DanAlbert wrote:
It's over 4 years old so unfortunately I'm probably not much help beyond what's
already written on the old review/commit. At the time we were still using GAS
though, so perhaps that made it more meaningful? If vfpv3 was already the
default FPU for armv7 at the time then it's p
https://github.com/DanAlbert approved this pull request.
https://github.com/llvm/llvm-project/pull/117624
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -311,3 +311,19 @@
// CHECK-X86_64-GCC: Found candidate GCC installation:
{{.*}}i686-linux-android{{[/\\]}}4.9
// CHECK-X86_64-GCC-NEXT: Found candidate GCC installation:
{{.*}}x86_64-linux-android{{[/\\]}}4.9
// CHECK-X86_64-GCC-NEXT: Selected GCC installation:
{{.*}}x86_
DanAlbert wrote:
I thought we'd decided to not do this yet.
https://github.com/llvm/llvm-project/pull/70251
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Author: Dan Albert
Date: 2020-12-02T13:26:28-08:00
New Revision: 0849047860a343d8bcf1f828a82d585e89079943
URL:
https://github.com/llvm/llvm-project/commit/0849047860a343d8bcf1f828a82d585e89079943
DIFF:
https://github.com/llvm/llvm-project/commit/0849047860a343d8bcf1f828a82d585e89079943.diff
LO
Author: danalbert
Date: Fri Mar 29 11:34:25 2019
New Revision: 357296
URL: http://llvm.org/viewvc/llvm-project?rev=357296&view=rev
Log:
[Driver] Use --warn-shared-textrel for Android.
Android does not allow shared text relocations. Enable the linker
warning to detect them by default.
Reviewers:
Author: danalbert
Date: Thu Mar 28 11:08:28 2019
New Revision: 357197
URL: http://llvm.org/viewvc/llvm-project?rev=357197&view=rev
Log:
[Driver] Default Android toolchains to noexecstack.
Android does not support executable stacks.
Reviewers: srhines, pirama
Reviewed By: pirama
Subscribers: cf
Author: danalbert
Date: Thu Feb 21 13:13:03 2019
New Revision: 354622
URL: http://llvm.org/viewvc/llvm-project?rev=354622&view=rev
Log:
[Driver] Fix float ABI default for Android ARMv8.
Summary: Android doesn't regress back to soft float after ARMv7 :)
Reviewers: srhines, pirama
Reviewed By: sr
Author: danalbert
Date: Fri Feb 15 12:31:54 2019
New Revision: 354166
URL: http://llvm.org/viewvc/llvm-project?rev=354166&view=rev
Log:
[Driver] Default all Android ARM targets to NEON.
Summary:
There are an insignificant number of ARM Android devices that don't
support NEON. Default to using NEO
Author: danalbert
Date: Tue Jan 8 14:33:59 2019
New Revision: 350668
URL: http://llvm.org/viewvc/llvm-project?rev=350668&view=rev
Log:
[Driver] Default to -fno-addrsig on Android.
Summary: The Android NDK still uses GNU binutils by default.
Reviewers: srhines, pirama
Reviewed By: srhines
Subs
Author: danalbert
Date: Tue Jan 8 14:31:19 2019
New Revision: 350664
URL: http://llvm.org/viewvc/llvm-project?rev=350664&view=rev
Log:
Android is not GNU, so don't claim that it is.
Reviewers: pirama, srhines
Reviewed By: srhines
Subscribers: kristina, cfe-commits
Differential Revision: https
Author: danalbert
Date: Tue Dec 18 15:29:35 2018
New Revision: 349570
URL: http://llvm.org/viewvc/llvm-project?rev=349570&view=rev
Log:
[Driver] Also obey -nostdlib++ when rewriting -lstdc++.
Reviewers: pirama
Reviewed By: pirama
Subscribers: cfe-commits
Differential Revision: https://reviews.
Author: danalbert
Date: Mon Nov 5 12:57:46 2018
New Revision: 346167
URL: http://llvm.org/viewvc/llvm-project?rev=346167&view=rev
Log:
[Driver] Reland again again: Default Android toolchains to libc++.
Landed more fixes to the compiler-rt Android tests.
Original review was https://reviews.llvm.
Author: danalbert
Date: Mon Oct 22 14:58:22 2018
New Revision: 344963
URL: http://llvm.org/viewvc/llvm-project?rev=344963&view=rev
Log:
Revert "[Driver] Reland again: Default Android toolchains to libc++."
