Author: nico
Date: Sat Jun 1 03:12:07 2019
New Revision: 362293
URL: http://llvm.org/viewvc/llvm-project?rev=362293&view=rev
Log:
msabi: Fix exponential mangling time for certain pathological inputs
Template back references used to be recursively recomputed, add a
memoization cache to cut down o
Author: nico
Date: Sun Jun 2 09:47:07 2019
New Revision: 362329
URL: http://llvm.org/viewvc/llvm-project?rev=362329&view=rev
Log:
Add test coverage for __pascal mangling
Modified:
cfe/trunk/test/CodeGenCXX/mangle-windows.cpp
Modified: cfe/trunk/test/CodeGenCXX/mangle-windows.cpp
URL:
http:
Author: nico
Date: Sun Jun 2 10:23:53 2019
New Revision: 362331
URL: http://llvm.org/viewvc/llvm-project?rev=362331&view=rev
Log:
Add mangling test coverage for non-volatile const member pointers
Modified:
cfe/trunk/test/CodeGenCXX/mangle-ms.cpp
Modified: cfe/trunk/test/CodeGenCXX/mangle-ms
Author: nico
Date: Tue Jun 4 16:27:40 2019
New Revision: 362560
URL: http://llvm.org/viewvc/llvm-project?rev=362560&view=rev
Log:
msabi: Fix exponential mangling time for even more contrived inputs
This is a follow-up to r362293 which fixed exponential time needed
for mangling certain templates.
Author: nico
Date: Thu Jun 13 21:05:17 2019
New Revision: 363352
URL: http://llvm.org/viewvc/llvm-project?rev=363352&view=rev
Log:
Revert 363295, it caused PR42276. Also revert follow-ups 363337, 363340.
Revert 363340 "Remove unused SK_LValueToRValue initialization step."
Revert 363337 "PR23833,
Hi,
this made clang segfault on some inputs seen in ANGLE. I filed PR42276 with
a reduced repro and reverted this (and the two dependent changes) in
r363352 for now.
Nico
On Thu, Jun 13, 2019 at 2:56 PM Richard Smith via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: rsmith
> Date:
Most builds seem to be happy. Since you're seeing this problem, maybe you
could debug it a bit?
On Fri, Jun 14, 2019 at 8:11 AM Sven van Haastregt via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: svenvh
> Date: Fri Jun 14 05:14:58 2019
> New Revision: 363376
>
> URL: http://llvm.org
Nevermind, just saw the discussion on phab. Reverting while debugging is of
course fine.
On Fri, Jun 14, 2019 at 8:26 AM Nico Weber wrote:
> Most builds seem to be happy. Since you're seeing this problem, maybe you
> could debug it a bit?
>
> On Fri, Jun 14, 2019 at 8:11 AM Sven van Haastregt vi
Author: nico
Date: Fri Jun 14 05:58:17 2019
New Revision: 363379
URL: http://llvm.org/viewvc/llvm-project?rev=363379&view=rev
Log:
Revert r363298 "[lit] Disable test on darwin when building shared libs."
It broke running check-lld on mac, see https://reviews.llvm.org/D61697
Modified:
cfe/tru
Author: nico
Date: Mon Jun 17 05:10:40 2019
New Revision: 363548
URL: http://llvm.org/viewvc/llvm-project?rev=363548&view=rev
Log:
Promote -fdebug-compilation-dir from a cc1 flag to clang and clang-cl driver
flags
The flag is useful when wanting to create .o files that are independent
from the a
the first place. clang-cl more or less accepts
> all options unless they conflict, so I applied that logic to O0 as well.
>
> I think the right thing to do is to make %cpp_compiler not expand to
> clang-cl if possible. I'll mark UNSUPPORTED for now and try to do that next.
>
> On W
Author: nico
Date: Mon Jul 15 10:12:08 2019
New Revision: 366084
URL: http://llvm.org/viewvc/llvm-project?rev=366084&view=rev
Log:
Use unique_ptr instead of manual delete in one place. No behavior change.
