Comment at: include/clang/Tooling/Tooling.h:79
@@ -63,1 +78,3 @@
+ bool runInvocation(clang::CompilerInvocation *Invocation,
+ FileManager *Files);
Does doxygen default to having the base class' comment here?
Comment at:
Hi Tim,
I added some CodeGen test and try to use clang-format to check code style.
I guess there will be less code format problem in future:)
Please review,thanks.
2013/11/1 Tim Northover t.p.northo...@gmail.com
Hi Kevin,
The Clang patch looks good, and as far as I can see the LLVM one is
Comment at: clang-query/QueryEngine.h:28
@@ +27,3 @@
+
+class QueryEngine {
+ llvm::raw_ostream Out;
I'm a very strong proponent of public-first. The reason is that a reader of
the class usually starts at the top, and a user of a class should not care
about
Comment at: lib/Tooling/CompilationDatabase.cpp:118
@@ +117,3 @@
+
+ void run(const driver::Action *A) {
+bool setCollect = false;
I'd change the signature to have a bool Collect parameter, and then remove the
member and do:
{
bool CollectChildren =
On Thu, Oct 31, 2013 at 6:02 PM, Doug Gregor dgre...@apple.com wrote:
Sorry for the slow reply, but it seems like a net loss to generalize the
simple identifier-based module naming scheme to full string literals,
because we're making the mental model more complex (when do I use / in a
Ping.
Original Message
Subject:[PATCH] PR10837 fix
Date: Wed, 30 Oct 2013 18:56:34 +0400
From: Ivan A. Kosarev i...@ivan-labs.com
Organization: Ivan Labs
To: cfe-commits@cs.uiuc.edu
Hello,
Attached is a fix for the PR10837 (
Author: d0k
Date: Mon Nov 4 04:29:56 2013
New Revision: 193986
URL: http://llvm.org/viewvc/llvm-project?rev=193986view=rev
Log:
Make ADLResult::iterator a proper iterator.
PR17796.
Modified:
cfe/trunk/include/clang/Sema/Lookup.h
Modified: cfe/trunk/include/clang/Sema/Lookup.h
URL:
Author: d0k
Date: Mon Nov 4 04:29:51 2013
New Revision: 193985
URL: http://llvm.org/viewvc/llvm-project?rev=193985view=rev
Log:
Driver: Add support for -march=bdver3 on x86.
Modified:
cfe/trunk/lib/Basic/Targets.cpp
cfe/trunk/test/Preprocessor/predefined-arch-macros.c
Modified:
On Sun, Nov 3, 2013 at 11:40 PM, Chris Wailes chris.wai...@gmail.com wrote:
Hi delesley, aaron.ballman, silvas,
Added a test case for the fix to bug 17632 in r193751
http://llvm-reviews.chandlerc.com/D2094
Files:
test/SemaCXX/warn-consumed-analysis.cpp
Index:
Author: aaronballman
Date: Mon Nov 4 06:55:56 2013
New Revision: 193989
URL: http://llvm.org/viewvc/llvm-project?rev=193989view=rev
Log:
Attributes which accept a type as their sole argument are no longer hard coded
into the parser. Instead, they are automatically listed through tablegen.
Agreed on the longer term solution. I've committed the shorter term
solution in r193989
~Aaron
On Thu, Oct 31, 2013 at 5:39 PM, Richard Smith rich...@metafoo.co.uk wrote:
Long term, I think we should define an enum of attribute argument forms, and
TableGen a mapping from arguments to forms,
Author: aaronballman
Date: Mon Nov 4 07:02:44 2013
New Revision: 193990
URL: http://llvm.org/viewvc/llvm-project?rev=193990view=rev
Log:
Updating the makefile to reflect CMake; addendum to r193989.
Modified:
cfe/trunk/include/clang/Parse/Makefile
Modified:
Author: chapuni
Date: Mon Nov 4 07:10:53 2013
New Revision: 193991
URL: http://llvm.org/viewvc/llvm-project?rev=193991view=rev
Log:
clang/lib/Parse/CMakeLists.txt: Fixup corresponding to r193989.