More compiler-rt test bot breakages...
Removed:
cfe/trunk/test/Driver/Inputs/basic_an
Author: danalbert
Date: Mon Oct 22 13:16:21 2018
New Revision: 344946
URL: http://llvm.org/viewvc/llvm-project?rev=344946&view=rev
Log:
[Driver] Reland again: Default Android toolchains to libc++.
Some of the test data went missing last time I tried to submit this,
causing the tests to fail when
Author: danalbert
Date: Fri Oct 19 12:23:01 2018
New Revision: 344806
URL: http://llvm.org/viewvc/llvm-project?rev=344806&view=rev
Log:
Revert "[Driver] Reland: Default Android toolchains to libc++."
This reverts commit 84677d5009d613232d360fda27e6e41fb5cb6700.
Modified:
cfe/trunk/lib/Driver
Author: danalbert
Date: Fri Oct 19 11:06:02 2018
New Revision: 344795
URL: http://llvm.org/viewvc/llvm-project?rev=344795&view=rev
Log:
[Driver] Reland: Default Android toolchains to libc++.
The sanitizer builder that was broken by this should now be fixed.
Original review was https://reviews.ll
Author: danalbert
Date: Fri Oct 12 10:06:31 2018
New Revision: 344367
URL: http://llvm.org/viewvc/llvm-project?rev=344367&view=rev
Log:
[Driver] Add defaults for Android ARM FPUs.
Summary:
Android mandates that devices have at least vfpv3-d16 until
Marshmallow and NEON after that. Still honor the
Author: danalbert
Date: Thu Oct 11 14:28:42 2018
New Revision: 344297
URL: http://llvm.org/viewvc/llvm-project?rev=344297&view=rev
Log:
Revert "[Driver] Default Android toolchains to libc++."
Breaks some of the Android bots because they aren't expecting to need
to explicitly set -stdlib.
This re
Author: danalbert
Date: Thu Oct 11 13:58:43 2018
New Revision: 344296
URL: http://llvm.org/viewvc/llvm-project?rev=344296&view=rev
Log:
[Driver] Default Android toolchains to libc++.
Reviewers: srhines, pirama, EricWF
Reviewed By: srhines
Subscribers: cfe-commits
Differential Revision: https:/
Author: danalbert
Date: Thu Oct 11 13:57:54 2018
New Revision: 344295
URL: http://llvm.org/viewvc/llvm-project?rev=344295&view=rev
Log:
[Driver] Default to `-z now` and `-z relro` on Android.
Summary:
RTLD_LAZY is not supported on Android (though failing to use `-z now`
will work since it is assu
Author: danalbert
Date: Thu Oct 11 13:39:32 2018
New Revision: 344293
URL: http://llvm.org/viewvc/llvm-project?rev=344293&view=rev
Log:
[Driver] Fix --hash-style choice for Android.
Summary:
Android supports GNU style hashes as of Marshmallow, so we should be
generating both styles for pre-M targ
Author: danalbert
Date: Wed May 2 12:38:37 2018
New Revision: 331390
URL: http://llvm.org/viewvc/llvm-project?rev=331390&view=rev
Log:
[Driver] Infer Android sysroot location.
Summary:
Android toolchains include their headers and libraries in a
self-contained directory within the toolchain.
Rev
Author: danalbert
Date: Wed May 2 12:31:01 2018
New Revision: 331389
URL: http://llvm.org/viewvc/llvm-project?rev=331389&view=rev
Log:
[Driver] Obey computed sysroot when finding libc++ headers.
Summary:
A handful of targets will try some default paths if --sysroot is not provided.
If that is th
Author: danalbert
Date: Wed Apr 25 14:26:06 2018
New Revision: 330873
URL: http://llvm.org/viewvc/llvm-project?rev=330873&view=rev
Log:
[Driver] Reland "Android triples are not aliases for other triples."
Fixed directory separators in tests to be compatible with both
Windows and !Windows.
This r
Author: danalbert
Date: Tue Apr 24 15:06:40 2018
New Revision: 330780
URL: http://llvm.org/viewvc/llvm-project?rev=330780&view=rev
Log:
Revert "[Driver] Android triples are not aliases for other triples."
Revering while I diagnose the failures.
