On Wed, May 8, 2019 at 6:28 AM Hongtao Liu wrote:
>
> Any other comments, i'll merge to trunk?
You are not allowed to merge patches without explicit approval from
the relevant maintainer. Please see [1].
However, there will be a problem with the approval of your patch.
AVX512 stuff has its own m
Hi,
Thanks Richard, Segher, Andrew and all.
Segher Boessenkool writes:
> Let me try to answer some of this...
>
> On Tue, May 07, 2019 at 03:31:27PM +0200, Richard Biener wrote:
>> On Mon, 6 May 2019, Jiufu Guo wrote:
>> > This patch implements the optimization in PR77820. The optimization
>>
On Tue, May 7, 2019 at 11:55 PM Jeff Law wrote:
>
> On 5/7/19 3:45 AM, Richard Biener wrote:
> > On Tue, May 7, 2019 at 11:13 AM Aldy Hernandez wrote:
> >>
> >> Hi.
> >>
> >> We seem to have numerous copies of the same EH propagation cleanups
> >> scattered throughout the compiler. The attached
This is https://gcc.gnu.org/ml/gcc-patches/2019-03/msg01299.html with
the fixes Segher requested, plus a few more:
- delete PREFERRED_RELOAD_CLASS changes
- adjust for recent register renumbering
- use defines rather than hard coding register numbers
- flip altivec/float test when dealing with move
This libgo patch by Cherry Zhang fixes the reflect package to
correctly handle direct interface typed receiver in Value.call. A
direct interface type's value method takes a value receiver now.
Don't pass a pointer to the method function. Bootstrapped and ran Go
testsuite on x86_64-pc-linux-gnu.
This patch to the Go frontend by Cherry Zhang avoids using a double
pointer for the value method of a direct interface type. For a direct
interface type T with a value method M, its pointer type (*T)'s method
table includes a stub method of M which takes a (*T) as the receiver
instead of a T. How
Any other comments, i'll merge to trunk?
On Tue, May 7, 2019 at 3:31 PM Hongtao Liu wrote:
>
> On Tue, May 7, 2019 at 3:03 PM Jakub Jelinek wrote:
> >
> > On Tue, May 07, 2019 at 01:38:49PM +0800, Hongtao Liu wrote:
> > > +2019-05-06 H.J. Lu
> > > + Hongtao Liu
> > > +
> > > + P
On Wed, May 8, 2019 at 2:33 AM Uros Bizjak wrote:
>
> On Tue, May 7, 2019 at 8:49 AM Hongtao Liu wrote:
>
> > > > > > > > > > This patch is about to enable support for bfloat16
> > > > > > > > > > which will be in Future Cooper Lake, Please refer to
> > > > > > > > > > https://software.inte
This patch to the Go frontend by Cherry Zhang adds a
-fgo-debug-optimization option to emit optimization diagnostics. This
can be used for testing optimizations. Apply this to the range clear
optimizations of maps and arrays. Bootstrapped and ran Go testsuite
on x86_64-pc-linux-gnu. Committed to
More Doxygenation.
Tested powerpc64le-linux. Committed to trunk.
commit 4fc9ce4407954cf6601e80685c30bb3b7318f025
Author: Jonathan Wakely
Date: Tue May 7 23:58:33 2019 +0100
Improve API docs for Filesystem TS and Networking TS
* include/experimental/bits/fs_path.h: Impro
More Doxygenation.
Tested powerpc64le-linux. Committed to trunk.
commit 591cd4fa016959cad3665719b44a40b87df86729
Author: Jonathan Wakely
Date: Tue May 7 23:33:31 2019 +0100
Improve API docs for std::pair
* include/bits/stl_pair.h: Improve docs.
* include/std
More Doxygenation.
Tested powerpc64le-linux. Committed to trunk.
commit 75601d987e84e686d433b5f96f838e42165ea97c
Author: Jonathan Wakely
Date: Tue May 7 22:33:01 2019 +0100
Improve API docs for and
* doc/doxygen/doxygroups.cc (std::literals): Add documentation for
This change ensures that std::common_type<> is a complete type (LWG
2408), and that std::common_type, std::common_type, and
std::common_type will use program-defined specializations
for std::common_type (LWG 2465).
