important.
On Sun, Jul 26, 2015 at 8:54 AM Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Hans and Nick,
I would like to propose to merge following patches to libunwind 3.7
branch:
r242642: libunwind: Introduce __libunwind_config.h.
r243073: unwind: Fix libc++abi and libgcc build
Hi Hans and Nick,
I would like to propose to merge following patches to libunwind 3.7 branch:
r242642: libunwind: Introduce __libunwind_config.h.
r243073: unwind: Fix libc++abi and libgcc build.
We need these changes to build libc++abi with libgcc (without
libunwind.so.) IMO, this is an
logan closed this revision.
logan added a comment.
Thanks for reviewing. Committed as http://reviews.llvm.org/rL243073.
http://reviews.llvm.org/D11190
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
logan added a comment.
Since there's no way to remove libraries from the list (because the compiler
adds them), there is no way a warning would be effective. Ie. the user would
get a warning and would be able to do nothing.
Although it is difficult to use, we can remove the default
logan abandoned this revision.
logan added a comment.
Stashing the change with http://reviews.llvm.org/D11190.
http://reviews.llvm.org/D11191
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
logan updated this revision to Diff 30299.
logan added a comment.
Stash the changes from http://reviews.llvm.org/D11191.
http://reviews.llvm.org/D11190
Files:
include/unwind.h
src/UnwindLevel1.c
Index: src/UnwindLevel1.c
===
logan added a comment.
Hi,
Please have a look on the revised version. This should address all of the
comments at the moment. I really wish to commit this change before 3.7
release. Thanks.
Logan
http://reviews.llvm.org/D11190
___
logan closed this revision.
logan added a comment.
Thanks for reviewing!
Committed as http://reviews.llvm.org/rL242642.
http://reviews.llvm.org/D11230
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
logan added inline comments.
Comment at: include/unwind.h:217
@@ -216,1 +216,3 @@
+#if !defined(__LIBUNWIND_DONT_INLINE_UNWIND_LEVEL1)
+#define __LIBUNWIND_EXPORT_UNWIND_LEVEL1 static __inline__
compnerd wrote:
The double negative hurts. Would you be opposed
logan added a comment.
@compnerd:
For your use case, please have a look at the upcoming
http://reviews.llvm.org/D11191. I believe it should work for you. Although I
am really hesitating, I have just confirmed that Savannah libunwind[1] is
working on ARM. So I am not strongly oppose to the
logan updated this revision to Diff 29802.
logan added a comment.
Address asl's comment.
http://reviews.llvm.org/D11190
Files:
include/unwind.h
src/UnwindLevel1.c
Index: src/UnwindLevel1.c
===
--- src/UnwindLevel1.c
+++
logan added a comment.
@danalbert:
May you further elaborate your idea? I don't quite understand why should we
install the headers of libunwind for libcxx?
http://reviews.llvm.org/D11230
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
logan created this revision.
logan added reviewers: asl, joerg, rengolin, danalbert, compnerd.
logan added a subscriber: cfe-commits.
Introduce __libunwind_config.h to avoid cross repository circular
dependency with libcxxabi.
http://reviews.llvm.org/D11230
Files:
include/__libunwind_config.h
logan added inline comments.
Comment at: include/unwind.h:224
@@ +223,3 @@
+
+static __inline__ uintptr_t _Unwind_GetGR(struct _Unwind_Context *context,
+ int index) {
asl wrote:
logan wrote:
asl wrote:
Why can't we
logan added a comment.
I have some concern on the design of this change. IMO, it is unintuitive that
one `-l` option will affect the other `-l` options. Can't we simply rely on
`--rtlib=compiler-rt`? Or, alternatively, can we simply emit a warning instead
of changing the behavior quietly?
logan created this revision.
logan added reviewers: rengolin, joerg, danalbert, asl, compnerd.
logan added a subscriber: cfe-commits.
logan added a dependency: D11190: unwind: Fix libc++abi and libgcc build..
Herald added subscribers: rengolin, aemerson.
I have removed the exported symbols of
logan created this revision.
logan added reviewers: rengolin, joerg, danalbert, asl, compnerd.
logan added a subscriber: cfe-commits.
To build libc++abi without libunwind, we should make sure that all
function calls to _Unwind_{Get,Set}{GR,IP}() are inlined as function
calls to
logan added inline comments.