This reverts commit 82dc3bf2157da280420f80e654292cb
Author: danalbert
Date: Tue Apr 24 14:18:37 2018
New Revision: 330770
URL: http://llvm.org/viewvc/llvm-project?rev=330770&view=rev
Log:
[Driver] Android triples are not aliases for other triples.
Summary:
Android targets should never use tools/libraries for non-Android
targets or vice versa.
Rev
Author: danalbert
Date: Wed Apr 4 14:28:34 2018
New Revision: 329234
URL: http://llvm.org/viewvc/llvm-project?rev=329234&view=rev
Log:
[Driver] Include the Android multiarch includes.
Summary:
Most Android headers live in a single directory, but a small handful
live in multiarch directories.
Re
Author: danalbert
Date: Wed Feb 21 14:36:51 2018
New Revision: 325733
URL: http://llvm.org/viewvc/llvm-project?rev=325733&view=rev
Log:
[Driver] Generate .eh_frame_hdr for static executables too.
Summary: libgcc won't unwind without an .eh_frame_hdr section.
Reviewers: srhines, chandlerc
Review
Author: danalbert
Date: Wed Feb 7 13:58:48 2018
New Revision: 324534
URL: http://llvm.org/viewvc/llvm-project?rev=324534&view=rev
Log:
Stop using __strtonum_fallback on Android.
Fallback implementations are now provided by bionic when necessary,
which these may conflict with.
Modified:
libc
Author: danalbert
Date: Mon Jan 8 14:57:12 2018
New Revision: 322039
URL: http://llvm.org/viewvc/llvm-project?rev=322039&view=rev
Log:
Revert "Make rehash(0) work with ubsan's unsigned-integer-overflow."
Seems to have broken some tests since I first wrote this a while
back. Will reland after che
Author: danalbert
Date: Mon Jan 8 13:49:12 2018
New Revision: 322031
URL: http://llvm.org/viewvc/llvm-project?rev=322031&view=rev
Log:
Make rehash(0) work with ubsan's unsigned-integer-overflow.
Reviewers: mclow.lists, EricWF
Reviewed By: mclow.lists
Subscribers: cfe-commits
Differential Revi
Author: danalbert
Date: Wed Nov 1 16:43:07 2017
New Revision: 317142
URL: http://llvm.org/viewvc/llvm-project?rev=317142&view=rev
Log:
Revert "[libc++] Don't alias quick_exit if __ANDROID_API__ < 21"
Broke the Darwin build bots.
This reverts commit f56f1bba1ade4a408d403ff050d50e837bae47df.
Mod
Author: danalbert
Date: Wed Nov 1 14:26:06 2017
New Revision: 317125
URL: http://llvm.org/viewvc/llvm-project?rev=317125&view=rev
Log:
[libunwind] Don't use dl_iterate_phdr if __ANDROID_API__ < 21
Summary:
On ARM, dl_iterate_phdr is only implemented in the Android NDK version 21 or
later:
https
Author: danalbert
Date: Wed Nov 1 14:17:56 2017
New Revision: 317124
URL: http://llvm.org/viewvc/llvm-project?rev=317124&view=rev
Log:
[libc++] Don't alias quick_exit if __ANDROID_API__ < 21
Summary:
quick_exit() and at_quick_exit() were introduced in android NDK 21:
https://android.googlesource
r311488 was submitted moments later and should fix this (one of the other
buildbots caught this really quickly).
On Tue, Aug 22, 2017 at 3:09 PM, Adrian Prantl wrote:
>
> > On Aug 22, 2017, at 2:05 PM, Dan Albert via cfe-commits <
> cfe-commits@lists.llvm.org> wrote:
> &g
Author: danalbert
Date: Tue Aug 22 14:16:22 2017
New Revision: 311488
URL: http://llvm.org/viewvc/llvm-project?rev=311488&view=rev
Log:
Degeneralize more tests.
As before, not every platform supports LTO. Make sure the platform
we're targeting is one that supports it (regardless of the *host*
pla
Author: danalbert
Date: Tue Aug 22 14:05:01 2017
New Revision: 311487
URL: http://llvm.org/viewvc/llvm-project?rev=311487&view=rev
Log:
Revert "Revert "Revert "Revert "Fix LLVMgold plugin name/path for non-Linux.