The implementation of common_type is changed to use
void_t, and the specialization
When fixing 90171 it struck me as undesirable to have so many separate
functions that all needed to know about the definition of a usual
deallocation function. So this patch condenses them into one. I left
destroying_delete_p because it is used by other files as well.
Tested x86_64-pc-linux-gnu,
There are a few things left in the rs6000 port that are unused now
that we do not support old reload anymore. This removes those.
Tested on powerpc64-linux {-m32,-m64}; committing.
Segher
2019-05-07 Segher Boessenkool
* config/rs6000/rs6000-protos.h (rs6000_legitimize_reload_addre
On 5/7/19 3:45 AM, Richard Biener wrote:
> On Tue, May 7, 2019 at 11:13 AM Aldy Hernandez wrote:
>>
>> Hi.
>>
>> We seem to have numerous copies of the same EH propagation cleanups
>> scattered throughout the compiler. The attached patch moves all the
>> logic into one class that allows for easy
2019-05-07 Uroš Bizjak
* config/i386/i386.md (cvt_mnemonic): New mode attribute.
(ashr3_cvt): Merge insn pattern from ashrsi3_cvt and
ashrdi3_cvt using SWI48 mode iterator.
Bootstrapped and regression tested on x86_64-linux-gnu {,-m32}.
Committed to mainline SVN.
Uros.
Index: conf
I don't remember which PR I was looking at when I stashed this change away for
GCC 10; it ended up not being necessary, but definitely seems correct.
Tested x86_64-pc-linux-gnu, applying to trunk.
* pt.c (type_dependent_expression_p): A non-type template parm with
a placeholder ty
* typeck.c (build_static_cast_1): Use cp_build_addr_expr.
For GCC 9 I fixed this bug with a patch to gimplify_cond_expr, but this
function was also doing the wrong thing.
Using build_address does not push the ADDR_EXPR down into the arms of a
COND_EXPR, which we need for proper handling o
Hi Dominique,
Many thanks - I had already found this after replenishing my tree and
regtesting. I don't quite know how it escaped but the fix is obvious.
Amicalement
Paul
On Tue, 7 May 2019 at 09:39, Dominique d'Humières wrote:
>
> Hi Paul,
>
> With your patch, I see
>
> FAIL: gfortran.dg/iso_
On Tue, May 7, 2019 at 8:49 AM Hongtao Liu wrote:
> > > > > > > > > This patch is about to enable support for bfloat16 which
> > > > > > > > > will be in Future Cooper Lake, Please refer to
> > > > > > > > > https://software.intel.com/en-us/download/intel-architecture-instruction-set-extens
On 5/7/19 10:31 AM, Christophe Lyon wrote:
After your commit, I'm seeing an ICE while building glibc headers:
: internal compiler error: Segmentation fault
0xc2eeaf crash_signal
/tmp/2191418_6.tmpdir/aci-gcc-fsf/sources/gcc-fsf/gccsrc/gcc/toplev.c:326
0x151ad0d munge
/tmp/219
I haven't had IA-64 hardware access for a few years, but I've been
trying to help out as much as I can without hardware. Now though, my
RISC-V work is keeping me very busy, and my IA-64 work has dwindled
down to nothing, so it is time for me to resign as an IA-64 maintainer.
Also, Intel has announ
Hello, gentle maintainer.
This is a message from the Translation Project robot.
A revised PO file for textual domain 'gcc' has been submitted
by the Ukrainian team of translators. The file is available at:
https://translationproject.org/latest/gcc/uk.po
(This file, 'gcc-9.1.0.uk.po', has j
Hello, gentle maintainer.
This is a message from the Translation Project robot. (If you have
any questions, send them to .)
A new POT file for textual domain 'gcc' has been made available
to the language teams for translation. It is archived as:
https://translationproject.org/POT-files/gcc
On Mon, May 6, 2019 at 7:24 AM Jiufu Guo wrote:
>
> Hi,
>
> This patch implements the optimization in PR77820. The optimization
> eliminates phi and phi's basic block, if the phi is used only by
> condition branch, and the phi's incoming value in the result of a
> CMP result.
