Georg-Johann Lay writes:
> This patch contains some unrelated tweaks
>
> - Supplying no-ldregs variant for andqi3, iorqi3 where a const_int mask
> affects
> only 1 bit
>
> - Some patterns that match situations with zero_extend that can be performed
> with less instructions / register
On Thu, Jul 14, 2016 at 12:14:10PM +0930, Alan Modra wrote:
> (Which is a bug, since the advent of multi-char constraints, and
> potentially affects us with our use of constraint strings like "?*wb".)
On looking into this, I see it is just a documentation bug.
--
Alan Modra
Australia
On Wed, Jul 13, 2016 at 06:46:22AM -0500, Segher Boessenkool wrote:
> On Wed, Jul 13, 2016 at 02:57:01PM +0930, Alan Modra wrote:
> > This is what I've bootstrapped and regression tested on
> > powerpc64le-linux. I'm using Peter's testcases from this thread
> > rather than the one in the original
On 01/07/16 19:15, Martin Sebor wrote:
+ /* Differentiate between an exact and inexact buffer overflow
+or truncation. */
+ const char *fmtstr;
+ if (res->number_chars < 0)
+ fmtstr = info->bounded
+ ? "output may
On 14 July 2016 at 02:23, Ville Voutilainen wrote:
> On 14 July 2016 at 02:14, Ville Voutilainen
> wrote:
>> * include/std/any (_Storage()): Make constexpr and have it
>> initialize _M_ptr.
>
>
> This bit had a stale left-over
On 7/14/2016 06:22, Jeff Law wrote:
> On 07/03/2016 05:56 AM, JonY wrote:
>> This patch allows OBJCOPY to be set by configure. It was missing in
>> Makefile.in.
>>
>> Patch OK?
> With a ChangeLog and verification that some host/target combination
> still builds this is OK.
>
> jeff
>
>
Hi
On 14 July 2016 at 02:14, Ville Voutilainen wrote:
> * include/std/any (_Storage()): Make constexpr and have it
> initialize _M_ptr.
This bit had a stale left-over comment in it. Removed in the updated patch.
P0032R3_2.diff.gz
Description: GNU Zip
Tested on Linux-x64.
2016-07-14 Ville Voutilainen
Implement P0032R3, Homogeneous interface for variant, any and optional,
for the parts concerning any and optional.
* include/std/any (_Storage()): Make constexpr and have it
initialize _M_ptr.
On Wed, Jul 13, 2016 at 03:35:10PM -0700, Bruce Korb wrote:
> Actually, it occurs to me:
>
> On Wed, Jul 13, 2016 at 11:23 AM, Marek Polacek wrote:
> > My current implementation warns here, but the warning can be suppressed
> > by adding /* FALLTHRU */ or [...]
>
> that the
Actually, it occurs to me:
On Wed, Jul 13, 2016 at 11:23 AM, Marek Polacek wrote:
> My current implementation warns here, but the warning can be suppressed
> by adding /* FALLTHRU */ or [...]
that the traditional "lint-ean" spelling is "/* FALLTHROUGH */", so
why would the
On 06/28/2016 09:56 AM, Martin Sebor wrote:
c++/71675 - __atomic_compare_exchange_n returns wrong type for
typed enum
__atomic_compare_exchange_n is documented to return bool but when
its operands are of one of the character types or derived from
it (such as a class enum in C++) it converts the
On 07/03/2016 05:56 AM, JonY wrote:
This patch allows OBJCOPY to be set by configure. It was missing in
Makefile.in.
Patch OK?
With a ChangeLog and verification that some host/target combination
still builds this is OK.
jeff
On 07/06/2016 01:42 PM, Andrew Burgess wrote:
* Richard Sandiford [2016-07-04 09:47:20 +0100]:
Thanks for removing the duplicated error check for unknown predicates.
I think that error gets reported later though, so we should check for
null here:
return pred
On 07/11/2016 05:44 AM, Dominik Vogt wrote:
On Thu, Jul 07, 2016 at 12:57:16PM +0100, Dominik Vogt wrote:
On Wed, Jul 06, 2016 at 02:01:06PM +0200, Bernd Schmidt wrote:
There's one thing I don't quite understand and which seems to have
changed since v1:
On 07/04/2016 02:19 PM, Dominik Vogt
On 06/30/2016 12:53 PM, David Malcolm wrote:
This is a port of the C frontend's r237714 [1] to the C++ frontend:
https://gcc.gnu.org/ml/gcc-patches/2016-06/msg01052.html
offering spelling suggestions for misspelled identifiers, macro names,
and some keywords (e.g. "singed" vs "signed" aka PR
On 06/30/2016 12:53 PM, David Malcolm wrote:
The followup patch implements lookup_name_fuzzy for the C++ frontend.