ClangAttrTypeArg is required here.
Modified:
cfe/trunk/lib/Parse/CMakeLists.txt
Modified:
I like the test, but the comment should probably be Test that simple
expressions are simplified at -O0, no? :-)
On 3 November 2013 22:22, Justin Bogner m...@justinbogner.com wrote:
Rafael Espíndola rafael.espind...@gmail.com writes:
That seems like a good idea, but I need some input on where
Is Xcode the only system with a 8 byte wchar? Is that valid according
to the standard?
You are changing 5 types (SizeType, PtrDiffType, IntPtrType,
WCharType, WIntType) but only updating tests for 2. Why?
On 21 October 2013 02:17, Robert Lytton rob...@xmos.com wrote:
Hi
Is this good to go?
Author: bogner
Date: Mon Nov 4 10:13:21 2013
New Revision: 193994
URL: http://llvm.org/viewvc/llvm-project?rev=193994view=rev
Log:
CodeGen: Use EmitBranchOnBool when generating For and CXXForRange
A while ago EmitForStmt was changed to explicitly evaluate the
condition expression and create a
Author: bogner
Date: Mon Nov 4 10:13:23 2013
New Revision: 193995
URL: http://llvm.org/viewvc/llvm-project?rev=193995view=rev
Log:
CodeGen: Test that simple expressions are simplified at -O0
Added:
cfe/trunk/test/CodeGen/branch-on-bool.c
Added: cfe/trunk/test/CodeGen/branch-on-bool.c
URL:
Author: bogner
Date: Mon Nov 4 10:13:18 2013
New Revision: 193993
URL: http://llvm.org/viewvc/llvm-project?rev=193993view=rev
Log:
CodeGen: Move an initialization away from an unrelated comment
An initialization somehow found its way in between a comment and the
block of code the comment is
Rafael Espíndola rafael.espind...@gmail.com writes:
I like the test, but the comment should probably be Test that simple
expressions are simplified at -O0, no? :-)
You're right, of course. Committed r193992, r193993, and r193994.
___
cfe-commits
We normally commit the tests in the same patch that changed the code...
On 4 November 2013 08:13, Justin Bogner m...@justinbogner.com wrote:
Author: bogner
Date: Mon Nov 4 10:13:23 2013
New Revision: 193995
URL: http://llvm.org/viewvc/llvm-project?rev=193995view=rev
Log:
CodeGen: Test
Are you talking about the function name
testIsRValueRefishAndCanonicalType? The function it is testing in
Consumed.cpp is called isRValueRefish. It tests for RValue references and
LValue references that are written as RValue references.
- Chris
On Mon, Nov 4, 2013 at 7:45 AM, Aaron Ballman
While looking into http://llvm.org/bugs/show_bug.cgi?id=17782, I noticed that
there was a chunk of code in locale for dealing with floating point numbers.
This code was duplicated three times, once for float, double, and long double.
Hoist this into a template, and make the original call sites
Comment at: clang-query/QueryParser.h:18
@@ +17,3 @@
+
+Query ParseQuery(const char *Line);
+
StringRef instead of const char*
Comment at: clang-query/QueryParser.cpp:22
@@ +21,3 @@
+
+static bool IsWhitespace(char C) {
+ return C == ' ' || C
Hi Rafael,
The changes are to bring clang inline with the current llvm-gcc port for XCore.
Hence, WCharType is an unsigned 8bit and WIntType is a signed 32bit.
I believe the test already in test/Preprocessor/stdint.c cover the other cases.
This does lead to question why the change was needed.
It
On Mon, Nov 4, 2013 at 11:19 AM, Chris Wailes chris.wai...@gmail.com wrote:
Are you talking about the function name testIsRValueRefishAndCanonicalType?
The function it is testing in Consumed.cpp is called isRValueRefish. It
tests for RValue references and LValue references that are written as
On Nov 4, 2013, at 11:35 AM, Marshall Clow mclow.li...@gmail.com wrote:
While looking into http://llvm.org/bugs/show_bug.cgi?id=17782, I noticed that
there was a chunk of code in locale for dealing with floating point numbers.