Modified:
cfe/trunk/lib/Driver/Compilation.cpp
Modified: cfe/trunk/lib/Driver/Compilati
Author: nico
Date: Mon Jul 15 10:20:34 2019
New Revision: 366085
URL: http://llvm.org/viewvc/llvm-project?rev=366085&view=rev
Log:
Use a unique_ptr instead of manual memory management for CustomDiagInfo
Modified:
cfe/trunk/include/clang/Basic/DiagnosticIDs.h
cfe/trunk/lib/Basic/Diagnostic
Author: nico
Date: Mon Jul 15 10:27:46 2019
New Revision: 366088
URL: http://llvm.org/viewvc/llvm-project?rev=366088&view=rev
Log:
Use a unique_ptr instead of manual memory management for LineTable
Modified:
cfe/trunk/include/clang/Basic/SourceManager.h
cfe/trunk/lib/Basic/SourceManager.c
Author: nico
Date: Tue Jul 23 10:34:18 2019
New Revision: 366828
URL: http://llvm.org/viewvc/llvm-project?rev=366828&view=rev
Log:
Improve clang-format-diff help output
The description in clang-format-diff.py is more useful than the one
in `clang-format-diff -h`, so use the same description in bo
Author: nico
Date: Tue Jul 23 10:49:45 2019
New Revision: 366831
URL: http://llvm.org/viewvc/llvm-project?rev=366831&view=rev
Log:
clang-format: Fix namespace end comments for namespaces with attributes and
macros.
Fixes PR39247.
While here, also make C++20 `namespace A::inline B::inline C` nes
Author: nico
Date: Wed Jul 24 10:01:57 2019
New Revision: 366926
URL: http://llvm.org/viewvc/llvm-project?rev=366926&view=rev
Log:
clang-format: Add a test that shows that some code I thought was dead is not
dead.
Modified:
cfe/trunk/unittests/Format/FormatTest.cpp
Modified: cfe/trunk/unit
Author: nico
Date: Wed Jul 24 10:15:47 2019
New Revision: 366929
URL: http://llvm.org/viewvc/llvm-project?rev=366929&view=rev
Log:
clang-format: Add another test like r366926
Modified:
cfe/trunk/unittests/Format/FormatTest.cpp
Modified: cfe/trunk/unittests/Format/FormatTest.cpp
URL:
http://
Author: nico
Date: Fri Jul 26 18:13:00 2019
New Revision: 367165
URL: http://llvm.org/viewvc/llvm-project?rev=367165&view=rev
Log:
driver: Don't warn about assembler flags being unused when not assembling;
different approach
This morally relands r365703 (and r365714), originally reviewed at
http
Author: nico
Date: Fri Jul 26 19:41:40 2019
New Revision: 367167
URL: http://llvm.org/viewvc/llvm-project?rev=367167&view=rev
Log:
clang-format: Support `if CONSTEXPR` if CONSTEXPR is a macro.
This is like r305666 (which added support for `if constexpr`) except
that it allows a macro name after t
Author: nico
Date: Sat Jul 27 09:57:19 2019
New Revision: 367176
URL: http://llvm.org/viewvc/llvm-project?rev=367176&view=rev
Log:
Attempt to make test in r367165 more robust.
Some people were seeing this failure:
```
: 'RUN: at line 83'; clang -mrelax-all -fno-integrated-as
/b/s/w/ir/k/llvm-
Author: nico
Date: Mon Jul 29 06:26:48 2019
New Revision: 367225
URL: http://llvm.org/viewvc/llvm-project?rev=367225&view=rev
Log:
clang-format clang/lib/Format
Modified:
cfe/trunk/lib/Format/BreakableToken.cpp
cfe/trunk/lib/Format/Format.cpp
cfe/trunk/lib/Format/TokenAnnotator.h
Author: nico
Date: Thu Aug 1 08:06:57 2019
New Revision: 367574
URL: http://llvm.org/viewvc/llvm-project?rev=367574&view=rev
Log:
Rename two clang tests from .cc to .cpp.
clang/test/lit.cfg.py doesn't list .cc as test extension, so these
tests never ran.
Tweak one of the two tests to actually p
The msan bot doesn't like this, it reports an uninitialized read a
t clang/lib/CrossTU/CrossTranslationUnit.cpp :
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/34087/steps/check-clang%20msan/logs/stdio
Testing: 0
FAIL: Clang :: Analysis/ctu-unknown-part
Author: nico
Date: Mon Aug 5 08:23:10 2019
New Revision: 367875
URL: http://llvm.org/viewvc/llvm-project?rev=367875&view=rev
Log:
Undo what looks like an unintentional change in r367829
The MSan bot was (rightfully) complaining that NumASTLoaded was
unitialized, so put the initialization removed
The problem is probably this part from the diff:
- unsigned NumASTLoaded{0u};
+
+ /// The number successfully loaded ASTs. Used to indicate, and - with
the
+ /// appropriate threshold value - limit the memory usage of the
+ /// CrossTranslationUnitContext.
