Daniel Dunbar [Tue, 13 Jan 2009 21:07:43 +0000 (21:07 +0000)]
ccc: Darwin: Implement some important general argument translations
for the Darwin tool chain.
- Ideally we would localize these to tool specific argument
processing but for now this matches gcc closely.
Douglas Gregor [Tue, 13 Jan 2009 19:47:12 +0000 (19:47 +0000)]
Turn off some Destroy calls that are currenly causing double-destruction of ScopedDecls. We will re-enable this later, when we have time to fully solve the ownership issue.
DeclContext::KindTrait was not meant to be used outside of DeclContext::CastTo (causes compilation error on MSVC).
Add DeclContext::getDeclKind() and use that instead of DeclContext::KindTrait.
Daniel Dunbar [Tue, 13 Jan 2009 05:54:38 +0000 (05:54 +0000)]
ccc: Add option groups.
- Simple mechanism for group together sets of options so the driver
can efficiently deal with them as a group (i.e., for forwarding -i*
to cc1).
- Use to finish off the major missing pieces of Darwin/CC1 support.
Zhongxing Xu [Tue, 13 Jan 2009 01:49:57 +0000 (01:49 +0000)]
Add KillStruct to region store.
- put the killed region in the kill set.
- set its default value to unknown.
- removes all bindings for its subregions.
Zhongxing Xu [Tue, 13 Jan 2009 01:29:24 +0000 (01:29 +0000)]
Add an initial framework of a DeclContextPrinter. It can print DeclContext and
its Decls in indented format. An Example:
$ cat t.cpp
class A {
int a;
void f();
};
void A::f() {
a = 3;
}
$ clang -print-decl-contexts t.cpp
[translation unit] 0x9754d7c
<typedef> __builtin_va_list
[class] A 0x9753310
<class> A 0x975ce20
<field> a
<c++ method> f
<c++ ctor> A
<c++ ctor> A
<c++ method> operator=
<c++ dtor> ~A
[c++ method] f [[0x9753310]]
Some comments: '<>' indicates a declaration, '[]' indicates a definition, '[[
]]' displays the semantic DeclContext which is different from the lexical
DeclContext. The symbols printed can definitely be changed in the future.
Ted Kremenek [Mon, 12 Jan 2009 21:45:02 +0000 (21:45 +0000)]
retain/release checker:
- Refactor a bunch of logic in the retain/release checker, making it more
condense and easier to read.
- Add support for "Create" methods in the DiskArbitration framework
retain/release tests:
- Rename CFDate.m to retain-release.m, and move test from CFString.c to
retain-release.m
- Add DiskArbitration framework tests cases.
- Add/refine and few more retain/release GC test cases.
Patch to supprt case of readonly property being
assigned to when it has user declared setter method
defined in the class implementation (but no declaration in
the class itself).
There is still some work to do in Objective-C++, but this requires
that each of the Objective-C entities (interfaces, implementations,
etc.) to be introduced into the context stack with
PushDeclContext/PopDeclContext. This will be a separate fix, later.
Nuno Lopes [Sun, 11 Jan 2009 23:22:37 +0000 (23:22 +0000)]
make ScalarExprEmitter::EmitCompare() emit the expression with the correct type instead of always zext it to an int
this fixes codegen of simple exprs in C++ like 'if (x != 0)'
Daniel Dunbar [Sun, 11 Jan 2009 23:13:15 +0000 (23:13 +0000)]
ccc: Add fairly complete argument translation for Darwin link step.
- Some things are still hardcoded, and macosx-version-min comparison
isn't implemented, but otherwise this very closely matches gcc.
- The one exception is that arguments (like -framework or -Wl,) which are
treated as linker inputs instead of options are not being
forwarded yet.
Chris Lattner [Sun, 11 Jan 2009 19:48:19 +0000 (19:48 +0000)]
make paste avoidance avoid pasting digraphs and :: only when digraphs or c++ is enabled
respectively. Inspired by a patch by Dan Villiom Podlaski Christiansen.