This code was duplicated three times, once for float, double, and
On Nov 2, 2013, at 3:38 , David Majnemer david.majne...@gmail.com wrote:
+ : getLangOpts().CPlusPlus
+ ? diag::ext_flexible_array_union_gnu
+ : diag::err_flexible_array_union;
This doesn't look right. Flexible array
Author: rafael
Date: Mon Nov 4 11:13:51 2013
New Revision: 193997
URL: http://llvm.org/viewvc/llvm-project?rev=193997view=rev
Log:
clang-format this if.
Modified:
cfe/trunk/lib/Driver/Tools.cpp
Modified: cfe/trunk/lib/Driver/Tools.cpp
URL:
The if was a bit hard to read, so a clang-formatted it. A rebased
patch (attached) lgtm.
On 16 October 2013 09:05, Robert Lytton rob...@xmos.com wrote:
Hi
Here is a patch to turn -fno-use-cxa-atexit on as default for the xcore
target.
Robert
On 4 November 2013 08:55, Robert Lytton rob...@xmos.com wrote:
Hi Rafael,
The changes are to bring clang inline with the current llvm-gcc port for
XCore.
Hence, WCharType is an unsigned 8bit and WIntType is a signed 32bit.
OK.
I believe the test already in test/Preprocessor/stdint.c cover
Rafael Espíndola rafael.espind...@gmail.com writes:
We normally commit the tests in the same patch that changed the code...
I thought about that, but in this case it's also testing changes from
2008 that didn't have a test with them (the actual emitBranchOnBool
logic, and specifically the 'if'
On 28.10.2013, at 15:37, Rafael Espíndola rafael.espind...@gmail.com wrote:
On 19 October 2013 18:58, Rafael Espíndola rafael.espind...@gmail.com wrote:
Ping.
ping.
Cheers,
Rafael
t.patch
I can't comment on whether this comment adheres to the limitations of aliases
on ELF, but
Looks good!
Comment at: lib/AST/VTableBuilder.cpp:3314-3316
@@ +3313,5 @@
+const ASTRecordLayout BaseLayout = Context.getASTRecordLayout(CurBase);
+if (!BaseLayout.hasVBPtr() ||
+DerivedVBPtrOffset != BaseOffset + BaseLayout.getVBPtrOffset())
+ break;
+
Author: rafael
Date: Mon Nov 4 12:38:59 2013
New Revision: 194000
URL: http://llvm.org/viewvc/llvm-project?rev=194000view=rev
Log:
Use aliases for more constructors and destructors.
With this patch we produce alias for cases like
templatetypename T
struct foobar {
foobar() {
}
};
template
Author: rikka
Date: Mon Nov 4 12:59:34 2013
New Revision: 194002
URL: http://llvm.org/viewvc/llvm-project?rev=194002view=rev
Log:
Try to correct a mistyped - or to - for some C++ cases.
Similar C code isn't caught as it seems to hit a different code path.
Also, as the check is only done for
On Mon, Nov 4, 2013 at 9:09 AM, Jordan Rose jordan_r...@apple.com wrote:
On Nov 2, 2013, at 3:38 , David Majnemer david.majne...@gmail.com wrote:
+ : getLangOpts().CPlusPlus
+ ? diag::ext_flexible_array_union_gnu
+ :
On Nov 4, 2013, at 11:13, David Majnemer david.majne...@gmail.com wrote:
On Mon, Nov 4, 2013 at 9:09 AM, Jordan Rose jordan_r...@apple.com wrote:
On Nov 2, 2013, at 3:38 , David Majnemer david.majne...@gmail.com wrote:
+ : getLangOpts().CPlusPlus
+
On Mon, Nov 4, 2013 at 11:14 AM, Jordan Rose jordan_r...@apple.com wrote:
On Nov 4, 2013, at 11:13, David Majnemer david.majne...@gmail.com wrote:
On Mon, Nov 4, 2013 at 9:09 AM, Jordan Rose jordan_r...@apple.com wrote:
On Nov 2, 2013, at 3:38 , David Majnemer david.majne...@gmail.com
Author: zaks
Date: Mon Nov 4 13:13:08 2013
New Revision: 194005
URL: http://llvm.org/viewvc/llvm-project?rev=194005view=rev
Log:
[analyzer] Track the count of NSOrderedSet similarly to other fast enumerations.