With tests fixed for Windows style paths now that they are going
through path ca
Author: danalbert
Date: Tue Aug 15 16:57:32 2017
New Revision: 310976
URL: http://llvm.org/viewvc/llvm-project?rev=310976&view=rev
Log:
Revert "Add a target for new LTO plugin name tests."
Modified:
cfe/trunk/test/Driver/lto-plugin-darwin.c
cfe/trunk/test/Driver/lto-plugin-linux.c
cfe
Author: danalbert
Date: Tue Aug 15 16:57:34 2017
New Revision: 310977
URL: http://llvm.org/viewvc/llvm-project?rev=310977&view=rev
Log:
Revert "Revert "Revert "Fix LLVMgold plugin name/path for non-Linux."""
Removed:
cfe/trunk/test/Driver/lto-plugin-darwin.c
cfe/trunk/test/Driver/lto-plug
Author: danalbert
Date: Tue Aug 15 15:10:44 2017
New Revision: 310966
URL: http://llvm.org/viewvc/llvm-project?rev=310966&view=rev
Log:
Add a target for new LTO plugin name tests.
Not all targets will use -plugin with -flto. Pick a fixed target so
this works regardless of the default target (rega
Author: danalbert
Date: Tue Aug 15 14:31:17 2017
New Revision: 310960
URL: http://llvm.org/viewvc/llvm-project?rev=310960&view=rev
Log:
Revert "Revert "Fix LLVMgold plugin name/path for non-Linux.""
Summary:
Relanding https://reviews.llvm.org/D35739 which was reverted because
it broke the tests o
Author: danalbert
Date: Mon Aug 14 17:31:44 2017
New Revision: 310903
URL: http://llvm.org/viewvc/llvm-project?rev=310903&view=rev
Log:
Revert "Fix LLVMgold plugin name/path for non-Linux."
Broke a test. Will fix the test and re-land later.
Modified:
cfe/trunk/lib/Driver/ToolChains/CommonArg
Author: danalbert
Date: Mon Aug 14 16:19:38 2017
New Revision: 310895
URL: http://llvm.org/viewvc/llvm-project?rev=310895&view=rev
Log:
Fix LLVMgold plugin name/path for non-Linux.
Summary:
It's only named LLVMgold.so on Linux. Fix the name for Windows and
Darwin.
Also fix the path for Windows s
Author: danalbert
Date: Tue Feb 7 15:04:19 2017
New Revision: 294350
URL: http://llvm.org/viewvc/llvm-project?rev=294350&view=rev
Log:
Use copy.deepcopy instead of doing it manually.
Reviewers: EricWF
Reviewed By: EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org
Author: danalbert
Date: Thu Feb 2 13:44:11 2017
New Revision: 293926
URL: http://llvm.org/viewvc/llvm-project?rev=293926&view=rev
Log:
Avoid implementation defined behavior in a test.
Summary:
num_put::put uses %p for pointer types, but the exact format of %p is
implementation defined behavior f
danalbert accepted this revision.
danalbert added a reviewer: danalbert.
danalbert added a comment.
This revision is now accepted and ready to land.
LGTM
https://reviews.llvm.org/D26385
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://
danalbert added a comment.
> It is defines with a value of 0. This allows you to actually do something
> better, IMO.
Can we stick with undefined? That's historically how things have been, and I'm
sure there's code out there depending on that (I had actually written a test
that would depend on
danalbert added a comment.
> This macro (along with ANDROID) should always be defined for Android targets.
What if only `arm-linux-androideabi` (without a version) is specified? We
should be falling back to the old behavior (don't defined `__ANDROID_API__`)
when that happens since that's what e
This revision was automatically updated to reflect the committed changes.
Closed by commit rL281936: Fix signatures of fallback tow(upper|lower)_l.
(authored by danalbert).
Changed prior to commit:
https://reviews.llvm.org/D24743?vs=71871&id=71876#toc
Repository:
rL LLVM
https://reviews.llv
Author: danalbert
Date: Mon Sep 19 15:42:57 2016
New Revision: 281936
URL: http://llvm.org/viewvc/llvm-project?rev=281936&view=rev
Log:
Fix signatures of fallback tow(upper|lower)_l.
Summary:
These functions take and return wint_t, not int:
http://pubs.opengroup.org/onlinepubs/9699919799/function
danalbert created this revision.
danalbert added reviewers: EricWF, mclow.lists.
danalbert added a subscriber: cfe-commits.
danalbert set the repository for this revision to rL LLVM.