Comment at: include/unwind.h:224
@@ +223,3 @@
+
+static __inline__ uintptr_t _Unwind_GetGR(struct _Unwind_Context *context,
+ int index) {
asl wrote:
Why can't we simply make them static
Author: logan
Date: Sun Jul 12 09:47:53 2015
New Revision: 241993
URL: http://llvm.org/viewvc/llvm-project?rev=241993view=rev
Log:
Remove include directive for the unused libunwind_ext.h.
Modified:
libcxxabi/trunk/src/cxa_personality.cpp
Modified: libcxxabi/trunk/src/cxa_personality.cpp
Hi,
Thanks for your review. Committed as http://reviews.llvm.org/rL240648.
http://reviews.llvm.org/D10720
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Hi kledzik, rengolin,
unw_getcontext() should return UNW_ESUCCESS on success. However, the assembly
for AArch64 is incorrect since ldr x0, #0 is a PC-relative load instead of an
immediate value load.
This patch should fix the problem by changing ldr to mov.
http://reviews.llvm.org/D10720
Note that we need to enable -funwind-frame to UnwindLevel1-gcc-ext.c
I guess you mean `-funwind-tables`?
Yes, `-funwind-tables` will be added to CFLAGS if we build the library with
`cmake`.
http://reviews.llvm.org/D9961
EMAIL PREFERENCES
Hi Anton and Renato,
Thanks for reviewing. Committed as http://reviews.llvm.org/rL238560.
http://reviews.llvm.org/D9961
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
Hi Anton and Renato,
Thanks for reviewing. Committed as http://reviews.llvm.org/rL238561.
http://reviews.llvm.org/D9962
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
Author: logan
Date: Fri May 29 10:34:24 2015
New Revision: 238561
URL: http://llvm.org/viewvc/llvm-project?rev=238561view=rev
Log:
libcxx: Switch to use __gnu_unwind_frame() for libunwind.
As a step to fix libunwind unw_step(), a new function
__gnu_unwind_frame() has been introduced to
In http://reviews.llvm.org/D9962#177558, @rengolin wrote:
And I imagine that this will also fix the problem we had earlier with
undefined __gnu_unwind_frame, right?
I haven't encountered this problem by myself. But I believe this should fix
your problem.
http://reviews.llvm.org/D9962
Hi asl, rengolin, danalbert, compnerd, joerg,
As a step to fix libunwind unw_step(), a new function
__gnu_unwind_frame() has been introduced to libunwind, and it is
required to use this function so that some libunwind internal data
structure can be updated properly.
Depends On: D9961
Hi asl, rengolin, danalbert, compnerd,
This commit fixes the unw_step() for ARM EHABI. However, this commit
also changes the implementation details for ARM EHABI.
The first change is that the personality function should call
__gnu_unwind_frame() for default (or de facto) frame unwinding based
Author: logan
Date: Sat May 16 08:10:39 2015
New Revision: 237519
URL: http://llvm.org/viewvc/llvm-project?rev=237519view=rev
Log:
Fix build when libunwind is disabled.
The previous commit breaks the builds when libc++abi is not built with
libunwind becuase the default value for
Hi rengolin, EricWF,
Print both the compiler command and linker command so that it will be
easier for developers to reproduce the failed test cases.
http://reviews.llvm.org/D9807
Files:
test/libcxx/compiler.py
Index: test/libcxx/compiler.py
Hi rengolin, danalbert, compnerd,
The test cases were crashing due to the mixed usage of the unwinding
functions from both libunwind and libgcc_s. The unwind functions are
mixed because the llvm_unwinder entry is not available in the
lit.site.cfg for libc++. As a result, -lgcc_s is picked
Author: logan
Date: Sat May 16 07:44:31 2015
New Revision: 237518
URL: http://llvm.org/viewvc/llvm-project?rev=237518view=rev
Log:
libcxx: Fix ARM libc++/abi and libunwind buildbot.
The test cases were crashing due to the mixed usage of the unwinding
functions from both libunwind and libgcc_s.
Thanks for your review. Committed as http://reviews.llvm.org/rL237518.
Let's see if it works.
http://reviews.llvm.org/D9808
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
Thanks for reviewing. Committed as http://reviews.llvm.org/rL237530.
http://reviews.llvm.org/D9807
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Author: logan
Date: Sat May 16 19:24:11 2015
New Revision: 237530
URL: http://llvm.org/viewvc/llvm-project?rev=237530view=rev
Log:
libcxx: Enhance lit test command in verbose mode.
Print both the compiler command and linker command so that it will be
easier for developers to reproduce the failed
Hi Saleem,
Please reconsider this change. IMHO, the correct solution is to avoid the
extension instead of marking LLVM-libunwind as mandatory. Thanks.