Modified:
cfe/trunk/lib/StaticAnalyzer/Checkers/BasicObjCFoundationChecks.cpp
Author: zaks
Date: Mon Nov 4 13:13:03 2013
New Revision: 194004
URL: http://llvm.org/viewvc/llvm-project?rev=194004view=rev
Log:
[analyzer] Suppress warnings coming out of std::basic_string.
The analyzer cannot reason about the internal invariances of the data structure
(radar://15194597).
On Nov 4, 2013, at 11:15, David Majnemer david.majne...@gmail.com wrote:
On Mon, Nov 4, 2013 at 11:14 AM, Jordan Rose jordan_r...@apple.com wrote:
On Nov 4, 2013, at 11:13, David Majnemer david.majne...@gmail.com wrote:
On Mon, Nov 4, 2013 at 9:09 AM, Jordan Rose jordan_r...@apple.com
Comment at: lib/AST/MicrosoftMangle.cpp:1849
@@ -1848,2 +1848,3 @@
MicrosoftCXXNameMangler Mangler,
raw_ostream Out) {
+ if (!Adjustment.Virtual.isEmpty()) {
This is complicated
Hi All,
Attached are some minor changes for IBM XLC++/AIX. It is based off the
checkout revision 193999. Thanks!
Diffs of the patch:
(See attached file: libcxx.diffs)
Thanks,
Xing Xue
XLC++ Compiler Development / IBM Software Solutions Toronto Lab.
Tie Line: 313-3972 Phone: 905-413-3972
Hi rsmith, doug.gregor,
This patch fixes Richard's testcase for r193815. Now we include non-explicit
submodules into the list of exports.
The test failed previously because:
- recursive_visibility_a1.inner is not imported (only recursive_visibility_a1
is),
- thus the 'inner' submodule is not
Thanks for explaining. I think the idea of handling the init-captures earlier
is right, but I don't like the direction of creating the `VarDecl`s in the
wrong context and patching them up later. Instead, can you:
* duplicate the `auto` deduction and handling of the initialization sequence
On Fri, Nov 1, 2013 at 6:57 PM, Dmitri Gribenko griboz...@gmail.com wrote:
On Fri, Nov 1, 2013 at 6:47 PM, Dmitri Gribenko griboz...@gmail.com wrote:
On Thu, Oct 31, 2013 at 6:47 PM, Richard Smith rich...@metafoo.co.uk wrote:
I had a go at writing a test for this (attached). However... my test
This is a small optimization on linux, but should help more on windows
where, if I understand it correctly, msvc only outputs one destructor
if there would be two identical ones.
Cheers,
Rafael
t.patch
Description: Binary data
___
cfe-commits mailing
This looks correct to me.
Since we're including non-explicit submodules in the result of
`getExportedModules`, please update `ASTReader::makeModuleVisible` to remove
the code that handles them there.
Alternatively, you could add a loop to `Module::isModuleVisible` to check
whether the
I'm a bit confused by this. We seem to stop emitting some weak_odr symbols
with this patch. Are those symbols not required by the ABI? (If not, why
were they weak_odr and not linkonce_odr?)
On Mon, Nov 4, 2013 at 12:04 PM, Rafael Espíndola
rafael.espind...@gmail.com wrote:
This is a small
Removed the now-redundant code in ASTReader to handle non-explicit submodules.
Hi rsmith, doug.gregor,
http://llvm-reviews.chandlerc.com/D2100
CHANGE SINCE LAST DIFF
http://llvm-reviews.chandlerc.com/D2100?vs=5339id=5341#toc
Files:
lib/Basic/Module.cpp
lib/Serialization/ASTReader.cpp
This LGTM, thanks!
http://llvm-reviews.chandlerc.com/D2100
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Richard, Doug,
Thank you for reviews! Committed r194018.
