Chandler Carruth [Mon, 22 Aug 2011 17:24:56 +0000 (17:24 +0000)]
Fix a crash-on-valid that has been here for a very long time:
const int &x = x;
This crashed by inifinetly recursing within the lvalue evaluation
routine. I've added a (somewhat) braindead way of preventing this
recursion. If folks have better suggestions for how to avoid it I'm all
ears.
That said, we have some work to do. This doesn't trigger a single
warning for uninitialized, self-initialized or otherwise completely
wrong code. In some senses, the crash was almost better.
Boost the efficiency of SourceManager::getMacroArgExpandedLocation.
Currently getMacroArgExpandedLocation is very inefficient and for the case
of a location pointing at the main file it will end up checking almost all of
the SLocEntries. Make it faster:
-Use a map of macro argument chunks to their expanded source location. The map
is for a single source file, it's stored in the file's ContentCache and lazily
computed, like the source lines cache.
-In SLocEntry's FileInfo add an 'unsigned NumCreatedFIDs' field that keeps track
of the number of FileIDs (files and macros) that were created during preprocessing
of that particular file SLocEntry. This is useful when computing the macro argument
map in skipping included files while scanning for macro arg FileIDs that lexed from
a specific source file. Due to padding, the new field does not increase the size
of SLocEntry.
Jordy Rose [Sun, 21 Aug 2011 21:04:38 +0000 (21:04 +0000)]
[analyzer] Remove FIXME; Ted reminded me that -init is not guaranteed to return its receiver and pretending that it does won't actually buy us anything. (Comment change only.)
Jordy Rose [Sun, 21 Aug 2011 19:41:36 +0000 (19:41 +0000)]
[analyzer] Migrate return value handling from CFRefCount to ExprEngine. This seems to result in a minor performance hit, but I think that will go away again once we eliminate TransferFuncs from function calls entirely.
Chandler Carruth [Sat, 20 Aug 2011 06:30:57 +0000 (06:30 +0000)]
Remove this test. It requires C++ standard library headers, which we'd
like to avoid within the Clang test suite, it doesn't verify the output
at all so it only servers as a test that Clang doesn't crash, and
finally all it does is declare a function that returns a vector and call
it. Probably the biggest thing being tested here is Clang's parsing of
the vector template, and we have lots of good template parsing tests. We
don't need another in codegen.
Ted Kremenek [Sat, 20 Aug 2011 06:00:03 +0000 (06:00 +0000)]
Start partitioning ExprEngine.cpp into separate .cpp files that handle different parts
of the analysis (e.g., analysis of C expressions, analysis of Objective-C expressions, and so on).
Douglas Gregor [Sat, 20 Aug 2011 05:09:43 +0000 (05:09 +0000)]
Make the loading of multiple records for the same identifier (from
different modules) more robust. It already handled (simple) merges of
the set of declarations attached to that identifier, so add a test
case that shows us getting two different declarations for the same
identifier (one struct, one function) from different modules, and are
able to use both of them.
Douglas Gregor [Sat, 20 Aug 2011 04:39:52 +0000 (04:39 +0000)]
Introduce a module visitation function that starts at the top-level
modules (those that no other module depends on) and performs a search
over all of the modules, visiting a new module only when all of the
modules that depend on it have already been visited. The visitor can
abort the search for all modules that a module depends on, which
allows us to minimize the number of lookups necessary when performing
a search.
Switch identifier lookup from a linear walk over the set of modules to
this module visitation operation. The behavior is the same for simple
PCH and chained PCH, but provides the proper search order for
modules. Verified with printf debugging, since we don't have enough in
place to actually test this.
Anna Zaks [Sat, 20 Aug 2011 01:27:22 +0000 (01:27 +0000)]
Static Analyzer Diagnostics: Move the responsibility for generating the endOfPath diagnostic piece from BugReport to BugReporterVisitor. Switch CFRefCount to use visitors in order to generate the endOfPath piece.
For assigning SourceLocations to macro arg tokens, reserve a single SLocEntry
for tokens that are lexed consecutively from the same FileID, instead of creating
a SLocEntry for each token. e.g for
assert(foo == bar);
there will be a single SLocEntry for the "foo == bar" chunk and locations
for the 'foo', '==', 'bar' tokens will point inside that chunk.
For parsing SemaExpr.cpp, this reduced the number of SLocEntries by 25%.