Logan
On Tue, Apr 28, 2015 at 2:19 AM, Saleem Abdulrasool compn...@compnerd.org
javascript:_e(%7B%7D,'cvml','compn...@compnerd.org'); wrote:
Hi Saleem,
Any updated comments on this? Thanks.
Logan
On Tue, Mar 24, 2015 at 5:46 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Saleem,
Sorry for the late reply.
Although I am not strongly oppose to the idea to provide both inline and
extern version, I am concerned
Ping?
On Tue, Mar 24, 2015 at 5:46 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Saleem,
Sorry for the late reply.
Although I am not strongly oppose to the idea to provide both inline and
extern version, I am concerned that exporting these symbols will further
fragmentize
of declaring their own function prototype.
Sincerely,
Logan
On Wed, Mar 18, 2015 at 10:14 AM, Saleem Abdulrasool compn...@compnerd.org
wrote:
On Tue, Mar 17, 2015 at 10:42 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Saleem,
Why would libc++abi have undefined references, it it doesn't
...@compnerd.org
wrote:
On Sat, Mar 14, 2015 at 8:43 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Saleem,
Similar change has been committed and reverted (see r219629 and r226818.)
This change WILL BREAK the libc++abi + libgcc build. With this change,
the libc++abi shared library will come
Hi Saleem,
Similar change has been committed and reverted (see r219629 and r226818.)
This change WILL BREAK the libc++abi + libgcc build. With this change, the
libc++abi shared library will come with undefined references to
_Unwind_{Get,Set}{IP,GR}, which are not available in libgcc.
AFAICT,
Author: logan
Date: Thu Mar 12 12:27:19 2015
New Revision: 232065
URL: http://llvm.org/viewvc/llvm-project?rev=232065view=rev
Log:
[docs] Replace the doxygen qch option properly.
Modified:
cfe/trunk/docs/Makefile
Modified: cfe/trunk/docs/Makefile
URL:
)
Thanks.
Logan
On Sat, Feb 28, 2015 at 4:27 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi,
I found that due to some setting problem (missing or outdated css and
templates), the Doxygen documentation looks strange at the moment.
In the attached patch sets, I have updated the doxygen
landed as below:
On Tue, Jan 27, 2015 at 8:50 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
r226818 - Remove _Unwind_{Get,Set}{GR,IP} from ARM EHABI build.
r227234
r226819 - Allow libc++abi to be built without unwinder.
This one had a merge conflict in CMakeLists.txt; please double
Hi,
I would like to propose to merge r226818, r226819, r226820, r226822,
r226823, and r226824 to the release_36 branch for libc++abi.
Here are the reasons for these changes:
1. Fixes the compilation of libc++abi for ARM without LLVM unwinder.
2. Fix the existing test case backtrace_test which
Thanks. http://reviews.llvm.org/rL226896 works for me. Abandon this change.
http://reviews.llvm.org/D5743
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Author: logan
Date: Thu Jan 22 07:27:36 2015
New Revision: 226819
URL: http://llvm.org/viewvc/llvm-project?rev=226819view=rev
Log:
Allow libc++abi to be built without unwinder.
This CL adds a new compilation flags LIBCXXABI_USE_LLVM_UNWINDER
to specify whether the LLVM unwinder is enabled.
Thanks. Committed as http://reviews.llvm.org/rL226820.
http://reviews.llvm.org/D7031
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Thanks. Committed as http://reviews.llvm.org/rL226823.
http://reviews.llvm.org/D7034
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Author: logan
Date: Thu Jan 22 07:38:11 2015
New Revision: 226822
URL: http://llvm.org/viewvc/llvm-project?rev=226822view=rev
Log:
Force unwind frame with user-defined personality.
If libcxxabi is compiled as a shared library, and the
executable references the user-defined personality routines
Author: logan
Date: Thu Jan 22 07:28:39 2015
New Revision: 226820
URL: http://llvm.org/viewvc/llvm-project?rev=226820view=rev
Log:
Fix _Unwind_Backtrace for libc++abi built with libgcc.
Implement an undocumented _US_FORCE_UNWIND flag for force
unwinding.
Modified:
Thanks. Committed as http://reviews.llvm.org/rL226819.
http://reviews.llvm.org/D7030
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Thanks. Committed with http://reviews.llvm.org/rL226818.
http://reviews.llvm.org/D7029
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Author: logan
Date: Thu Jan 22 07:25:55 2015
New Revision: 226818
URL: http://llvm.org/viewvc/llvm-project?rev=226818view=rev
Log:
Remove _Unwind_{Get,Set}{GR,IP} from ARM EHABI build.
