers to diagnose ambiguity in the results. Other diagnostics
(e.g. access control and deprecation) will be moved to automatically trigger
during lookup as part of this same mechanism.
This abstraction makes it much easier to encapsulate aliasing declarations
(e.g. using declarations) inside the lookup system: eventually, lookup will
just produce the aliases in the LookupResult, and the standard access methods
will naturally strip the aliases off.
Douglas Gregor [Tue, 17 Nov 2009 00:55:50 +0000 (00:55 +0000)]
When querying type qualifiers on QualType via one of the "non-local"
interfaces (which are used throughout the front end), combine the
qualifiers on the QualType instance with the qualifiers on the
canonical type to produce the set of qualifiers that, semantically,
apply to that type. This should design away a large category of
"qualifier-hidden-behind-a-typedef" buts like we saw in PR5383.
Performance-wise, this caused a regression of ~0.5% on Cocoa.h, but
it's totally worth it. We may actually be able to get a little more
performance back by using CanQualType more often.
Douglas Gregor [Mon, 16 Nov 2009 21:35:15 +0000 (21:35 +0000)]
First part of changes to eliminate problems with cv-qualifiers and
sugared types. The basic problem is that our qualifier accessors
(getQualifiers, getCVRQualifiers, isConstQualified, etc.) only look at
the current QualType and not at any qualifiers that come from sugared
types, meaning that we won't see these qualifiers through, e.g.,
typedefs:
typedef const int CInt;
typedef CInt Self;
Self.isConstQualified() currently returns false!
Various bugs (e.g., PR5383) have cropped up all over the front end due
to such problems. I'm addressing this problem by splitting each
qualifier accessor into two versions:
- the "local" version only returns qualifiers on this particular
QualType instance
- the "normal" version that will eventually combine qualifiers from this
QualType instance with the qualifiers on the canonical type to
produce the full set of qualifiers.
This commit adds the local versions and switches a few callers from
the "normal" version (e.g., isConstQualified) over to the "local"
version (e.g., isLocalConstQualified) when that is the right thing to
do, e.g., because we're printing or serializing the qualifiers. Also,
switch a bunch of
Sebastian Redl [Mon, 16 Nov 2009 21:03:45 +0000 (21:03 +0000)]
Repair broken FindCompositePointerType. Correct early termination condition. Get CVR qualifiers from canonical types. Traverse collected qualifiers in reverse order on rebuilding the pointer, so that we don't swap inner and outer qualifiers. That last one fixes PR5509.
Handle case of missing '@end' in implementation context
gracefully, on par with gcc, by: Issuing a warning,
doing final sematinc check of its definitions and generating
its meta-data.
Ken Dyck [Mon, 16 Nov 2009 16:36:33 +0000 (16:36 +0000)]
Parameterize the constant-generating macros in stdint.h with new built-in
__INTn_C_SUFFIX__ macros that are defined for types with corresponding
constant suffixes (i.e. long and long long).
Eli Friedman [Mon, 16 Nov 2009 05:55:46 +0000 (05:55 +0000)]
Make GetAddrOfConstantStringFromLiteral return a constant of the correct type.
This doesn't have any visible effects at the moment because normally the
implicit cast code forces the type to the expected type.
Zhongxing Xu [Mon, 16 Nov 2009 04:49:44 +0000 (04:49 +0000)]
* Do the same thing to the basicstore as in r84163.
* Add a load type to GRExprEngine::EvalLoad().
* When retrieve from 'theValue' of OSAtomic funcitions, use the type of the
region instead of the argument expression as the load type.
* Then we can convert CastRetrievedSVal to a pure assertion. In the future
we can let all Retrieve() methods simply return SVal.
Anders Carlsson [Sun, 15 Nov 2009 16:43:15 +0000 (16:43 +0000)]
If we find a deallocation function in the class scope, but it is a placement function we should not look for a deallocation function in the global scope.
Douglas Gregor [Sun, 15 Nov 2009 09:20:52 +0000 (09:20 +0000)]
When performing a static downcast as part of a static_cast, make sure
that we're dealing with canonical types like the documentation say
(yay, CanQualType). Alas, this is another instance where using
getQualifiers() on a non-canonical QualType got us in trouble.
Good news: with this fix, Clang can now parse all of its own headers!
Douglas Gregor [Sun, 15 Nov 2009 08:11:13 +0000 (08:11 +0000)]
When adding the underlying declaration of a decl to a lookup-results
set, expand overloaded function declarations. Long-term, this should
actually be done by the name-lookup code rather than here, but this
part of the code (involving using declarations) is getting a makeover
now and the test-case is useful.
Douglas Gregor [Sun, 15 Nov 2009 07:48:03 +0000 (07:48 +0000)]
When looking for operator() to type-check a call to an object of class
type, use full qualified name lookup rather than the poking the
declaration context directly. This makes sure that we see operator()'s
in superclasses. Also, move the complete-type check before this name
lookup.
Daniel Dunbar [Sun, 15 Nov 2009 00:27:43 +0000 (00:27 +0000)]
Add a trivial example plugin, which prints the names of the top-level decls.
- The build scriptage is about twice as long as the code, which is nice. :)
Daniel Dunbar [Sun, 15 Nov 2009 00:12:04 +0000 (00:12 +0000)]
Add pluggable action support to clang-cc, via -plugin command line option.
- Expects the plugin has been loaded with -load.
- Using this may require disabling TOOL_NO_EXPORTS in the clang-cc Makefile, this breaks the llvm::Registry way of working (static constructors are bad, kids). This should be replaced with a "real" plugin model that has explicit plugin interfaces.
Daniel Dunbar [Sat, 14 Nov 2009 22:04:54 +0000 (22:04 +0000)]
Add clang -mcpu=native support, patch by Roman Divacky, varioustweaks by me.
- We still need support for detecting the target features, since the name
doesn't actually do a good job of decribing what the CPU supports (for LLVM).
Sebastian Redl [Sat, 14 Nov 2009 21:15:49 +0000 (21:15 +0000)]
- Have TryStaticImplicitCast set the cast kind to NoOp when binding a reference. CheckReferenceInit already inserts implicit casts to the necessary types. This fixes an assertion in CodeGen for some casts and brings a fix for PR5453 close, if I understand that bug correctly.
- Also, perform calculated implicit cast sequences if they're determined to work. This finally diagnoses static_cast to ambiguous or implicit bases and fixes two long-standing fixmes in the test case. For the C-style cast, this requires propagating the access check suppression pretty deep into other functions.
- Pass the expressions for TryStaticCast and TryStaticImplicitCast by reference. This should lead to a better AST being emitted for such casts, and also fixes a memory leak, because CheckReferenceInit and PerformImplicitConversion wrap the node passed to them. These wrappers were previously lost.
Benjamin Kramer [Sat, 14 Nov 2009 12:08:24 +0000 (12:08 +0000)]
Change *BugReport constructors to take StringRefs.
- Eliminates many calls to std::string.c_str()
- Fixes an invalid read in ReturnStackAddressChecker due to an unsafe call to
StringRef.data() which doesn't guarantee null-termination.
projects / clang / log