>
> This optimizatio
Let me try to answer some of this...
On Tue, May 07, 2019 at 03:31:27PM +0200, Richard Biener wrote:
> On Mon, 6 May 2019, Jiufu Guo wrote:
> > This patch implements the optimization in PR77820. The optimization
> > eliminates phi and phi's basic block, if the phi is used only by
> > condition br
Great, committed in rev. 270975
Alejandro
> -Original Message-
> From: Richard Sandiford
> Sent: 07 May 2019 17:18
> To: Alejandro Martinez Vicente
> Cc: James Greenhalgh ; GCC Patches patc...@gcc.gnu.org>; nd ; Richard Biener
>
> Subject: Re: [Aarch64][SVE] Vectorise sum-of-absolute-
Alejandro Martinez Vicente writes:
> Thanks for your comments Richard. I think this patch addresses them.
Yeah, this is OK to install, thanks.
Richard
>
> Alejandro
>
>> -Original Message-
>> From: Richard Sandiford
>> Sent: 07 May 2019 15:46
>> To: Alejandro Martinez Vicente
>> Cc: J
I'm backporting these 5 patches to gcc-8.
Marek
2019-02-27 Marek Polacek
PR c++/88857 - ICE with value-initialization of argument in template.
* call.c (convert_like_real): Don't call build_value_init in template.
--- gcc/cp/call.c
+++ gcc/cp/call.c
@@ -7005,7 +7005,8 @@ conve
Thanks for your comments Richard. I think this patch addresses them.
Alejandro
> -Original Message-
> From: Richard Sandiford
> Sent: 07 May 2019 15:46
> To: Alejandro Martinez Vicente
> Cc: James Greenhalgh ; GCC Patches patc...@gcc.gnu.org>; nd ; Richard Biener
>
> Subject: Re: [Aar
On Wed, Nov 28, 2018 at 12:38 PM Richard Sandiford
wrote:
>
> Uros Bizjak writes:
> > diff --git a/gcc/cfgexpand.c b/gcc/cfgexpand.c
> > index 21bdcdaeaa35..691e0c7c1b0b 100644
> > --- a/gcc/cfgexpand.c
> > +++ b/gcc/cfgexpand.c
> > @@ -2981,6 +2981,14 @@ expand_asm_stmt (gasm *stmt)
> >
* doc/xml/manual/intro.xml: Fix DR 2537 and DR 2566 confusion.
Committed to trunk.
commit 99929bcbbb2c8e0234dbdaf11866d2399f1f039e
Author: Jonathan Wakely
Date: Tue May 7 15:41:00 2019 +0100
Fix incorrect DR numbers in libstdc++ manual
* doc/xml/manual/intro.xml:
* include/bits/regex.h: Improve docs.
* include/bits/regex.tcc: Do not document implementation details.
Tested powerpc64le-linux. Committed to trunk.
commit abbe3b64701718dc3327039127a07547391e29d2
Author: Jonathan Wakely
Date: Tue May 7 15:10:12 2019 +0100
Improve API do
On 07/05/19 12:22 +0100, Jonathan Wakely wrote:
On 07/05/19 12:01 +0100, Nina Dinka Ranns wrote:
Ack. I've put the use of _Alloc_traits::is_always_equal within #if
__cplusplus >= 201703L block since it is officially a C++17 feature.
Let me know if you think that's an overkill.
Yes, that's over
On 04/05/19 15:36 +0100, Jonathan Wakely wrote:
On 03/05/19 23:42 +0100, Jonathan Wakely wrote:
On 23/03/17 17:49 +, Jonathan Wakely wrote:
On 12/03/17 13:16 +0100, Daniel Krügler wrote:
The following is an *untested* patch suggestion, please verify.
Notes: My interpretation is that hash
Alejandro Martinez Vicente writes:
> +;; Helper expander for aarch64_abd_3 to save the callers
> +;; the hassle of constructing the other arm of the MINUS.