+ unsigned NumASTLoaded;
i.e. yo
This breaks the sphinx bot:
http://lab.llvm.org:8011/builders/clang-sphinx-docs/builds/45204/steps/docs-clang-html/logs/stdio
Warning, treated as error:
/home/buildbot/llvm-build-dir/clang-sphinx-docs/llvm/src/tools/clang/docs/UsersManual.rst:995:
WARNING: unknown option: -Wno-c++98-compat
On Mo
Author: nico
Date: Mon Aug 5 12:35:59 2019
New Revision: 367912
URL: http://llvm.org/viewvc/llvm-project?rev=367912&view=rev
Log:
Fix another uninit read found by msan after r367829
Modified:
cfe/trunk/include/clang/CrossTU/CrossTranslationUnit.h
Modified: cfe/trunk/include/clang/CrossTU/Cr
There was yet another uninit read:
Testing: 0 .. 10.. 20.. 30.. 40.. 50.. 60.. 70.. 80..
FAIL: Clang-Unit ::
CrossTU/./CrossTUTests/CrossTranslationUnit.RespectsLoadThreshold (13891 of
15323)
TEST 'Clang-Unit ::
CrossTU/./CrossTUTests/CrossTranslationUnit.
Still sad, now with a different message:
http://lab.llvm.org:8011/builders/clang-sphinx-docs/builds/45220
/home/buildbot/llvm-build-dir/clang-sphinx-docs/llvm/src/tools/clang/docs/UsersManual.rst:1002:
WARNING: unknown option: -Wall
On Mon, Aug 5, 2019 at 3:46 PM JF Bastien wrote:
> Fixed. I gu
Thanks!
On Mon, Aug 5, 2019 at 3:58 PM JF Bastien wrote:
> Ugh this is silly… fixed again.
>
> On Aug 5, 2019, at 12:55 PM, Nico Weber wrote:
>
> Still sad, now with a different message:
> http://lab.llvm.org:8011/builders/clang-sphinx-docs/builds/45220
>
> /home/buildbot/llvm-build-dir/clang-s
Author: nico
Date: Thu Aug 8 10:58:32 2019
New Revision: 368322
URL: http://llvm.org/viewvc/llvm-project?rev=368322&view=rev
Log:
clang: Diag running out of file handles while looking for files
clang would only print "file not found" when it's unable to find a
header file. If the reason for tha
Thanks for taking care of this.
I think having a list of "allowed" error codes is more robust long-term
than having an explicit list of unexpected error codes, even though if it
requires some short-term whack-a-mole.
The original change landed after the 9.0 branch, so we have some time for
whack-
http://lab.llvm.org:8011/builders/clang-with-lto-ubuntu/builds/14045
FAILED:
tools/clang/lib/Frontend/CMakeFiles/obj.clangFrontend.dir/ASTConsumers.cpp.o
/home/buildslave/buildslave1a/clang-with-lto-ubuntu/install/stage1/bin/clang++
-DGTEST_HAS_RTTI=0 -D_DEBUG -D_GNU_SOURCE -D__STDC_CONSTANT_MA
This fixes `+ X(std::unique_ptr up) : pointee(up.get()),
pointer(std::move(up)) {}` as well, right?
On Fri, Aug 9, 2019 at 8:31 PM Gabor Horvath via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: xazax
> Date: Fri Aug 9 17:32:29 2019
> New Revision: 368501
>
> URL: http://llvm.org/v
Up to you :)
On Fri, Aug 9, 2019 at 8:54 PM Gábor Horváth wrote:
> It does! Do you want me to commit that test as well?
>
> On Fri, 9 Aug 2019 at 17:50, Nico Weber wrote:
>
>> This fixes `+ X(std::unique_ptr up) : pointee(up.get()),
>> pointer(std::move(up)) {}` as well, right?
>>
>> On Fri, A
Hi Gabor,
this makes clang warn on this program:
$ cat test.cc
#include
#include
struct S {
bool operator==(const S &rhs) const;
};
const S &f(const std::vector &v) {
const auto &it = std::find(v.rbegin(), v.rend(), S());
return *it;
}
$ out/gn/bin/clang -c test.cc -isysroot $(xcrun -
Confirmed, after r368534 the Chromium builds looks good again. (r368528
without r368534 regressed something else, but r368534 fixed things. I'm
assuming you don't need a repro for the breakage in r368528
without r368534.)
