a7c44113bc
Changed behavior of how we handle "NULL" summaries: just call
...
GRSimpleVals::EvalCal(), and don't change reference counts.
Remove "getDoNothingSummary()", as a NULL summary does the same thing.
Added temporary hack for the "Get" rule for objects that return a pointer type:
treat them as non-owned CF objects.
Added test case to detect the release of a non-owned object.
llvm-svn: 49555
2008-04-11 20:11:19 +00:00
4b77209694
Fixed some logic errors in the CF ref count checker; we now can detect simple
...
use-after-release errors. Added test case.
llvm-svn: 49509
2008-04-10 23:44:06 +00:00
4a78c3ae11
Refactored all logic to run the GRSimpleVals and CFRef checker into a common
...
code path in the clang driver.
Renamed options --grsimple to -checker-simple and -check-cfref to -checker-cfref.
llvm-svn: 49500
2008-04-10 22:16:52 +00:00
0e91b41902
typedef void T;
...
void f(T);
is only invalid in C++ mode, not C89 mode.
llvm-svn: 49460
2008-04-10 02:26:16 +00:00
58258246ec
Several improvements from Doug Gregor related to default
...
argument handling. I'll fix up the c89 (void) thing next.
llvm-svn: 49459
2008-04-10 02:22:51 +00:00
aa9c7aed0f
Add support for C++ default arguments, and rework Parse-Sema
...
interaction for function parameters, fixing PR2046.
Patch by Doug Gregor!
llvm-svn: 49369
2008-04-08 04:40:51 +00:00
3c7961597f
merge compatibility testing of qualified/unqualified interfaces together
...
and fix a bug.
llvm-svn: 49322
2008-04-07 06:38:24 +00:00
c9062d01d7
MyOtherClass<MyProtocol>* is compatible with MyClass*
...
if MyClass is a superclass of MyOtherClass, there is no need for
an exact interface match.
llvm-svn: 49320
2008-04-07 06:06:56 +00:00
56a7bf5a7e
remove the old non-parallel test system. 'make' in clang/test now runs
...
the tests in parallel.
llvm-svn: 49290
2008-04-06 22:32:01 +00:00
285a3e47be
reject 'int test(x, x) int x; {}'
...
llvm-svn: 49271
2008-04-06 06:50:56 +00:00
67b450cb48
reject 'typedef int y; int test(x, y)'.
...
llvm-svn: 49270
2008-04-06 06:47:48 +00:00
9d51f2b9d4
Fix handling of implicit int, resolving PR2012 and reverting (and
...
subsuming) my patch for PR1999.
llvm-svn: 49251
2008-04-05 06:32:51 +00:00
94fc8063b4
Step #1 to fixing PR2012: c89 allows declspecs to be completely
...
missing from function definitions only. If we see a function
definiton with missing declspecs, just fudge in an int.
llvm-svn: 49250
2008-04-05 05:52:15 +00:00
579a05d767
Fix PR2049, updating Eli's patch that fixes to mainline. This produces incorrect
...
code, but the codegen doesn't crash. I'll file a bugzilla for the AST being wrong.
llvm-svn: 49226
2008-04-04 18:42:16 +00:00
f3bc75afcf
Since isComplexType() no longer returns true for _Complex integers, the code
...
generator needs to call isAnyComplexType(). This fixes PR1960.
llvm-svn: 49220
2008-04-04 16:54:41 +00:00
3913ea4bfb
Add explicit support for diagnosing implicit function decls.
...
Without this, the diagnostic is very confusing. The diag is now consistent with gcc as well.
llvm-svn: 49214
2008-04-04 14:32:09 +00:00
055c94f2f8
PR1963: Address of function is a constant expression
...
llvm-svn: 49212
2008-04-04 09:45:30 +00:00
e799eb5daa
Fix PR2081 (problems codegen'ing some recursive structures) patch
...
by Lauro Venancio!
llvm-svn: 49144
2008-04-03 05:50:42 +00:00
778927532d
Change ObjCInterfaceDecl to inherit from NamedDecl (not TypeDecl). While ObjCInterfaceDecl is arguably a TypeDecl, it isn't a ScopedDecl. Since TypeDecl's are scoped, it makes sense to simply treat them as NamedDecl's. I could have fiddled a bit more with the hierarchy (in terms of creating a non-scoped TypeDecl), however this probably isn't worth the effort.