Dmitri
http://llvm-reviews.chandlerc.com/D2100
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Author: gribozavr
Date: Mon Nov 4 15:51:33 2013
New Revision: 194018
URL: http://llvm.org/viewvc/llvm-project?rev=194018view=rev
Log:
Include non-explicit submodules in exported module list
This change fixes Richard's testcase for r193815. Now we include non-explicit
submodules into the list
On 4 November 2013 12:21, Richard Smith rich...@metafoo.co.uk wrote:
I'm a bit confused by this. We seem to stop emitting some weak_odr symbols
with this patch. Are those symbols not required by the ABI? (If not, why
were they weak_odr and not linkonce_odr?)
They were weak_odr aliases to
Ping.
It's easy to get clang to trigger this bug which results in an invalid
iterator to be returned (which the current code happens to ignore, but
that's just a lucky coincidence), as this regularly occurs during
execution of the lit tests.
On a related note, any suggestions on how to create a
How do you deal with the case where not all references to the aliased
entity have been generated yet? For example:
Emit call to derived ctor
Emit derived ctor
RAUW derived ctor with base ctor
Emit call to derived ctor, create new declaration for it
I attempted this before, and this is the issue
On 4 November 2013 14:37, Reid Kleckner r...@google.com wrote:
How do you deal with the case where not all references to the aliased entity
have been generated yet? For example:
Emit call to derived ctor
Emit derived ctor
RAUW derived ctor with base ctor
Emit call to derived ctor, create
Comment at: include/clang/Parse/Parser.h:638
@@ -637,3 +637,3 @@
/// SkipToTok is specified, it calls SkipUntil(SkipToTok). Finally, true is
/// returned.
bool ExpectAndConsume(tok::TokenKind ExpectedTok, unsigned Diag,
Please describe what NoCount
Hi Kevin
Looks good to me.
I added some CodeGen test and try to use clang-format to check code style.
Very good idea. I do love that tool.
Cheers.
Tim
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Hi Hao,
I think that looks good. Nicely done.
Cheers.
Tim.
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
This all looks fine and obvious to me.
Me too. I've gone through and checked the Constraints which is the
fiddly bit, and they match up.
Cheers.
Tim
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
- Added doxygen, test
Hi klimek, sbenza,
http://llvm-reviews.chandlerc.com/D2095
CHANGE SINCE LAST DIFF
http://llvm-reviews.chandlerc.com/D2095?vs=5334id=5342#toc
Files:
include/clang/ASTMatchers/ASTMatchers.h
include/clang/ASTMatchers/ASTMatchersInternal.h
Author: fjahanian
Date: Mon Nov 4 18:28:21 2013
New Revision: 194031
URL: http://llvm.org/viewvc/llvm-project?rev=194031view=rev
Log:
ObjectiveC. Method implementations should only check for
Missing call to Super in the overriding method and
not in the method itself. // rdar://15385981.
On Mon, Nov 4, 2013 at 2:48 PM, Rafael Espíndola rafael.espind...@gmail.com
wrote:
On 4 November 2013 14:37, Reid Kleckner r...@google.com wrote:
How do you deal with the case where not all references to the aliased
entity
have been generated yet? For example:
Emit call to derived
Why doesn't this apply to inline constructors? They are linkonce_odr.
Sorry, it does too. But they are also from EmitDeferred.
EmitDeferred has a fixpoint loop, though, so it could cause more calls to
the replaced alias to be emitted.
I feel like this needs some defense against some other
We already reject linkonce_odr declarations, so it should fail the
verifier. Any ideas on how to assert earlier?
I decided to use a debugger to step over
struct foo {
~foo() {}
};
struct bar {
~bar() {foo x;}
};
bar y;
foo x;
What happens is that at the end of GetOrCreateLLVMFunction we
Author: kevinqin
Date: Mon Nov 4 20:05:44 2013
New Revision: 194042
URL: http://llvm.org/viewvc/llvm-project?rev=194042view=rev
Log:
Implemented aarch64 neon intrinsic vcopy_lane with float type.