It's cleaner for that implementation to return a const char *, so this
patch updates the implementation in the C frontend to have the same
return type.
Successfully bootstrapped on
On 07/01/2016 03:56 AM, Bernd Schmidt wrote:
On 07/01/2016 10:34 AM, Dodji Seketeli wrote:
The patch below was bootstrapped and tested on x86_64-linux, without
issues,
The patch looks good to me, thanks.
Alright. I think we need a C frontend maintainer or maybe Jason to
approve it.
Jeff Law writes:
> On 06/26/2016 07:50 PM, Andi Kleen wrote:
>> From: Andi Kleen
>>
>> Autofdo outputs to different dump files and doesn't support some
>> transformation that normal profiling. Add dg-final-scan-autofdo
>> and dg-final-scan-not-autofdo
Tom Tromey writes:
> > "Matthew" == Matthew Fortune writes:
>
> Matthew> Tested on: x86_64-pc-linux-gnu (default and -m32), mips-linux-gnu
> Matthew> mipsel-linux-gnuabi64 with no regressions. The new test only failed
> Matthew> on mips-linux-gnu
Tom Tromey writes:
> > "Matthew" == Matthew Fortune writes:
>
> Matthew> Sorry for the long delay...
>
> No problem.
>
> >> This is ok.
> >> Could you check? I think a -m32 build ought to show it. Maybe your
> >> x86-64 build already did
On 06/27/2016 08:10 PM, Eric Gallager wrote:
The last time I ran ./contrib/download_prerequisites, I already had
previous symlinks set up from a previous run of the script, so `ln`
followed the existing symlinks and created the new ones in the
directories to which the symlinks pointed. This
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
This fixes some basic issues with the profile test cases with autofdo.
- Disable checking for value transformations that autofdo does not
support.
- Disable checking for fixed hit counts which autofdo does not
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
Print some information about indirect call promotions in the afdo dump
file. Do it in the same format as the instrumented profiler so that
the test suite can match on it.
gcc/:
2016-06-26 Andi Kleen
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
autofdo create_gcov creates an extra .imports file. Always remove that
too when running an autofdo test case.
gcc/testsuite/:
* 2016-06-26 Andi Kleen
* lib/profopt.exp
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
not just when verbose. This ensures all command lines needed to
reproduce the test case are always logged
gcc/testsuite/:
* 2016-06-26 Andi Kleen
* lib/profopt.exp
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
Autofdo outputs to different dump files and doesn't support some
transformation that normal profiling. Add dg-final-scan-autofdo
and dg-final-scan-not-autofdo statements to the test suite
so that the test cases
On 06/26/2016 07:50 PM, Andi Kleen wrote:
From: Andi Kleen
The pass to transform gimple based on value profiling runs with autofdo
on, but currently every transformation fails. For indirect calls autofdo
does it on its own, and it doesn't suppport other value profiling.
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
Hi,
This transformation tries two cost models, one estimating the number
of insns to use, one estimating the RTX cost of the transformed sequence.
This is inconsistent with the other cost models used in ifcvt.c and
unneccesary - eliminate the
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
Hi,
This patch is rewrites the cost model for bb_ok_for_noce_multiple_sets
to use the max_seq_cost heuristic added in earlier patch revisions.
As with the previous patch, I've used the new parameters to ensure that
the testsuite is still testing
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
Hi,
This patch clears up the cost model for noce_try_cmove_arith. We lose
the "??? FIXME: Magic number 5" comment, and gain a more realistic cost
model for if-converting memory accesses.
This is the patch that has the chance to cause the largest
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
Hi,
This patch pulls the comparisons between if_info->branch_cost and a magic
number representing an instruction count to a common function. While I'm
doing it, I've documented the instructions that the magic numbers relate
to, and updated them
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
Hi,
This patch removes what is left of branch_cost uses, moving them to use
the new hook and tagging each left over spot with a TODO to revisit them.
All these uses are in rtx costs units, so we don't have more work to do at
this point.