Your warnings even found a bug:
https://bugs.chromium.org/p/openscreen/issue
Author: nico
Date: Tue Aug 13 10:37:09 2019
New Revision: 368728
URL: http://llvm.org/viewvc/llvm-project?rev=368728&view=rev
Log:
clang: Don't warn on unused momit-leaf-frame-pointer when frame pointers are
off.
This fixes a regression from r365860: As that commit message
states, there are 3 va
>From the review: This is for gcc 4.8 compat, not for the other reasons
mentioned in the CL description.
On Tue, Aug 13, 2019 at 3:31 PM Alexey Bataev via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: abataev
> Date: Tue Aug 13 12:32:36 2019
> New Revision: 368739
>
> URL: http://llv
Hi,
this results in a false positive on webrtc, on this code:
https://cs.chromium.org/chromium/src/third_party/libwebp/src/enc/picture_csp_enc.c?q=picture_csp_enc.c&sq=package:chromium&dr&l=1002
The code does this:
#ifdef WORDS_BIGENDIAN
#define ALPHA_OFFSET 0 // uint32_t 0xff00 is 0xff,0
Author: nico
Date: Wed Jun 26 10:51:47 2019
New Revision: 364453
URL: http://llvm.org/viewvc/llvm-project?rev=364453&view=rev
Log:
Make AddLastArg() variadic and use it more. No behavior change.
Modified:
cfe/trunk/lib/Driver/ToolChains/Clang.cpp
Modified: cfe/trunk/lib/Driver/ToolChains/Cla
Turns out this doesn't actually work: -fdump-records-layout is a cc1
option, not a clang driver option. The correct td entry would be
def _SLASH_d1reportAllClassLayout : CLFlag<"d1reportAllClassLayout">,
HelpText<"Dump record layout information">,
Alias, AliasArgs<["-fdump-record-layouts"]>;
Author: nico
Date: Tue Jul 2 03:22:11 2019
New Revision: 364901
URL: http://llvm.org/viewvc/llvm-project?rev=364901&view=rev
Log:
clang-cl: Make /d1reportAllClassLayout actually work and improve test
See review thread for r301567.
Modified:
cfe/trunk/include/clang/Driver/CLCompatOptions.td
Ah, Hans is away for a while. I landed the simple-but-unfulfilling fix in
r364901 to unblock my other change for now. Maybe we can tweak it once Hans
is back.
On Mon, Jul 1, 2019 at 9:21 PM Nico Weber wrote:
> Turns out this doesn't actually work: -fdump-records-layout is a cc1
> option, not a c
Author: nico
Date: Fri Jul 5 04:45:24 2019
New Revision: 365186
URL: http://llvm.org/viewvc/llvm-project?rev=365186&view=rev
Log:
Make joined instances of JoinedOrSeparate flags point to the unaliased args,
like all other arg types do
This fixes an 8-year-old regression. r105763 made it so that
Author: nico
Date: Fri Jul 5 08:12:31 2019
New Revision: 365212
URL: http://llvm.org/viewvc/llvm-project?rev=365212&view=rev
Log:
Add a comment explaining why a function exists
Modified:
cfe/trunk/lib/Driver/Driver.cpp
Modified: cfe/trunk/lib/Driver/Driver.cpp
URL:
http://llvm.org/viewvc/l
Author: nico
Date: Mon Jul 8 17:02:23 2019
New Revision: 365411
URL: http://llvm.org/viewvc/llvm-project?rev=365411&view=rev
Log:
clang-cl: Port cl.exe's C4659 to clang-cl
Differential Revision: https://reviews.llvm.org/D64349
Modified:
cfe/trunk/include/clang/Basic/Attr.td
cfe/trunk/in
Author: nico
Date: Mon Jul 8 17:34:08 2019
New Revision: 365413
URL: http://llvm.org/viewvc/llvm-project?rev=365413&view=rev
Log:
Let unaliased Args track which Alias they were created from, and use that in
Arg::getAsString() for diagnostics
With this, `clang-cl /source-charset:utf-16 test.cc`
Author: nico
Date: Mon Jul 8 17:36:18 2019
New Revision: 365414
URL: http://llvm.org/viewvc/llvm-project?rev=365414&view=rev
Log:
Use `ln -n` to prevent forming a symlink cycle, instead of rm'ing the source
This is a better fix for the problem fixed in r334972.