Modified:
cfe/trunk/include/clang/Basic/arm_neon.td
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
On 4 Nov 2013 11:05, Kaelyn Uhrain ri...@google.com wrote:
Author: rikka
Date: Mon Nov 4 12:59:34 2013
New Revision: 194002
URL: http://llvm.org/viewvc/llvm-project?rev=194002view=rev
Log:
Try to correct a mistyped - or to - for some C++ cases.
Similar C code isn't caught as it seems
Friendly ping.
2013/10/28 Serge Pavlov sepavl...@gmail.com
Warn on duplicate function specifier
It looks like things are more complicated here.
Thank you for the comment and for the nice example.
In the updated patch __forceinline is considered as a separate specifier,
but its
On Mon, Nov 4, 2013 at 6:06 PM, Rafael Espíndola rafael.espind...@gmail.com
wrote:
We already reject linkonce_odr declarations, so it should fail the
verifier. Any ideas on how to assert earlier?
I decided to use a debugger to step over
struct foo {
~foo() {}
};
struct bar {
Author: rafael
Date: Mon Nov 4 23:22:36 2013
New Revision: 194046
URL: http://llvm.org/viewvc/llvm-project?rev=194046view=rev
Log:
Produce direct calls instead of alias to linkonce_odr functions.
This is a small optimization on linux, but should help more on windows
where msvc only outputs one
Author: rafael
Date: Mon Nov 4 23:29:54 2013
New Revision: 194047
URL: http://llvm.org/viewvc/llvm-project?rev=194047view=rev
Log:
Revert Produce direct calls instead of alias to linkonce_odr functions.
This reverts commit r194046.
Debugging a bootstrap issue.
Modified:
- Add tests
Hi klimek, sbenza,
http://llvm-reviews.chandlerc.com/D2096
CHANGE SINCE LAST DIFF
http://llvm-reviews.chandlerc.com/D2096?vs=5335id=5345#toc
Files:
include/clang/AST/ASTTypeTraits.h
lib/AST/ASTTypeTraits.cpp
unittests/AST/ASTTypeTraitsTest.cpp
lg
http://llvm-reviews.chandlerc.com/D2096
BRANCH
clangquery2
ARCANIST PROJECT
clang
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Comment at: include/clang/Parse/Parser.h:638
@@ -637,3 +637,3 @@
/// SkipToTok is specified, it calls SkipUntil(SkipToTok). Finally, true is
/// returned.
bool ExpectAndConsume(tok::TokenKind ExpectedTok, unsigned Diag,
hfin...@anl.gov wrote:
Please
Author: rafael
Date: Tue Nov 5 00:44:57 2013
New Revision: 194049
URL: http://llvm.org/viewvc/llvm-project?rev=194049view=rev
Log:
Add a testcase from the recent bootstrap failure.
Modified:
cfe/trunk/test/CodeGenCXX/destructors.cpp
Modified: cfe/trunk/test/CodeGenCXX/destructors.cpp
URL:
On Mon, Nov 4, 2013 at 10:44 PM, Rafael Espindola
rafael.espind...@gmail.com wrote:
Author: rafael
Date: Tue Nov 5 00:44:57 2013
New Revision: 194049
URL: http://llvm.org/viewvc/llvm-project?rev=194049view=rev
Log:
Add a testcase from the recent bootstrap failure.
Modified:
Author: rafael
Date: Tue Nov 5 01:11:42 2013
New Revision: 194050
URL: http://llvm.org/viewvc/llvm-project?rev=194050view=rev
Log:
Expand the test a bit.
Modified:
cfe/trunk/test/CodeGenCXX/destructors.cpp
Modified: cfe/trunk/test/CodeGenCXX/destructors.cpp
URL:
Tests that test that we simply don't crash seem inadequate - could we test
that we have the correct behavior rather than just don't crash (probably
an IRGen test)?
OK. Expanded the test in r194050.
Cheers,
Rafael
___
cfe-commits mailing list
79 matches
Mail list logo