This commit partially reverts r219629.
This functions are not a part of ARM EHABI specification, and AFAIK,
Recover the previous patch set.
This differential is incorrectly covered by incorrect patch set.
http://reviews.llvm.org/D7031
Files:
include/unwind.h
src/cxa_personality.cpp
Index: include/unwind.h
===
--- include/unwind.h
Thanks. Committed as http://reviews.llvm.org/rL226823.
http://reviews.llvm.org/D7099
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Author: logan
Date: Thu Jan 22 07:40:16 2015
New Revision: 226824
URL: http://llvm.org/viewvc/llvm-project?rev=226824view=rev
Log:
Enable backtrace_test for ARM.
Modified:
libcxxabi/trunk/test/backtrace_test.pass.cpp
Modified: libcxxabi/trunk/test/backtrace_test.pass.cpp
URL:
Update.
http://reviews.llvm.org/D7032
Files:
src/Unwind/Unwind-EHABI.cpp
src/Unwind/Unwind-EHABI.h
src/Unwind/UnwindCursor.hpp
src/Unwind/UnwindLevel1-gcc-ext.c
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Index: src/Unwind/Unwind-EHABI.cpp
Author: logan
Date: Thu Jan 22 07:39:08 2015
New Revision: 226823
URL: http://llvm.org/viewvc/llvm-project?rev=226823view=rev
Log:
Add -funwind-tables to CMAKE_C_FLAGS.
Without -funwind-tables, the compiler won't generate the unwinding
table for these C functions. However, the functions in
Thanks. Committed as rL226822.
http://reviews.llvm.org/D7032
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Hi danalbert, rengolin, jroelofs,
Without -funwind-tables, the compiler won't generate the unwinding
table for these C functions. However, the functions in libunwind,
such as `_Unwind_Backtrace()`, WILL unwind stack to get the backtrace.
http://reviews.llvm.org/D7099
Files:
CMakeLists.txt
OK. Please refer to http://reviews.llvm.org/D7099 for the `-funwind-tables`
change instead.
http://reviews.llvm.org/D7033
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
Make Unwind-EHABI.h become C compatible.
http://reviews.llvm.org/D7031
Files:
src/Unwind/Unwind-EHABI.cpp
src/Unwind/Unwind-EHABI.h
src/Unwind/UnwindCursor.hpp
src/Unwind/UnwindLevel1-gcc-ext.c
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Index:
Comment at: src/Unwind/UnwindLevel1-gcc-ext.c:103
@@ +102,3 @@
+#if LIBCXXABI_ARM_EHABI
+static inline uint32_t readPREL31(const uint32_t *data) {
+ uint32_t base = (uintptr_t) data;
jroelofs wrote:
would be nice to re-use signExtendPrel31 instead.
Done.
Address the review comments.
http://reviews.llvm.org/D7031
Files:
include/unwind.h
src/cxa_personality.cpp
Index: include/unwind.h
===
--- include/unwind.h
+++ include/unwind.h
@@ -63,6 +63,8 @@
static const _Unwind_State
I am using `cmake` and `clang 3.5`.
An alternative solution is to add `-funwind-tables` to CFLAGS.
I guess that it is due to the fact that these C functions are considered as
`nothrow`.
http://reviews.llvm.org/D7033
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Comment at: src/cxa_personality.cpp:1105
@@ +1104,3 @@
+// Check the undocumented force unwinding behavior
+const unsigned _US_FORCE_UNWIND = 8u;
+bool is_force_unwinding = static_castunsigned(state) _US_FORCE_UNWIND;
jroelofs wrote:
Why not add this
Fix the conditional compilation guard.
It seems that I forgot to delete this when I was rebasing.
http://reviews.llvm.org/D7029
Files:
include/unwind.h
src/Unwind/Unwind-EHABI.cpp
Index: include/unwind.h
===
---
Address the review comments.
http://reviews.llvm.org/D7032
Files:
src/Unwind/Unwind-EHABI.cpp
src/Unwind/Unwind-EHABI.h
src/Unwind/UnwindCursor.hpp
src/Unwind/UnwindLevel1-gcc-ext.cpp
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
Index:
@rengolin I have tried to compile with `arm-linux-gnueabihf-g++-4.7` but it
doesn't compile libcxxabi. After applying some hacks, I found that 3 tests
failed. However, they seem unrelated to this change (it looks like the code
generation issue.)