Also remove the rm'ing of the sym
Author: nico
Date: Wed Jul 10 07:49:36 2019
New Revision: 365638
URL: http://llvm.org/viewvc/llvm-project?rev=365638&view=rev
Log:
Remove two unused member variables.
They were added over 10 years ago in r66575 and have never been used as
far as I can tell.
(r67087 added similar fields to Compil
Author: nico
Date: Wed Jul 10 15:29:35 2019
New Revision: 365703
URL: http://llvm.org/viewvc/llvm-project?rev=365703&view=rev
Log:
driver: Don't warn about assembler flags being unused when not assembling
clang currently warns when passing flags for the assembler (e.g.
-Wa,-mbig-obj) to an invoca
Author: nico
Date: Wed Jul 10 16:50:16 2019
New Revision: 365714
URL: http://llvm.org/viewvc/llvm-project?rev=365714&view=rev
Log:
Attempt to fix bots after r365703
Modified:
cfe/trunk/test/Driver/as-options.s
Modified: cfe/trunk/test/Driver/as-options.s
URL:
http://llvm.org/viewvc/llvm-pro
Author: nico
Date: Wed Jul 10 18:13:38 2019
New Revision: 365721
URL: http://llvm.org/viewvc/llvm-project?rev=365721&view=rev
Log:
Various minor tweaks to CLCompatOptions.td
- Add back indentation I accidentally removed in r364901
- Wrap two lines to 80 cols
- Slightly tighten up help text for se
Author: nico
Date: Wed Jul 10 18:18:05 2019
New Revision: 365724
URL: http://llvm.org/viewvc/llvm-project?rev=365724&view=rev
Log:
clang-cl: Remove -O0 option
cl.exe doesn't understand it; there's /Od instead. See also the review
thread for r229575.
Update lots of compiler-rt tests to use -Od in
Author: nico
Date: Thu Jul 11 08:26:45 2019
New Revision: 365789
URL: http://llvm.org/viewvc/llvm-project?rev=365789&view=rev
Log:
Fix a few 'no newline at end of file' warnings that Xcode emits
(Xcode even has a snazzy "Fix" button, but clicking that inserts two
newlines. So close!)
Modified:
Author: nico
Date: Fri May 10 06:56:56 2019
New Revision: 360427
URL: http://llvm.org/viewvc/llvm-project?rev=360427&view=rev
Log:
Add name comments for bools passed to getObjCEncodingForTypeImpl()
No behavior change. Medium term, probably want to use a bitmask instead
of 8 distinct bool paramete
Author: nico
Date: Tue May 14 05:32:37 2019
New Revision: 360668
URL: http://llvm.org/viewvc/llvm-project?rev=360668&view=rev
Log:
Make getObjCEncodingForTypeImpl() take a bitmask instead of 8 bools
Slightly easier to read, uses slightly less stack space, and makes it
impossible to mix up the ord
Hello,
this breaks building with -DLLVM_ENABLE_PIC=OFF. Maybe the new target
shouldn't be build in those builds?
Also, if I read this right, this makes static libraries for clang always be
object libraries. Is that correct? If so, this likely makes the normal
clang binary larger and less efficie
Author: nico
Date: Thu May 16 18:42:37 2019
New Revision: 360973
URL: http://llvm.org/viewvc/llvm-project?rev=360973&view=rev
Log:
Revert r360946 "Add Clang shared library with C++ exports"
It breaks LLVM_ENABLE_PIC=OFF builds, and it's not clear
if the object library approach doesn't impact the
To not keep the build broken over night, and since it's a small change that
should be easy to reland, I've reverted this for now in r360973.
On Thu, May 16, 2019 at 8:52 PM Nico Weber wrote:
> Hello,
>
> this breaks building with -DLLVM_ENABLE_PIC=OFF. Maybe the new target
> shouldn't be build i
Thanks! Bots are happy with the reland. Thanks for taking the time to
explain the object library stuff!
*From: *Chris Bieneman
*Date: *Thu, May 16, 2019 at 11:24 PM
*To: *Nico Weber
*Cc: *cfe-commits
I’ll re-land disabling if PIC is off.