> +(define_expand "abd_3"
> + [(use (match_operand:SVE_I 0 "register_operand"))
> + (USMAX:SVE_I (match_operand:SVE_I 1 "register_operand")
On May 7, 2019 4:33:08 PM GMT+02:00, "Martin Liška" wrote:
>On 5/7/19 2:56 PM, Richard Biener wrote:
>> But that can use the existing get_hot_bb_threshold since we never
>want
>> to dump -1 in case min_count was never initialized.
>
>Yes. But the function will call:
>
>get_hot_bb_threshold ()
>{
>
On 5/7/19 2:56 PM, Richard Biener wrote:
> But that can use the existing get_hot_bb_threshold since we never want
> to dump -1 in case min_count was never initialized.
Yes. But the function will call:
get_hot_bb_threshold ()
{
if (min_count == -1)
{
gcov_type t = profile_info->sum_max
Hi Nathan,
On Tue, 7 May 2019 at 14:39, Nathan Sidwell wrote:
>
> This patch reimplements the header dependency data structures. We can
> now use a vector class, rather than cut-n-paste 3 sets of bespoke
> C-style array handling. Sadly, simply using vec.h didn't work, so I do
> have one interna
Hi,
I updated the patch after the dot product went in. This is the new covet letter:
This patch adds support to vectorize sum of abslolute differences (SAD_EXPR)
using SVE.
Given this input code:
int
sum_abs (uint8_t *restrict x, uint8_t *restrict y, int n)
{
int sum = 0;
for (int i = 0; i
On Tue, May 07, 2019 at 10:44:34AM +0200, Dominique d'Humières wrote:
> Hi Steve,
>
> > Ping.
>
> AFAICT this has been committed as revision r270495.
>
Whoops, you're correct.
I have a few too many diffs lying around.
--
Steve
On Mon, 6 May 2019, Jiufu Guo wrote:
> Hi,
>
> This patch implements the optimization in PR77820. The optimization
> eliminates phi and phi's basic block, if the phi is used only by
> condition branch, and the phi's incoming value in the result of a
> CMP result.
>
> This optimization eliminate
Another partial merge.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
>From 3daa15805e831d77d5f1420ec6543a67b1870119 Mon Sep 17 00:00:00 2001
From: Richard Guenther
Date: Mon, 21 Jan 2019 13:48:32 +0100
Subject: [PATCH] allow-some-invariant-conds-with-SLP
* tr
On Tue, 7 May 2019, Jakub Jelinek wrote:
> On Tue, May 07, 2019 at 01:50:23PM +0200, Marc Glisse wrote:
> > > And actually it seems that we could optimize the plus1 == plus2 cases
> > > even if HONOR_SIGN_DEPENDENT_ROUNDING (type), because even in fesetenv
> > > (FE_DOWNWARD) mode the testcase pri
On Tue, May 07, 2019 at 01:50:23PM +0200, Marc Glisse wrote:
> > And actually it seems that we could optimize the plus1 == plus2 cases
> > even if HONOR_SIGN_DEPENDENT_ROUNDING (type), because even in fesetenv
> > (FE_DOWNWARD) mode the testcase prints the first two (in all other modes all
> > 4).
On Tue, May 7, 2019 at 2:01 PM Martin Liška wrote:
>
> On 5/6/19 1:35 PM, Richard Biener wrote:
> > On Mon, May 6, 2019 at 10:00 AM Martin Liška wrote:
> >>
> >> Hi.
> >>
> >> The patch is about support of a new GIMPLE expr.
> >>
> >> Patch can bootstrap on x86_64-linux-gnu and survives regressio
On Tue, May 7, 2019 at 2:00 PM Martin Liška wrote:
>
> On 5/6/19 4:02 PM, Richard Biener wrote:
> > On Mon, May 6, 2019 at 9:59 AM Martin Liška wrote:
> >>
> >> On 5/2/19 2:31 PM, Richard Biener wrote:
> >>> On Mon, Apr 29, 2019 at 2:51 PM Martin Liška wrote:
>
> On 4/26/19 3:18 PM, Ri
> When installing gcc 9.1.0 on Solaris 10 with CONFIG_SHELL=/bin/ksh, it
> failed in the same way as originally fixed by
>
> https://gcc.gnu.org/ml/gcc-patches/2016-05/msg00087.html
>
> While the patch still is on the gcc-5 and gcc-6 branches, it has been
> lost (inadvertently, I assume) on
This patch reimplements the header dependency data structures. We can
now use a vector class, rather than cut-n-paste 3 sets of bespoke
C-style array handling. Sadly, simply using vec.h didn't work, so I do
have one internal vector class.