On 06/21/2016 09:50 AM, James Greenhalgh wrote:
On Fri, Jun 03, 2016 at 12:39:42PM +0200, Richard Biener wrote:
On Thu, Jun 2, 2016 at 6:53 PM, James Greenhalgh
wrote:
Hi,
This patch introduces a new target hook, to be used like BRANCH_COST but
with a guaranteed
On 05/25/2016 05:22 AM, Bin Cheng wrote:
Hi, As analyzed in PR68303 and PR69710, vectorizer generates
duplicated computations in loop's pre-header basic block when
creating base address for vector reference to the same memory object.
Not a huge surprise. Loop optimizations generally have a
On 04/14/2016 12:45 AM, Hurugalawadi, Naveen wrote:
Hi,
>> I think we should handle at least INTEGER_CST and SSA_NAME
>> with VRP, and it seems natural to add a VRP check
The check should be added in the tree_single_nonzero_warnv_p
for SSA_NAME case for tree_expr_nonzero_p.
However, for
On 07/13/16 11:14, Kyrill Tkachov wrote:
Hi all,
The most common way to load and store TImode value in aarch64 is to
perform an LDP/STP of two X-registers.
This is the *movti_aarch64 pattern in aarch64.md.
There is a bug in the logic in aarch64_classify_address where it
validates the offset
On 02/22/2016 11:10 AM, ro...@nextmovesoftware.com wrote:
It has been a while since my last contribution. The following patch allows
GCC's optimizers
to more aggressively eliminate and optimize java array bounds checks. The
results are
quite impressive, for example producing a 26%
On 01/30/2016 06:35 AM, Prasad Ghangal wrote:
Hi!
This is my first proposed patch for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=17896. I was willing to
do it using "APPEARS_TO_BE_BOOLEAN_EXPR_P(CODE, ARG)" to check
booleans but gcc doesn't allow (bootstraping fails). Hence I am using
On 07/06/2016 01:30 PM, David Malcolm wrote:
This might be a bit confusing when more tests are added, since
pointer equality is only useful in certain specific cases (e.g.
when you know you're dealing with CONST_INTs or pseudo registers).
How about making ASSERT_RTX_EQ check for rtx_equal_p
On Wed, 13 Jul 2016, Marek Polacek wrote:
On Wed, Jul 13, 2016 at 08:39:28PM +0200, Marc Glisse wrote:
On Wed, 13 Jul 2016, Marek Polacek wrote:
Does "__attribute__((fallthrough));" have any advantages over
"__builtin_fallthrough()"?
Not a strong argument, but compilers that don't know the
On Wed, Jul 13, 2016 at 11:39 AM, Marek Polacek wrote:
> Most likely what you saw was in cxx_pretty_printer::declaration_specifiers.
I only saw it once and, of course, it was once too often. ;)
If you fix it, it would sooth my sensibilities as the fixincludes maintainer,
On Wed, Jul 13, 2016 at 08:39:28PM +0200, Marc Glisse wrote:
> On Wed, 13 Jul 2016, Marek Polacek wrote:
>
> > Does "__attribute__((fallthrough));" have any advantages over
> > "__builtin_fallthrough()"?
>
> Not a strong argument, but compilers that don't know the construct will give
> an error
On Wed, Jul 13, 2016 at 11:30:34AM -0700, Bruce Korb wrote:
> > On Mon, Jul 11, 2016 at 01:36:02PM -0700, Bruce Korb wrote:
> > [[putrid code deleted]]
> >> Does this patch mean that the above got fixed? I mean, if you're
> >> going to fret over linguistic tags to make falling through explicit,
>
On Wed, 13 Jul 2016, Marek Polacek wrote:
Does "__attribute__((fallthrough));" have any advantages over
"__builtin_fallthrough()"?
Not a strong argument, but compilers that don't know the construct will
give an error on the builtin, and just a warning on the attribute (and
ignore the
On 13 July 2016 at 21:25, Daniel Krügler wrote:
> How would you feel about the introduction of an internal trait
> __is_boolean_testable, that would test both is_convertible bool> and is_constructible for now, so that we could
> reuse that at places like
> On Mon, Jul 11, 2016 at 01:36:02PM -0700, Bruce Korb wrote:
> [[putrid code deleted]]
>> Does this patch mean that the above got fixed? I mean, if you're
>> going to fret over linguistic tags to make falling through explicit,
>> it would seem the above code is pretty sore-thumby, yes?
