their bindings are not removed.

llvm-svn: 98705

since RemoveDeadBinding mistakenly remove the binding to CXXThisRegion. 

llvm-svn: 98629

Tested: make CPPFLAGS=-m64 CXXFLAGS=-m64 -j8 && (cd tools/clang;make test)
llvm-svn: 98399

llvm-svn: 98176

llvm-svn: 98175

llvm-svn: 98161

llvm-svn: 98160

Really apply (unnoticed weird git-svn merge conflict in 98144): Refactor RegionStore::RemoveDeadBindings to use the same core cluster analysis algorithm as RegionStore::InvalidateRegions().

Beyond simplifying the algorithm significantly, we no longer
need to build subregion maps in RemoveDeadBindings().  This
and other changes cause a significant speedup: the time to
analyze sqlite3.c (single core) drops by 14%.

llvm-svn: 98159

cluster analysis algorithm as RegionStore::InvalidateRegions().

Beyond simplifying the algorithm significantly, we no longer
need to build subregion maps in RemoveDeadBindings().  This
and other changes cause a significant speedup: the time to
analyze sqlite3.c (single core) drops by 14%.

llvm-svn: 98144

into a 'ClusterAnalysis' parent class.  The idea is
to potentially reuse this for reworking RemoveDeadBindings.

llvm-svn: 98143

llvm-svn: 98136

with an 'unknown' memory space.

llvm-svn: 98110

of the class name.

llvm-svn: 97943

llvm-svn: 97941

information in ExplodedNode.

llvm-svn: 97785

we now may have identical states with different analysis context.

Set the right AnalysisContext in state when entering and leaving a callee.

With both of the above changes, we can pass the test case.

llvm-svn: 97724

After discussion with Zhongxing, don't force the initializer of DeclStmts to be
block-level expressions.

This led to some interesting fallout:

[UninitializedValues]

Always visit the initializer of DeclStmts (do not assume they are block-level expressions).

[BasicStore]

With initializers of DeclStmts no longer block-level expressions, this causes self-referencing initializers (e.g. 'int x = x') to no longer cause the initialized variable to be live before the DeclStmt.  While this is correct, it caused BasicStore::RemoveDeadBindings() to prune off the values of these variables from the initial store (where they are set to uninitialized).  The fix is to back-port some (and only some) of the lazy-binding logic from RegionStore to
BasicStore.  Now the default values of local variables are determined lazily as opposed
to explicitly initialized.

llvm-svn: 97591

llvm-svn: 97590

no need to store a type with SymbolRegionValue.

llvm-svn: 97437

llvm-svn: 97436

definition.

llvm-svn: 97373

Use this information to find the returned value and bind it to CallExpr in
ProcessCallExit.
And there is no need to remove dead bindings in ProcessCallExit, because
a. it would clean up the return value bound to CallExpr
b. we still would do it in the next ProcessStmt(), where we would not misclean
   up the return value.

llvm-svn: 97225

llvm-svn: 97207

Restore Zhongxing's commits r97122 r97127 r97129 r97131 which were reverted due to a Clang-on-Clang failure

llvm-svn: 97162

They were breaking clang-x86_64-darwin10-selfhost

llvm-svn: 97138

llvm-svn: 97131

llvm-svn: 97129

llvm-svn: 97127

This patch implements the CallEnter/CallExit idea of Ted.

Add two interfaces to GRSubEngine: ProcessCallEnter, ProcessCallExit.

The CallEnter program point uses caller's location context. The
CallExit program point uses callee's location context.

CallEnter is built by GRStmtNodeBuilder. CallExit is built by
GREndPathNodeBuilder.

llvm-svn: 97122

to various MacOS X functions.  The checks in BasicObjCFoundationChecks.cpp will
gradually be migrated here.

As a first check, check that when 'dispatch_once()' is passed a predicate value
that has non-local storage.

llvm-svn: 97116

llvm-svn: 97115

when it is explicitly marked as unused via __attribute__((unused)).

llvm-svn: 97104

to various unix/posix functions, e.g. 'open()'.

As a first check, check that when 'open()' is passed 'O_CREAT' that it has
a third argument.

llvm-svn: 97086

llvm-svn: 97085

llvm-svn: 97084

llvm-svn: 97083

llvm-svn: 97082

pointer (for defensive programming).  This matches the behavior with
assigning NULL to a regular pointer.  Fixes <rdar://problem/7631278>.

llvm-svn: 96985

in the static analyzer.

llvm-svn: 96539

llvm-svn: 96537