The other change is that, rather than apply quoting o
When installing gcc 9.1.0 on Solaris 10 with CONFIG_SHELL=/bin/ksh, it
failed in the same way as originally fixed by
https://gcc.gnu.org/ml/gcc-patches/2016-05/msg00087.html
While the patch still is on the gcc-5 and gcc-6 branches, it has been
lost (inadvertently, I assume) on trunk befor
On Tue, 7 May 2019 at 12:07, Jonathan Wakely wrote:
>
> On 07/05/19 10:37 +0100, Jonathan Wakely wrote:
> >On 07/05/19 11:05 +0200, Christophe Lyon wrote:
> >>On Sat, 4 May 2019 at 16:36, Jonathan Wakely wrote:
> >>>
> >>>On 03/05/19 23:42 +0100, Jonathan Wakely wrote:
> On 23/03/17 17:49 +00
On 5/6/19 4:02 PM, Richard Biener wrote:
> On Mon, May 6, 2019 at 9:59 AM Martin Liška wrote:
>>
>> On 5/2/19 2:31 PM, Richard Biener wrote:
>>> On Mon, Apr 29, 2019 at 2:51 PM Martin Liška wrote:
On 4/26/19 3:18 PM, Richard Biener wrote:
> On Wed, Apr 10, 2019 at 10:12 AM Martin Li
On 5/6/19 1:35 PM, Richard Biener wrote:
> On Mon, May 6, 2019 at 10:00 AM Martin Liška wrote:
>>
>> Hi.
>>
>> The patch is about support of a new GIMPLE expr.
>>
>> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
>>
>> Ready to be installed?
>
> Can you please avoid using/
On Tue, 7 May 2019, Jakub Jelinek wrote:
On Tue, May 07, 2019 at 09:55:21AM +0200, Jakub Jelinek wrote:
On Tue, May 07, 2019 at 09:48:13AM +0200, Richard Biener wrote:
Will leave the "correctness check" for other folks
but the above is
BTW, as I wanted to be sure about the correctness, I wro
First merge from the vectorize-with-SLP branch.
The limiting is ineffective for catching permute-caused growth because
this issue exists across instances as well. The original exponential
growth has been fixed by making the SLP tree a graph. The permute-caused
growth is limited to O(n^2).
Thi
Hi Iain,
>> > I've just given building gcc a try in an OpenIndiana VM, and get the
>> > following:
>> >
>> > ld: fatal: option -z has illegal argument 'relax=transtls'
>> > ld: fatal: flags processing errors
>> > collect2: error: ld returned 1 exit status
>> >
>> > $ ld --version
>> > ld: Software
On 07/05/19 12:22 +0100, Jonathan Wakely wrote:
On 07/05/19 12:01 +0100, Nina Dinka Ranns wrote:
Ack. I've put the use of _Alloc_traits::is_always_equal within #if
__cplusplus >= 201703L block since it is officially a C++17 feature.
Let me know if you think that's an overkill.
Yes, that's over
On Tue, 7 May 2019 at 12:22, Jonathan Wakely wrote:
>
> I can remove that #if and test and commit the result for you though,
> no need for another revision of the patch.
Thanks ! :)
Best,
Nina
On 07/05/19 12:01 +0100, Nina Dinka Ranns wrote:
Ack. I've put the use of _Alloc_traits::is_always_equal within #if
__cplusplus >= 201703L block since it is officially a C++17 feature.
Let me know if you think that's an overkill.
Yes, that's overkill, we provide is_always_equal unconditionally
Well, not actually measurable but we can save get_continuation_for_phi
calls from translate_vuse_through_block if we do not end up using
the result.