>
> This doesn’t replace the static libraries with object
It looks like this change is making gcc-7 crash on these (and other
http://lab.llvm.org:8011/console) bots:
http://lab.llvm.org:8011/builders/clang-cmake-aarch64-quick/builds/18639
http://lab.llvm.org:8011/builders/clang-ppc64be-linux/builds/33837/
[100/212] Building CXX object
tools/clang/unitte
Looks like your r361042 fixed the ppc64be bot nonetheless. Maybe it was
related after all :)
*From: *Kristof Umann
*Date: *Fri, May 17, 2019 at 11:10 AM
*To: *Nico Weber
*Cc: *cfe-commits
http://lab.llvm.org:8011/builders/clang-cmake-aarch64-quick seems to be
> doing fine now, http://lab.llvm.or
The test fails on Windows:
*From: *Richard Smith via cfe-commits
*Date: *Fri, May 17, 2019 at 3:58 AM
*To: *
Author: rsmith
> Date: Fri May 17 01:01:34 2019
> New Revision: 360998
>
> URL: http://llvm.org/viewvc/llvm-project?rev=360998&view=rev
> Log:
> Fix crash if, during evaluation of __bui
-- Testing: 49966 tests, 32 threads --
Testing: 0 .. 10.. 20..
FAIL: Clang :: SemaCXX/builtin-object-size-cxx14.cpp (14324 of 49966)
TEST 'Clang :: SemaCXX/builtin-object-size-cxx14.cpp'
FAILED
Script:
--
: 'RUN: at line 1';
c:\b\s\w\ir\k\src\third_party\ll
Also on the LLVM waterfall:
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/6937/steps/stage%201%20check/logs/stdio
(Same error.)
*From: *Nico Weber
*Date: *Fri, May 17, 2019 at 1:14 PM
*To: *Richard Smith
*Cc: *cfe-commits
Author: nico
Date: Fri May 17 10:33:54 2019
New Revision: 361054
URL: http://llvm.org/viewvc/llvm-project?rev=361054&view=rev
Log:
Attempt to fix test on Windows after r360998
Modified:
cfe/trunk/test/SemaCXX/builtin-object-size-cxx14.cpp
Modified: cfe/trunk/test/SemaCXX/builtin-object-size-
I attempted to fix this in 361054.
*From: *Nico Weber
*Date: *Fri, May 17, 2019 at 1:16 PM
*To: *Richard Smith
*Cc: *cfe-commits
Also on the LLVM waterfall:
> http://lab.llvm.org:8011/builders/clang-x64-windows-msvc
>
> http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/6937/steps/s
Author: nico
Date: Fri May 17 11:31:24 2019
New Revision: 361059
URL: http://llvm.org/viewvc/llvm-project?rev=361059&view=rev
Log:
Revert r361033 "Add a Visit overload for DynTypedNode to ASTNodeTraverser"
It fails to build on some bots.
Also revert follow-up r361055.
Removed:
cfe/trunk/uni
*From: *Stephen Kelly via cfe-commits
*Date: *Fri, May 17, 2019 at 1:56 PM
*To: *Yitzhak Mandelbaum
*Cc: *cfe-commits
Hmm, thanks for letting me know!
>
> Is that an old compiler?
>
> I'm not near a computer for the weekend.
>
> Can someone fix or revert?
>
Sure, reverted in 361059.
>
> Thanks
Hello, it seems this broke building some iOS programs.
$ cat test.m
# define UI_APPEARANCE_SELECTOR
__attribute__((annotate("ui_appearance_selector")))
@class UIColor;
@interface Test
@property(null_resettable, nonatomic, strong) UIColor *onTintColor
UI_APPEARANCE_SELECTOR;
@end
@implementatio
Author: nico
Date: Mon May 20 17:27:26 2019
New Revision: 361209
URL: http://llvm.org/viewvc/llvm-project?rev=361209&view=rev
Log:
Try to fix build with older gccs after r361152
Also merge the cmake change there to the gn build.
Modified:
cfe/trunk/unittests/Tooling/RangeSelectorTest.cpp
Mo
+ LLVMSupport
+ LLVMTestingSupport
clang/unittests/Tooling/CMakeLists already has this at the top:
set(LLVM_LINK_COMPONENTS
${LLVM_TARGETS_TO_BUILD}
Support
)
I think we link Support twice now. Did adding TestingSupport up there not
work? I'd expect that your addition would break shared l
I'm suggesting you make the below change. For testing, I'd just try
building locally.