>
> My
2016-07-13 12:05 GMT+02:00 Ville Voutilainen :
> On 13 July 2016 at 01:31, Jonathan Wakely wrote:
>> On 11/07/16 23:41 +0300, Ville Voutilainen wrote:
>>>
>>> @@ -785,41 +785,60 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
>>> }
>>> };
>>>
>>> +
On Mon, Jul 11, 2016 at 01:36:02PM -0700, Bruce Korb wrote:
> I'm curious about this. In the process of developing a code analysis
> tool, I found some GCC code that was, basically:
>
> switch (v) {
> case 0:
> if (e) {
> do_something();
> } else {
> case 1:
> do_something_else();
>
On Mon, Jul 11, 2016 at 04:28:32PM -0400, NightStrike wrote:
> On Mon, Jul 11, 2016 at 3:43 PM, Marek Polacek wrote:
> > But then the [[fallthrough]] attribute was
> > approved for C++17 [1], and that's what has got me to do all this.
> > ...
> > I added a new builtin,
> >
On 22/06/16 22:05 +0200, François Dumont wrote:
Hi
Here is eventually the so long promized patch to introduce Debug
algos similarly to Debug containers.
I'm trying to decide how much benefit this really gives us, and
whether the obfuscation to the code (even more namespaces involved,
and
On 07/13/2016 05:29 PM, Dominik Vogt wrote:
Unfortunately this patch (or whatever got actually committed) has
broken the gcc.target/s390/pr679443.c test case, which is a bit
fishy (see code snippet below). I assign most registers to global
variables and then use some complicated arithmetics
This fixes a conflict between how Parallel Mode has always used the
_GLIBCXX_ASSERTIONS macro and the new meaning we gave it for GCC 6
(enabling the lightweight debug checks).
It doesn't make sense for Parallel Mode to own that macro, and it
might be useful to enable Parallel Mode assertions
Currently the overload of basic_string::replace taking an
initializer_list forwards to the overload taking two const _Char*
arguments, which does some debug checks, then forwards to the overload
taking a const _Char* and a size_type, which repeats the debug checks.
The initializer_list overload
These functions should have been static members all along.
tested x86_64-linux, committed to trunk.
commit 421ccb84e555efcd7e3e0123dc85dd40155f1e82
Author: Jonathan Wakely
Date: Wed Jul 13 17:34:04 2016 +0100
Make __allocated_ptr::_S_raw_ptr static
*
On 07/13/2016 08:47 AM, Martin Liška wrote:
Hello.
As mentioned in [1], one slsr transformation is gone, thus we need to change
expected number
of multiplications.
Ready to be installed?
Thanks,
Martin
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71490#c5
Isn't that a code quality
> --- a/gcc/doc/invoke.texi
> +++ b/gcc/doc/invoke.texi
> @@ -7260,7 +7260,9 @@ any called function. In that case it is not necessary
> to save and restore
> them around calls. This is only possible if called functions are part of
> same compilation unit as current function and they are
Hello.
I accidentally learned that -fipa-ra option is disabled if:
/* Do not use IPA optimizations for register allocation if profiler is active
or port does not emit prologue and epilogue as RTL. */
if (profile_flag || !targetm.have_prologue () || !targetm.have_epilogue ())
Hi all,
The most common way to load and store TImode value in aarch64 is to perform an
LDP/STP of two X-registers.
This is the *movti_aarch64 pattern in aarch64.md.
There is a bug in the logic in aarch64_classify_address where it validates the
offset in the address used
to load a TImode value.
On Fri, Jul 1, 2016 at 11:33 AM, Richard Biener
wrote:
> On Tue, Jun 28, 2016 at 8:18 AM, Bin Cheng wrote:
>> Hi,
>> At the moment, loop niter analyzer depends on simple_iv to understand
>> control induction variable in order to do further niter
Hi,
Dump information of IVOPT has been improved, while scanning string in
gcc.dg/tree-ssa/scev-8.c was left over because it appears as PASS. This patch
changes test string to what should be tested.
Test result checked. This is an obvious change.
Thanks,
bin
gcc/testsuite/ChangeLog
2016-07-12
On 13/07/16 15:59 +0100, Jonathan Wakely wrote:
On 13/07/16 12:12 +0100, Jonathan Wakely wrote:
Also fix confusion between pointer and _CharT*, so that allocators with
fancy pointers work correctly.