...
I also finished unifying access to scope decl change by converting Sema::getObjCInterfaceDecl() to use Sema::LookupDecl(). This is much cleaner now:-)
llvm-svn: 49107
2008-04-02 18:30:49 +00:00
6f6d226f63
improve error to be something end users will actually understand :)
...
llvm-svn: 49097
2008-04-02 17:15:17 +00:00
018ba60f8e
Added path-sensitive null dereference test case.
...
llvm-svn: 49095
2008-04-02 16:54:39 +00:00
daaa8ba77e
1) Enforce C99 6.7.3p2: "Types other than pointer types derived from
...
object or incomplete types shall not be restrict-qualified."
2) Warn about qualifiers on function types: C99 6.7.3p8: "If the
specification of a function type includes any type qualifiers, the
behavior is undefined."
3) Implement restrict on C++ references.
4) fix some locations for various C++ reference diagnostics.
llvm-svn: 49081
2008-04-02 06:50:17 +00:00
a21ad8058a
Fix several bugs in array -> pointer decomposition.
...
First, we got several CVR propagation cases wrong, which Eli pointed
out in PR2039.
Second, we didn't propagate address space qualifiers correctly, leading
to incorrect lowering of code in CodeGen/address-space.c.
Third, we didn't uniformly propagate the specifier in the array to the
pointer ("int[restrict 4]" -> "int *restrict").
This adds an ASTContext::getArrayDecayedType member that handles the
non-trivial logic for this seemingly simple operation.
llvm-svn: 49078
2008-04-02 05:18:44 +00:00
81cb9e8637
Fix PR2017 and silence some bogus errors.
...
llvm-svn: 49068
2008-04-02 01:05:10 +00:00
257520b1b4
Fairly large "cleaup" related to changing ObjCCompatibleAliasDecl superclass (to inherit from NamedDecl, instead of ScopedDecl).
...
- Added a DenseMap to associate an IdentifierInfo with the ObjCCompatibleAliasDecl.
- Renamed LookupScopedDecl->LookupDecl and changed it's return type to Decl. Also added lookup for ObjCCompatibleAliasDecl's.
- Removed Sema::LookupInterfaceDecl(). Converted clients to used LookupDecl().
- Some minor indentation changes.
Will deal with ObjCInterfaceDecl and getObjCInterfaceDecl() in a separate commit...
llvm-svn: 49058
2008-04-01 23:04:06 +00:00
f646774f32
Added path-sensitive check for return statements that return the address
...
of a stack variable. This is the path-sensitive version of a check that
is already done during semantic analysis.
llvm-svn: 48980
2008-03-31 15:02:58 +00:00
59a2594f3f
rename Decl::CompatibleAlias -> ObjCCompatibleAlias.
...
Fix objc ivar lookup. Ivar lookup should occur between lookup
of method-local values and lookup of globals. Emulate this with
some logic in the handling of Sema::ActOnIdentifierExpr.
Two todo's left:
1) sema shouldn't turn a bare reference to an ivar into "self->ivar"
in the AST. This is a hack.
2) The new ScopedDecl::isDefinedOutsideFunctionOrMethod method does
not correctly handle typedefs and enum constants yet.
llvm-svn: 48972
2008-03-31 00:36:02 +00:00
f9e7c90129
Only have the rewriter produce a file when there are no errors.
...
llvm-svn: 48926
2008-03-28 22:26:09 +00:00
0de4199ca0
Make sure Sema::ActOnClassMessage() correctly diagnoses "super".
...
llvm-svn: 48924
2008-03-28 21:37:05 +00:00
276278e5d2
Expanded NSString checking to check for nil for a few more methods.
...
llvm-svn: 48898
2008-03-27 22:05:32 +00:00
2e4e7ccb22
Add line SourceLocation to NSString checks.