$ git diff
diff --git a/clang/unittests/Tooling/CMakeLists.txt
b/clang/unittests/Tooling/CMakeLists.txt
index 8c383be2d74..af8a35d9251 100644
--- a/clang/unittests/Tooling/CMakeLists.txt
+++ b/clang/unittests/Too
Author: nico
Date: Thu May 23 10:58:33 2019
New Revision: 361518
URL: http://llvm.org/viewvc/llvm-project?rev=361518&view=rev
Log:
lld-link, clang: Treat non-existent input files as possible spellos for option
flags
OptTable treats arguments starting with / that aren't a known option
as filename
Author: nico
Date: Fri May 24 17:27:19 2019
New Revision: 361684
URL: http://llvm.org/viewvc/llvm-project?rev=361684&view=rev
Log:
Rename clangToolingRefactor to clangToolingRefactoring for consistency with its
directory
See "[cfe-dev] The name of clang/lib/Tooling/Refactoring".
Differential Re
Author: nico
Date: Fri May 24 17:27:19 2019
New Revision: 361684
URL: http://llvm.org/viewvc/llvm-project?rev=361684&view=rev
Log:
Rename clangToolingRefactor to clangToolingRefactoring for consistency with its
directory
See "[cfe-dev] The name of clang/lib/Tooling/Refactoring".
Differential Re
Author: nico
Date: Tue May 28 07:04:48 2019
New Revision: 361824
URL: http://llvm.org/viewvc/llvm-project?rev=361824&view=rev
Log:
Add release note entries for recent typo correction changes
Differential Revision: https://reviews.llvm.org/D62523
Modified:
cfe/trunk/docs/ReleaseNotes.rst
Mod
Author: nico
Date: Wed May 29 06:48:19 2019
New Revision: 361959
URL: http://llvm.org/viewvc/llvm-project?rev=361959&view=rev
Log:
clang-cl: Fix mangling of catchable types with names longer than 4kiB
The mangling used to contain the MD5 name of both the RTTI type
descriptor and the name of the c
Have you had a chance to look at making this change?
On Mon, Apr 9, 2018 at 9:08 AM Yuka Takahashi wrote:
> Sounds good!
>
> 2018-04-09 15:03 GMT+02:00 Nico Weber :
>
>> Yes.
>>
>> On Mon, Apr 9, 2018 at 9:00 AM, Yuka Takahashi wrote:
>>
>>> Hi Nico,
>>>
>>> Thanks for your comment!
>>>
>>> I d
Author: nico
Date: Wed Dec 26 08:06:26 2018
New Revision: 350067
URL: http://llvm.org/viewvc/llvm-project?rev=350067&view=rev
Log:
Pass a concrete triple for two OpenMP tests that depend on TLS
Not all %itanium_abi_triple values support TLS. Makes
OpenMP/declare_reduction_codegen.cpp, OpenMP/para
Should this be in -Wextra-semi? It seems weird to have to pass both
-Wextra-semi and -Wextra-semi-stmt.
(Also, we generally try to not add DefaultIgnore warnings. I think this is
a useful warning though, and other than others in this thread I think it'd
be good to have -Wextra-semi in -Wall. But I
Author: nico
Date: Wed Jan 2 16:17:02 2019
New Revision: 350278
URL: http://llvm.org/viewvc/llvm-project?rev=350278&view=rev
Log:
Make test/Driver/darwin-sdk-version.c pass if the host triple is 32-bit
For some reason, the cmake build on my macbook has
LLVM_HOST_TRIPLE:STRING=i386-apple-darwin16
Author: nico
Date: Thu Jan 3 10:26:06 2019
New Revision: 350340
URL: http://llvm.org/viewvc/llvm-project?rev=350340&view=rev
Log:
Validate -add-plugin arguments.
-plugin already prints an error if the name of an unknown plugin is passed.
-add-plugin used to silently ignore that, now it errors to
Author: nico
Date: Fri Jan 4 17:10:20 2019
New Revision: 350451
URL: http://llvm.org/viewvc/llvm-project?rev=350451&view=rev
Log:
Move -add-plugin validation after -load was executed.
Moves the code added in r350340 around a bit, to hopefully make the existing
plugin tests pass when clang is bui
m.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/41673
>
> . . .
> Failing Tests (2):
> Clang :: CodeGenCXX/debug-info-class-limited-plugin.test
> Clang :: Frontend/plugin-vs-debug-info.cpp
>
> Please have a look ASAP?