Bootstrap / regtest running on x86_64-unknown-linux-gnu.
Richard.
2019-05-07 Richard Biener
PR tree-optimization/90316
* tree
This is an attempt to improve compile-time for PR90316 further
where PRE (in particular PHI translation) takes too much time
by walking aliases. Currently translate_vuses_through_block
walking isn't in any way limited, the following patch makes
it so and also limits get_continuation_for_phi walk
Ack. I've put the use of _Alloc_traits::is_always_equal within #if
__cplusplus >= 201703L block since it is officially a C++17 feature.
Let me know if you think that's an overkill.
New changelog below. I didn't change the description of
operator+(basic_string&&,basic_string&&) as it's still technic
On 07/05/19 10:37 +0100, Jonathan Wakely wrote:
On 07/05/19 11:05 +0200, Christophe Lyon wrote:
On Sat, 4 May 2019 at 16:36, Jonathan Wakely wrote:
On 03/05/19 23:42 +0100, Jonathan Wakely wrote:
On 23/03/17 17:49 +, Jonathan Wakely wrote:
On 12/03/17 13:16 +0100, Daniel Krügler wrote:
>
> No need for the else {} and thus indenting the rest since the if ()
> arm always returns from the function.
>
> OK with eliding this else { } wrapping.
Ah, right, I changed the function bit too many times :)
Here is updated patch I re-tested and comitted.
* tree-ssa-alias.c (aliasin
On Tue, May 7, 2019 at 11:13 AM Aldy Hernandez wrote:
>
> Hi.
>
> We seem to have numerous copies of the same EH propagation cleanups
> scattered throughout the compiler. The attached patch moves all the
> logic into one class that allows for easy marking of statements and
> automatic cleanup onc
On 07/05/19 07:06 +0200, François Dumont wrote:
Hi
I just prefer to make the tests implementation-agnostic using reserve.
I check that without the patch to initiate the hashtable with 11
buckets I reproduce the failures and that with this patch it is fine.
PR libstdc++/90277
On 07/05/19 11:05 +0200, Christophe Lyon wrote:
On Sat, 4 May 2019 at 16:36, Jonathan Wakely wrote:
On 03/05/19 23:42 +0100, Jonathan Wakely wrote:
>On 23/03/17 17:49 +, Jonathan Wakely wrote:
>>On 12/03/17 13:16 +0100, Daniel Krügler wrote:
>>>The following is an *untested* patch suggesti
On Tue, May 07, 2019 at 09:55:21AM +0200, Jakub Jelinek wrote:
> On Tue, May 07, 2019 at 09:48:13AM +0200, Richard Biener wrote:
> > Will leave the "correctness check" for other folks
> > but the above is
BTW, as I wanted to be sure about the correctness, I wrote a simple program
(below).
And actu
Hi.
We seem to have numerous copies of the same EH propagation cleanups
scattered throughout the compiler. The attached patch moves all the
logic into one class that allows for easy marking of statements and
automatic cleanup once it goes out of scope.
Tested on x86-64 Linux.
OK for trunk?
On Sat, 4 May 2019 at 16:36, Jonathan Wakely wrote:
>
> On 03/05/19 23:42 +0100, Jonathan Wakely wrote:
> >On 23/03/17 17:49 +, Jonathan Wakely wrote:
> >>On 12/03/17 13:16 +0100, Daniel Krügler wrote:
> >>>The following is an *untested* patch suggestion, please verify.
> >>>
> >>>Notes: My in
The following fixes duplicate temporary file usage by lto-wrapper
with -save-temps.
Bootstrapped / tested on x86_64-unknown-linux-gnu, applied.
Richard.
2019-05-07 Richard Biener
PR lto/90369
* lto-wrapper.c (debug_objcopy): Use the original filename
including archi
Hi Steve,
> Ping.
AFAICT this has been committed as revision r270495.