These functions take and return wint_t, not int:
http://pubs.opengroup.org/onlinepubs/9699919799/functions/towupper
This revision was automatically updated to reflect the committed changes.
Closed by commit rL281921: Replace __ANDROID__ with __BIONIC__. (authored by
danalbert).
Changed prior to commit:
https://reviews.llvm.org/D24690?vs=71714&id=71852#toc
Repository:
rL LLVM
https://reviews.llvm.org/D246
Author: danalbert
Date: Mon Sep 19 13:00:45 2016
New Revision: 281921
URL: http://llvm.org/viewvc/llvm-project?rev=281921&view=rev
Log:
Replace __ANDROID__ with __BIONIC__.
Summary:
None of these checks are specific to Android devices. If libc++ was
used with Bionic on a normal Linux system these
danalbert added inline comments.
Comment at: include/__config:90
@@ +89,3 @@
+#if defined(__linux__)
+#include
+#if !defined(__GLIBC_PREREQ)
EricWF wrote:
> Does MUSL libc have a `features.h`? This include was previously guarded by
> if `!defined(_LIBCPP_HAS_MU
danalbert added a comment.
In https://reviews.llvm.org/D24690#545523, @compnerd wrote:
> So, the only thing that Im confused about is where does `__BIONIC__` get
> defined?
It's in Bionic's ``, which gets pulled in via ``.
Comment at: include/__config:340
@@ -339,3 +344,1 @@
danalbert created this revision.
danalbert added reviewers: EricWF, mclow.lists.
danalbert added a subscriber: cfe-commits.
danalbert set the repository for this revision to rL LLVM.
Herald added subscribers: srhines, danalbert, tberghammer.
None of these checks are specific to Android devices. If
danalbert added a comment.
Maybe `__attribute__((noinline))` every function instead?
https://reviews.llvm.org/D23524
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
danalbert added a comment.
The configure-time check LGTM. Android has it for all recent versions (since
JB), but that will cover the case of GB and ICS.
Comment at: src/cxa_exception.cpp:127
@@ +126,3 @@
+// on 32 bit targets.
+ptr = std::malloc(size);
+#endif
-
Author: danalbert
Date: Wed Jun 15 15:20:32 2016
New Revision: 272827
URL: http://llvm.org/viewvc/llvm-project?rev=272827&view=rev
Log:
Add an Android version check for GNU strerror_r.
Summary:
Android didn't gain GNU's strerror_r until Marshmallow. If we're
building libc++ against something olde
This revision was automatically updated to reflect the committed changes.
Closed by commit rL272827: Add an Android version check for GNU strerror_r.
(authored by danalbert).
Changed prior to commit:
http://reviews.llvm.org/D21402?vs=60887&id=60893#toc
Repository:
rL LLVM
http://reviews.llv
danalbert created this revision.
danalbert added reviewers: mclow.lists, EricWF.
danalbert added a subscriber: cfe-commits.
Herald added subscribers: srhines, danalbert, tberghammer.
Android didn't gain GNU's strerror_r until Marshmallow. If we're
building libc++ against something older (we build
This revision was automatically updated to reflect the committed changes.
Closed by commit rL272443: Add `REQUIRES: c++experimental` where appropriate.
(authored by danalbert).
Changed prior to commit:
http://reviews.llvm.org/D21247?vs=60420&id=60421#toc
Repository:
rL LLVM
http://reviews.l
Author: danalbert
Date: Fri Jun 10 17:45:11 2016
New Revision: 272443
URL: http://llvm.org/viewvc/llvm-project?rev=272443&view=rev
Log:
Add `REQUIRES: c++experimental` where appropriate.
Summary:
I haven't added it to all the tests, just those that fail without it
(those that aren't header only).
danalbert created this revision.
danalbert added reviewers: EricWF, mclow.lists.
danalbert added a subscriber: cfe-commits.
I haven't added it to all the tests, just those that fail without it
(those that aren't header only).
http://reviews.llvm.org/D21247
Files:
test/libcxx/experimental/memo
danalbert accepted this revision.
danalbert added a comment.
This revision is now accepted and ready to land.
LGTM. Sorry for the delay in response.
http://reviews.llvm.org/D17416
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.
danalbert added inline comments.