...
Added test case to test warning about passing 'nil' to NSString's compare: method.
llvm-svn: 48896
2008-03-27 21:23:57 +00:00
c10c2e339f
Fix rewriter tests to work with new output file logic. This changes forces the output to standard out (rather than generate a .cpp file, which is the new default for the rewriter).
...
llvm-svn: 48847
2008-03-26 22:28:40 +00:00
9dd6ddec63
Added test case for static analysis to warn about passing uninitialized values as arguments (not the receiver) in message expressions.
...
llvm-svn: 48845
2008-03-26 22:00:11 +00:00
a1d9ed137a
Added test case for static analyzer to detect uses of uninitialized pointers as receivers in Objective-C message expressions.
...
Added test case directory "Analysis-Apple" which is only run on Apple (darwin) machines.
llvm-svn: 48844
2008-03-26 21:51:44 +00:00
623207d603
allow the format attribute to be specified in function pointer prototypes
...
# this is my first commit here, so please be gentle :)
llvm-svn: 48807
2008-03-25 23:01:48 +00:00
c506c78117
Extend QualType::getAddressSpace to do the right thing for array types, and in
...
the future, RecordTypes.
llvm-svn: 48784
2008-03-25 18:36:32 +00:00
ac92a13c86
The newest version of valgrind, 3.3.0, no longer appends PIDs to log filenames
...
by default. Adding a %p works with this new version and is backwards
compatible, though it looks a little funny.
llvm-svn: 48698
2008-03-23 03:42:18 +00:00
fb475ec504
Changed merge operation for uninitialized values analysis to "intersect" (previous union).
...
The effect is that if a variable is uninitialized along a branch (but initialized along another), at merge points it is considered uninitialized. Previously we had the opposite behavior. The new behavior is more conservative, and more in line with gcc's behavior.
llvm-svn: 48689
2008-03-22 20:11:00 +00:00
97d496c182
Teach the rewriter how to respect the -o option.
...
llvm-svn: 48669
2008-03-22 00:08:40 +00:00
4f55259237
fix this test.
...
llvm-svn: 48668
2008-03-22 00:07:50 +00:00
f4303edfa3
Only invoke the preprocessor; it's the only thing being tested. Also, check
...
the output to make sure that the macros and comments are being treated as such.
llvm-svn: 48638
2008-03-21 08:05:55 +00:00
558447416d
Use the -E option when invoking clang. Only the preprocessor is
...
required to perform this test.
llvm-svn: 48636
2008-03-21 07:02:04 +00:00
56c563ad3f
Add ability to run clang tests under valgrind. Patch by Sam Bishop.
...
llvm-svn: 48621
2008-03-20 22:50:54 +00:00
f5eb286707
reduce chattyness of makefiles
...
llvm-svn: 48597
2008-03-20 14:28:22 +00:00
4fa19e4047
scan testdirs only once, patch by Mike Stump, thanks!
...
llvm-svn: 48589
2008-03-20 08:09:10 +00:00
c6a8352820
Convert tabs to spaces.
...
llvm-svn: 48539
2008-03-19 07:31:52 +00:00
831c7c52c7
add 'clean' targets, make sure that 'make test' works after 'make clean'
...
llvm-svn: 48488
2008-03-18 07:03:00 +00:00
8bdc7b057d
let the %t magic happen, thanks Chris!
...
llvm-svn: 48487
2008-03-18 06:42:43 +00:00
1ac5215798
finishing touches: teminate the dots with a newline
...
llvm-svn: 48462
2008-03-17 18:36:14 +00:00
5434053f6d
Implemented verbose mode.
...
make test VERBOSE=1
should now produce the same output like
the Makefile normally does.
Without VERBOSE only the names of failed tests
are printed.
I am pretty comfortable with this version, and
I think it is time to switch over to Makefile.parallel
to be the default.
What do you think? Did I forget some piece of feedback?
llvm-svn: 48458
2008-03-17 16:29:58 +00:00
Ted Kremenek