Cheers,
Dominique
Hi Paul,
With your patch, I see
FAIL: gfortran.dg/iso_c_binding_char_1.f90 -O (test for errors, line 8)
FAIL: gfortran.dg/iso_c_binding_char_1.f90 -O (test for errors, line 9)
FAIL: gfortran.dg/iso_c_binding_char_1.f90 -O (test for excess errors)
This is due to a bad location of the e
On Tue, 7 May 2019 at 10:15, Rainer Orth wrote:
>
> Hi Iain,
>
> > On Tue, 19 Feb 2019 at 13:58, Rainer Orth
> > wrote:
> >>
> >> Hi Iain,
> >>
> >> >> Thanks. This will have to wait for
> >> >>
> >> >> [libphobos] Use sections_elf_shared.d on Solaris 11.5 (PR
> >> >> d/88150)
> >> >>
Hi Iain,
> On Tue, 19 Feb 2019 at 13:58, Rainer Orth
> wrote:
>>
>> Hi Iain,
>>
>> >> Thanks. This will have to wait for
>> >>
>> >> [libphobos] Use sections_elf_shared.d on Solaris 11.5 (PR d/88150)
>> >> https://gcc.gnu.org/ml/gcc-patches/2019-01/msg01661.html
>> >
>> > I'll m
On Tue, May 07, 2019 at 09:48:13AM +0200, Richard Biener wrote:
> Will leave the "correctness check" for other folks but the above is
> better written as
>
> + (outer_op (inner_op @0 REAL_CST@1) REAL_CST@2)
> +(if (real_zerop (@1)
> + && real_zerop (@2)
>
> because that gets code-ge
On Tue, 7 May 2019, Jakub Jelinek wrote:
> Hi!
>
> fold_real_zero_addition_p will fold x + (-0.0) or x - 0.0 to x
> when not -frounding-math, but not the rest of the options when
> -fsigned-zeros, and not when -fsignaling-nans.
> If we have (x + 0.0) + 0.0, we can fold that to just x + 0.0 even
>
On Tue, May 7, 2019 at 9:32 AM Rainer Orth
wrote:
>
> Hi Richard,
>
> > On Mon, May 6, 2019 at 10:39 PM Jakub Jelinek wrote:
> >>
> >> On Mon, May 06, 2019 at 08:46:05PM +0200, Richard Biener wrote:
> >> > >Fixed as follows. i386-pc-solaris2.11 bootstrap has completed with
> >> > >this
> >> > >
Hi Richard,
> On Mon, May 6, 2019 at 10:39 PM Jakub Jelinek wrote:
>>
>> On Mon, May 06, 2019 at 08:46:05PM +0200, Richard Biener wrote:
>> > >Fixed as follows. i386-pc-solaris2.11 bootstrap has completed with
>> > >this
>> > >patch, sparc-sun-solaris2.11 is running the testsuite and
>> > >x86_6
On Tue, May 7, 2019 at 3:03 PM Jakub Jelinek wrote:
>
> On Tue, May 07, 2019 at 01:38:49PM +0800, Hongtao Liu wrote:
> > +2019-05-06 H.J. Lu
> > + Hongtao Liu
> > +
> > + PR Target/89750
> > + PR Target/86444
>
> target, not Target. Various people handle these in various scri
On Mon, May 6, 2019 at 10:39 PM Jakub Jelinek wrote:
>
> On Mon, May 06, 2019 at 08:46:05PM +0200, Richard Biener wrote:
> > >Fixed as follows. i386-pc-solaris2.11 bootstrap has completed with
> > >this
> > >patch, sparc-sun-solaris2.11 is running the testsuite and
> > >x86_64-pc-linux-gnu is bui
Hi!
fold_real_zero_addition_p will fold x + (-0.0) or x - 0.0 to x
when not -frounding-math, but not the rest of the options when
-fsigned-zeros, and not when -fsignaling-nans.
If we have (x + 0.0) + 0.0, we can fold that to just x + 0.0 even
when honoring signed zeros, and IMNSHO even when honori
On Tue, May 07, 2019 at 01:38:49PM +0800, Hongtao Liu wrote:
> +2019-05-06 H.J. Lu
> + Hongtao Liu
> +
> + PR Target/89750
> + PR Target/86444
target, not Target. Various people handle these in various scripts,
so it is better to use consistency and exact spelling of the cate
84 matches
Mail list logo