Chris Lattner [Sun, 11 Jan 2009 19:28:34 +0000 (19:28 +0000)]
this patch "adds support for specifying multiple dependancy targets using ‘-MT’. Using this patch, I was able to use ‘ccc’ as C compiler for the few C files in LLVM. In my brief testing, the output remains identical to that of GCC."
Sebastian Redl [Sun, 11 Jan 2009 00:38:46 +0000 (00:38 +0000)]
Convert some more actions to smart pointers.
No performance regression in my basic test.
Also fixed a type error in ActOnFinishSwitchStmt's arguments (body is a stmt).
This patch fixes the code gen failures which was a fallout from
not merging protocol properties into the classes which
use those protocols. With this patch, all my exceutable
test pass again.
Daniel Dunbar [Sat, 10 Jan 2009 02:07:54 +0000 (02:07 +0000)]
ccc: Introduce ToolChains for mapping Actions to Tools which can
perform them.
- A ToolChain is a coherent set of tools use in a compilation
process. The idea is that a ToolChain holds roughly the information
(specs, search paths, etc.) that is in a single gcc binary.
- The default ToolChain is selected by the host and will generally
correspond to what the default system compiler would do. However,
this can be over-riden for a variety of purposes, for example the
by the driver driver or for testing.
Douglas Gregor [Fri, 9 Jan 2009 22:42:13 +0000 (22:42 +0000)]
When we see a reference to a struct, class, or union like "struct X"
that is neither a definition nor a forward declaration and where X has
not yet been declared as a tag, introduce a declaration
into the appropriate scope (which is likely *not* to be the current
scope). The rules for the placement of the declaration differ slightly
in C and C++, so we implement both and test the various corner
cases. This implementation isn't 100% correct due to some lingering
issues with the function prototype scope (for a function parameter
list) not being the same scope as the scope of the function
definition. Testcase is FIXME'd; this probably isn't an important issue.
Daniel Dunbar [Fri, 9 Jan 2009 22:39:43 +0000 (22:39 +0000)]
Add utils/SummarizeErrors.
- Little script for scanning a compile log and summarizing warnings,
errors, assertions, and crashes.
- Is very slow, and stack trace regexs probably only work on Darwin.
Ted Kremenek [Fri, 9 Jan 2009 22:05:30 +0000 (22:05 +0000)]
Enhance PTH 'getSpelling' caching:
- Refactor caching logic into a helper class PTHSpellingSearch
- Allow "random accesses" in the spelling cache, thus catching the remaining
cases where 'getSpelling' wasn't hitting the PTH cache
For -Eonly, PTH, Cocoa.h:
- This reduces wall time by 3% (user time unchanged, sys time reduced)
- This reduces the amount of paged source by 1112K.
The remaining 1112K still being paged in is from somewhere else
(investigating).
Douglas Gregor [Fri, 9 Jan 2009 19:42:16 +0000 (19:42 +0000)]
Replace DeclContext's vector of ScopedDecl pointers with a linked list
of ScopedDecls (using the new ScopedDecl::NextDeclInScope
pointer). Performance-wise:
- It's a net win in memory utilization, since DeclContext is now one
pointer smaller than it used to be (std::vectors are typically 3
pointers; we now use 2 pointers) and
- Parsing Cocoa.h with -fsyntax-only (with a Release-Asserts Clang)
is about 1.9% faster than before, most likely because we no longer
have the memory allocations and copying associated with the
std::vector.
I'll re-enable serialization of DeclContexts once I've sorted out the
NextDeclarator/NextDeclInScope question.
Ted Kremenek [Fri, 9 Jan 2009 18:20:21 +0000 (18:20 +0000)]
Always print out SourceManager stats with 'Stats' is true. This revealed that PTH always pulls in the source pages with -fsyntax-only (investigating further).
projects / clang / log