Comment at: docs/UndefinedBehaviorSanitizer.rst:128
@@ +127,3 @@
+ ``-fsanitize=undefined``.
+ - ``-fsanitize=integer``: Checks for undefined or suspicious integer
+ behavior.
rsmith wrote:
> danalbert wrote:
> > Suspicio
danalbert added inline comments.
Comment at: docs/UndefinedBehaviorSanitizer.rst:128
@@ +127,3 @@
+ ``-fsanitize=undefined``.
+ - ``-fsanitize=integer``: Checks for undefined or suspicious integer
+ behavior.
Suspicious here meaning that it also checks u
Yeah, what's the motivation for this? I'd actually prefer that these didn't
work in C because I'd like to know if my build system is broken.
On Nov 6, 2015 03:05, "Joerg Sonnenberger via cfe-commits" <
cfe-commits@lists.llvm.org> wrote:
> On Fri, Nov 06, 2015 at 06:30:12AM -, Eric Fiselier via
danalbert accepted this revision.
danalbert added a comment.
This revision is now accepted and ready to land.
LGTM.
http://reviews.llvm.org/D13331
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo
Author: danalbert
Date: Wed Sep 16 13:10:47 2015
New Revision: 247827
URL: http://llvm.org/viewvc/llvm-project?rev=247827&view=rev
Log:
Add endianness configuration block for GCC.
Previously GCC using libc++ would just leak endian.h for every
include.
Modified:
libcxx/trunk/include/__config
Android has posix_memalign too.
On Sep 5, 2015 12:00, "Joerg Sonnenberger" wrote:
> On Fri, Sep 04, 2015 at 09:28:39PM +, Eric Fiselier via cfe-commits
> wrote:
> > EricWF added a comment.
> >
> > In http://reviews.llvm.org/D12512#237175, @joerg wrote:
> >
> > > Please don't commit this as is
danalbert added a comment.
Android has posix_memalign too.
http://reviews.llvm.org/D12512
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
danalbert added reviewers: mclow.lists, EricWF.
danalbert added a comment.
FYI this was found because it can cause issues when used with GCC.
http://reviews.llvm.org/D12247
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.or
danalbert accepted this revision.
danalbert added a comment.
I think I prefer the haiku style, but I couldn't give you any concrete reasons
for that.
http://reviews.llvm.org/D12129
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://list
danalbert added inline comments.
Comment at: lib/Driver/Tools.cpp:3025
@@ -3024,3 +3024,3 @@
- // Note that these flags are trump-cards. Regardless of the order w.r.t. the
- // PIC or PIE options above, if these show up, PIC is disabled.
+ // Note that this flag is a trump-ca
Because we consider the fact that clang and gcc do this when those
functions are not marked with nonnull by the libc to be a bug. Adding it to
libc++ would just make another place that we have to fix that bug.
What is the objection to using _Nullable instead of __attribute__(nonnull)?
The original
My testing was varied. I could not get GCC or clang to optimize it away for
Linux, but both did for ARM Android.
Regardless, the fact that GCC is already doing this doesn't mean it's
desirable. We end up hunting and fixing bugs from this optimization this
every release, and our time could be bette
danalbert added inline comments.
Comment at: CMakeLists.txt:304
@@ +303,3 @@
+configure_file(
+ include/__config_site.in
+ ${CMAKE_BINARY_DIR}/include/c++/v1/__config_site
Did you forget to upload this?
http://reviews.llvm.org/D11963
___
Yeah, those sound like exactly what we want. Helping people find UB is
good, but optimizing assuming we've fixed all of the UB isn't something we
can do. Our bugs end up being rather permanent so we need to be defensive.
On Tue, Aug 11, 2015 at 1:18 PM, Aaron Ballman
wrote:
> On Tue, Aug 11, 201
>
> Would you be opposed to annotations that tell the programmer they have
> UB in their code, but *do not* effect the code generation?
Not on our end. This would be great.
On Tue, Aug 11, 2015 at 12:56 PM, Aaron Ballman via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> On Tue, Aug 11, 201
danalbert added a comment.
In http://reviews.llvm.org/D11948#221936, @joerg wrote:
> I'm against doing this unconditionally. IMO it creates bugs without
> reasonable compensation. Just because glibc wants to hurt people doesn't mean
> anyone should get hurt.
+1. We don't want this in Android.
82 matches
Mail list logo