http://reviews.llvm.org/D7099
EMAIL
Comment at: cmake/config-ix.cmake:7
@@ -6,2 +6,3 @@
check_cxx_compiler_flag(-fstrict-aliasing
LIBCXXABI_HAS_FSTRICT_ALIASING_FLAG)
+check_cxx_compiler_flag(-funwind-tables LIBCXXABI_HAS_FUNWIND_TABLES)
check_cxx_compiler_flag(-nodefaultlibs
Address the review comments from Dan Albert.
http://reviews.llvm.org/D7099
Files:
CMakeLists.txt
cmake/config-ix.cmake
Index: CMakeLists.txt
===
--- CMakeLists.txt
+++ CMakeLists.txt
@@ -164,6 +164,7 @@
endif()
endmacro()
Ping?
On Sat, Jan 17, 2015 at 4:02 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi danalbert, rengolin, jroelofs,
The efforts to build LLVM ARM EHABI unwinder breaks the possibility
to build libcxxabi with libgcc. This commit fix the problem by
introducing
Ping?
On Sat, Jan 17, 2015 at 3:58 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi danalbert, rengolin, jroelofs,
I found that declaring _Unwind_{Get,Set}{GR,IP} as the external
functions will break the build with libgcc, which does not define
these functions in their libraries
Hi danalbert, rengolin, jroelofs,
The efforts to build LLVM ARM EHABI unwinder breaks the possibility
to build libcxxabi with libgcc. This commit fix the problem by
introducing the LIBCXXABI_USE_LLVM_UNWINDER conditional compilation
flag and guard the unwinder-specific changes in libc++abi.
Hi danalbert, rengolin, jroelofs,
Implement an undocumented _US_FORCE_UNWIND state for force
unwinding.
http://reviews.llvm.org/D7031
Files:
src/cxa_personality.cpp
Index: src/cxa_personality.cpp
===
--- src/cxa_personality.cpp
Hi danalbert, rengolin, jroelofs,
If libcxxabi is compiled as a shared library, and the
executable references the user-defined personality routines
(e.g. __gxx_personality_v0), then the pointer comparison in
Unwind-EHABI.cpp won't work. This is due to the fact that
the PREL31 will point to the
Hi danalbert, rengolin, jroelofs,
If we compile these *.c files with arm-linux-gnueabi-gcc, gcc won't
generate the unwind table by default. As the result, the _Unwind_Backtrace
can't unwind through itself.
This change fix the problem by renaming *.c to *.cpp.
(p.s. This is the 5/6 patch to fix
Hi danalbert, rengolin, jroelofs,
With previous changes, I am able to run backtrace_test with
shared library build of libc++abi and libunwind.
I believe we can enable this test for ARM now!
(p.s. This is the 6/6 patch to fix the backtrace_test and libgcc support.)
http://reviews.llvm.org/D7034
Hi danalbert, rengolin, jroelofs,
I found that declaring _Unwind_{Get,Set}{GR,IP} as the external
functions will break the build with libgcc, which does not define
these functions in their libraries.
Besides, IMO, since these are not part of the ARM EHABI specification
it will be better to avoid
Author: logan
Date: Thu Jan 8 07:19:07 2015
New Revision: 225442
URL: http://llvm.org/viewvc/llvm-project?rev=225442view=rev
Log:
Frontend: Fix SourceColumnMap assertion failure on non-ascii characters.
If there are some non-ascii character in the input source code, the
column index might be
Thanks! Committed as http://reviews.llvm.org/rL225442 with revised commit
message and test case comments.
http://reviews.llvm.org/D6746
EMAIL PREFERENCES
http://reviews.llvm.org/settings/panel/emailpreferences/
___
cfe-commits mailing list
Ping?
On Sat, Jan 3, 2015 at 2:41 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Ping?
I have found that SourceColMap may trigger assertion failure when the
input source code contains some non-ascii characters. This is due to the
fact that some column is consisted of multiple bytes
. Please
have a look. Thanks!
http://reviews.llvm.org/D6746
Logan
On Tue, Dec 30, 2014 at 8:11 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Ping?
On Thu, Dec 25, 2014 at 12:54 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Richard,
I found that some assertion failure
Ping?