>
> Thanks
>
> Gali
Author: nico
Date: Fri Jan 4 17:19:14 2019
New Revision: 350453
URL: http://llvm.org/viewvc/llvm-project?rev=350453&view=rev
Log:
Let new test from r350340 still pass even after r350451.
Modified:
cfe/trunk/lib/Frontend/FrontendAction.cpp
Modified: cfe/trunk/lib/Frontend/FrontendAction.cpp
Nice, this is finding bugs:
https://bugs.chromium.org/p/chromium/issues/detail?id=919262
However, I noticed that for that case, the same warning is printed twice:
../../third_party/crashpad/crashpad/util/win/process_info.cc(227,36):
error: expression result unused [-Werror,-Wunused-value]
N
Author: nico
Date: Sun Jan 6 07:57:18 2019
New Revision: 350495
URL: http://llvm.org/viewvc/llvm-project?rev=350495&view=rev
Log:
Fix bug in test found by the diagnostic added in r350340.
I meant to commit this change in 350341 but failed to do so (since it's
in test/CodeGenCXX, not in test/Fron
ilds/41673
>>
>> . . .
>> Failing Tests (2):
>> Clang :: CodeGenCXX/debug-info-class-limited-plugin.test
>> Clang :: Frontend/plugin-vs-debug-info.cpp
>>
>> Please have a look ASAP?
>>
>> Thanks
>>
>> Galina
>>
>> On T
It looks like clang now looks for libc++ headers in -internal-isystem
Release+Asserts/bin/include/c++/v1 , compared to -internal-isystem
Release+Asserts/include/c++/v1. `make install` puts the libc++ headers in
Release+Asserts/include, the old location. Was this an intentional change?
As-is, this
That looks like it should help. Thanks for the quick fix!
On Tue, Jan 8, 2019 at 1:11 PM Ilya Biryukov wrote:
> Hi Nico,
>
> This is clearly a bug, it's supposed to search in a sibling directory.
> Are you running clang as './clang' in the scripts? The code seems to
> break in that case, https:
Author: nico
Date: Wed Jan 9 06:19:16 2019
New Revision: 350724
URL: http://llvm.org/viewvc/llvm-project?rev=350724&view=rev
Log:
Fix typo in comment
Modified:
cfe/trunk/include/clang/Sema/Sema.h
Modified: cfe/trunk/include/clang/Sema/Sema.h
URL:
http://llvm.org/viewvc/llvm-project/cfe/tru
Aren't C names sometimes slightly mangled too? For example, on macOS
they're prefixed by a _, doesn't that have to be undone there?
On Wed, Jan 9, 2019 at 6:58 PM Nick Desaulniers via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: nickdesaulniers
> Date: Wed Jan 9 15:54:55 2019
> New
../../clang/lib/Lex/PPExpressions.cpp:852:43: warning: missing field
'ExprRange' initializer [-Wmissing-field-initializers]
return {false, DT.IncludedUndefinedIds};
^
On Thu, Jan 10, 2019 at 4:26 PM Aaron Ballman via cfe-commits <
cfe-commits@lists.llv
Author: nico
Date: Fri Sep 13 06:18:55 2019
New Revision: 371854
URL: http://llvm.org/viewvc/llvm-project?rev=371854&view=rev
Log:
clang-format: Add support for formatting (some) lambdas with explicit template
parameters.
This patch makes cases work where the lambda's template list doesn't
conta
This looks similar to https://reviews.llvm.org/D67542
On Sat, Sep 14, 2019 at 1:59 AM Fangrui Song via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: maskray
> Date: Fri Sep 13 23:01:22 2019
> New Revision: 371918
>
> URL: http://llvm.org/viewvc/llvm-project?rev=371918&view=rev
> Log:
CodeGen/ARM/CGP/arm-gcp-casts.ll is broken too.
On Mon, Sep 16, 2019 at 1:44 PM Roman Lebedev via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: lebedevri
> Date: Mon Sep 16 10:46:08 2019
> New Revision: 372015
>
> URL: http://llvm.org/viewvc/llvm-project?rev=372015&view=rev
> Log:
>
Bots look happy again. Thanks for the quick fix :)
On Mon, Sep 16, 2019 at 2:27 PM Roman Lebedev wrote:
> Oh, that looked like it was due to r372012 but apparently it wasn't.
> Fixed in r372019.
> Sorry for the breakage.
>
> Roman
>
> On Mon, Sep 16, 2019 at 9:11 PM Nico Weber wrote:
> >
> > Co
201 - 300 of 1693 matches
Mail list logo