Anna Zaks [Fri, 19 Aug 2011 22:33:38 +0000 (22:33 +0000)]
Static Analyzer Diagnostics: Kill the addVisitorCreator(callbackTy, void*) API in favor of addVisitor(BugReporterVisitor*).
1) Create a header file to expose the predefined visitors. And move the parent(BugReporterVisitor) there as well.
2) Remove the registerXXXVisitor functions - the Visitor constructors/getters can be used now to create the object. One exception is registerVarDeclsLastStore(), which registers more then one visitor, so make it static member of FindLastStoreBRVisitor.
objective-c: Bring objective-c handling of decl context
to modernity. Instead of passing down individual
context objects from parser to sema, establish decl
context in parser and have sema access current context
as needed. I still need to take of Doug's comment for
minor cleanups.
Douglas Gregor [Fri, 19 Aug 2011 02:29:29 +0000 (02:29 +0000)]
Teach ModuleManager::addModule() to check whether a particular module
has already been loaded before allocating a new Module structure. If
the module has already been loaded (uniquing based on file name), then
just return the existing module rather than trying to load it again.
This allows us to load a DAG of modules. Introduce a simple test case
that forms a diamond-shaped module graph, and illustrates that a
source file importing the bottom of the diamond can see declarations
in all four of the modules that make up the diamond.
Note that this version moves the file-opening logic into the module
manager, rather than splitting it between the module manager and the
AST reader. More importantly, it properly handles the
weird-but-possibly-useful case of loading an AST file from "-".
Anna Zaks [Fri, 19 Aug 2011 01:57:09 +0000 (01:57 +0000)]
Fix a memory leak in the analyzer - BugReports didn't get freed. Plus, remove invalid assert from the destructor which wasn't called previously due to the leak.
Anna Zaks [Thu, 18 Aug 2011 22:37:56 +0000 (22:37 +0000)]
Static Analyzer Diagnostics: Move custom diagnostic visitors from BugReporterContext to BugReport.
One API change: I added BugReporter as an additional parameter to the BugReporterVisitor::VisitNode() method to allow visitors register other visitors with the report on the fly (while processing a node). This functionality is used by NilReceiverVisitor, which registers TrackNullOrUndefValue when the receiver is null.
Chad Rosier [Thu, 18 Aug 2011 19:06:24 +0000 (19:06 +0000)]
Temporarily revert r137925 to appease buildbots. Original commit message:
Teach ModuleManager::addModule() to check whether a particular module
has already been loaded before allocating a new Module structure. If
the module has already been loaded (uniquing based on file name), then
just return the existing module rather than trying to load it again.
This allows us to load a DAG of modules. Introduce a simple test case
that forms a diamond-shaped module graph, and illustrates that a
source file importing the bottom of the diamond can see declarations
in all four of the modules that make up the diamond.
Chandler Carruth [Thu, 18 Aug 2011 09:09:59 +0000 (09:09 +0000)]
Always mark friend function declarations in class templates as
implicitly instantiable, even if we don't see a body on the friend
function declaration. The body may simply have not yet been attached.
This fixes PR10666.
There may be an alternate, preferred implementation strategy, see my
FIXME. Review would definitely be appreciated Doug. =D
Douglas Gregor [Thu, 18 Aug 2011 04:41:58 +0000 (04:41 +0000)]
Teach ModuleManager::addModule() to check whether a particular module
has already been loaded before allocating a new Module structure. If
the module has already been loaded (uniquing based on file name), then
just return the existing module rather than trying to load it again.
This allows us to load a DAG of modules. Introduce a simple test case
that forms a diamond-shaped module graph, and illustrates that a
source file importing the bottom of the diamond can see declarations
in all four of the modules that make up the diamond.
Douglas Gregor [Thu, 18 Aug 2011 04:12:04 +0000 (04:12 +0000)]
Keep track of which modules have been loaded directly (e.g., via
-import-module) vs. loaded because some other module depends on
them. As part of doing this, pass down the module that caused a module
to be loaded directly, rather than assuming that we're loading a
chain. Finally, write out all of the directly-loaded modules when
serializing an AST file (using the new IMPORTS record), so that an AST
file can depend on more than one other AST file, all of which will be
loaded when that AST file is loaded. This allows us to form and load a
tree of modules, but we can't yet load a DAG of modules.
projects / clang / log