On Thu, Dec 25, 2014 at 12:54 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi Richard,
I found that some assertion failure will be raised from the
lib/Frontend/TextDiagnostic.cpp since the assertion incorrectly compares
the byte index with the number of columns, which is a problem
.) The patch should fix the problem, please have a look. Thanks.
http://reviews.llvm.org/D6746
Sincerely,
Logan
On Sat, Dec 20, 2014 at 5:17 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi rsmith,
http://reviews.llvm.org/D6746
Files:
lib/Frontend/TextDiagnostic.cpp
test/Frontend
Author: logan
Date: Sun Dec 21 08:22:00 2014
New Revision: 224690
URL: http://llvm.org/viewvc/llvm-project?rev=224690view=rev
Log:
Silence warnings in libunwind.
* Remove the embedded directive undefined behavior by moving the
the #ifdef out of the macro arguments. [-Wembedded-directive]
*
Author: logan
Date: Sat Dec 20 02:51:22 2014
New Revision: 224680
URL: http://llvm.org/viewvc/llvm-project?rev=224680view=rev
Log:
Frontend: Fix typo in comments.
Modified:
cfe/trunk/lib/Frontend/TextDiagnostic.cpp
Modified: cfe/trunk/lib/Frontend/TextDiagnostic.cpp
URL:
Hi rsmith,
http://reviews.llvm.org/D6746
Files:
lib/Frontend/TextDiagnostic.cpp
test/Frontend/source-col-map.c
Index: lib/Frontend/TextDiagnostic.cpp
===
--- lib/Frontend/TextDiagnostic.cpp
+++ lib/Frontend/TextDiagnostic.cpp
Hi rafael,
The assertion failure will be raised by TryEmitDefinitionAsAlias()
because it will call Entry-replaceAllUsesWith(Alias) and Entry
might be hold by the AssertingVH, which will become a dangling
pointer after RAUW.
This commit fixes the problem by replacing AssertingVH with
TrackingVH.
, İsmail Dönmez ism...@donmez.ws wrote:
On Thu, Jul 17, 2014 at 8:47 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi,
FYI, from the backtrace, it seems that this is related to:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20140707/109506.html
It seems that the types after
Hi,
FYI, from the backtrace, it seems that this is related to:
http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20140707/109506.html
It seems that the types after the packed parameters are not canonicalized
in some cases.
I am working on this (my patch is incorrect at the moment), but
Ping.
On Mon, Jul 7, 2014 at 9:01 PM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Ping?
This patch tries to fix an assertion failure related to the template alias
with two (or more) parameter packs. For example,
template typename... T struct tuple;
template typename... T struct
... A, typename... B
inline auto test(tupleA... xs, B... ys) - extractA..., B... { }
Please have a look, and feel free to let me know if there is any problem.
Thanks!
Sincerely,
Logan
On Mon, Jun 30, 2014 at 1:10 AM, Logan Chien tzuhsiang.ch...@gmail.com
wrote:
Hi rsmith,
If the template has
LGTM.
http://reviews.llvm.org/D4359
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
Also, I feel that it will be good to detect for `stdexcept` and `typeinfo`
headers to ensure that libcxx include path has been correctly specified.
Comment at: CMakeLists.txt:10
@@ +9,3 @@
+
+ # Rely on llvm-config.
+ set(CONFIG_OUTPUT)
Is there any reason
Author: logan
Date: Tue Jul 1 06:47:10 2014
New Revision: 212097
URL: http://llvm.org/viewvc/llvm-project?rev=212097view=rev
Log:
Remove CleanupHackLevel from CGException.
This patch removes the dead code, and refines the
getEHResumeBlock() slightly.
The CleanupHackLevel was a hack to the old
Thanks. Committed as r212097.
On Tue, Jul 1, 2014 at 2:05 PM, John McCall rjmcc...@apple.com wrote:
On Jun 30, 2014, at 11:31 AM, Reid Kleckner r...@google.com wrote:
This is really a question for John, but I have a feeling that he'll be
happy to see this code go.
Yes, killing the hack
Hi Dan,
Sorry for the late reply. May you send a new patch with file attachment?
Thanks.
Logan
On Tue, Jul 1, 2014 at 9:06 AM, Eric Fiselier e...@efcs.ca wrote:
Hi Daniel,
I applied the patch and build libcxxabi with it on linux64 Ubuntu.
A couple of notes and questions:
- It seems
Hi Albert,
IMO, we can simply use #ifdef to provide different code for ARM EHABI.
There should be similar problem with _Unwind_GetRegionStart(), and I feel
it will be
less straightforward to introduce more and more arguments to scan_eh_tab().
Sincerely,
Logan
http://reviews.llvm.org/D4326
1 - 100 of 250 matches
Mail list logo