The _M_data() function returns pointer, but we were using it where
_CharT* was required. This
On Wednesday 13 July 2016 17:14:52 Christophe Lyon wrote:
> Hi Thomas,
Hi Christophe,
>
> I'm seeing:
> gcc.target/arm/pr42574.c: syntax error in target selector
> "arm_thumb1_ok && { ! arm_thumb1_movt_ok }" for " dg-do 1 compile {
> arm_thumb1_ok && { ! arm_thumb1_movt_ok } } "
Oops. I
On Fri, Feb 19, 2016 at 11:03:02PM +0100, Bernd Schmidt wrote:
> In this PR, we generate unnecessarily bad code for code that
> declares a global register var. Since global regs get added to
> fixed_regs, IRA never considers them as candidates. However, we do
> seem to have proper data flow
On 07/13/2016 04:15 PM, Richard Biener wrote:
> so not exactly a 50/50 prediction. It is predicted by early return
> predictor it seems. If I make the predictor not apply we predict
> it as 50/50 chance. Hmm. Honza?
Yes, following test-case eliminates the early return predictor:
int foo(int
Hi Thomas,
On 13 July 2016 at 12:01, Kyrill Tkachov wrote:
> Hi Thomas,
>
>
> On 07/07/16 15:32, Thomas Preudhomme wrote:
>>
>> Hi Kyrill,
>>
>> Please find an updated version in attachment. Please note I made quite a
>> few
>> other changes as well. The most
On Fri, Jul 1, 2016 at 1:06 PM, Richard Biener
wrote:
> On Tue, Jun 28, 2016 at 8:18 AM, Bin.Cheng wrote:
>> On Tue, Jun 14, 2016 at 1:57 PM, Richard Biener
>> wrote:
>>> On Mon, Jun 13, 2016 at 12:01 PM, Bin Cheng
On 13/07/16 12:12 +0100, Jonathan Wakely wrote:
Also fix confusion between pointer and _CharT*, so that allocators with
fancy pointers work correctly.
The _M_data() function returns pointer, but we were using it where
_CharT* was required. This introduces a new _M_c_str() function which
returns
Hello.
As mentioned in [1], one slsr transformation is gone, thus we need to change
expected number
of multiplications.
Ready to be installed?
Thanks,
Martin
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71490#c5
>From ac130165f6c8166c28227fec2a6fa3afbccadb27 Mon Sep 17 00:00:00 2001
From:
On 06/24/2016 02:41 PM, Evgeny Stupachenko wrote:
Hi,
Fix ICE when IPA-cp and target_clones are applied to the same function.
Is the patch ok for trunk?
Thanks,
Evgeny
2016-06-24 Evgeny Stupachenko
gcc/
* ipa-cp.c (determine_versionability): Do not create
On 07/12/2016 03:31 AM, Martin Jambor wrote:
Hi,
On Fri, Jun 24, 2016 at 01:41:05PM -0700, Evgeny Stupachenko wrote:
Hi,
Fix ICE when IPA-cp and target_clones are applied to the same function.
Is the patch ok for trunk?
I can't approve anything but since I wrote most of IPA-CP, it may
count
Hi!
On Wed, 13 Jul 2016 11:25:46 +0200, I wrote:
> Working on something else regarding the C/C++ OpenACC routine directive,
> I couldn't but untangle [...]
> (Another C/C++ OpenACC routine
> cleanup patch is emerging, depending on this one.)
Here it is; likewise, OK for trunk? (Further cleanup
On Wed, 13 Jul 2016, Alexander Monakov wrote:
> Hi,
>
> On Wed, 13 Jul 2016, Richard Biener wrote:
> > The following adds the ability to transform
> >
> > if (x != 0)
> >x = x / 10;
> >
> > to
> >
> > x = x / 10;
> >
> > as requested by PR. Plus it adds some more ops where such
The following adds a simple ad-hoc (matching other code) way to
perform sinking and merging of common stores to a CFG merger
to the SSA code sinking pass.
On cc1-files (from the GCC 4.7 branch head) this performs
930 such merges out of which 366 are stores with the same value
(and thus require
Hi,
On Wed, 13 Jul 2016, Richard Biener wrote:
> The following adds the ability to transform
>
> if (x != 0)
>x = x / 10;
>
> to
>
> x = x / 10;
>
> as requested by PR. Plus it adds some more ops where such transform
> is possible.
In the bugzilla, you said,
> Only for -Os, it's
On 07/13/2016 12:32 PM, Richard Biener wrote:
> Changelog has a swapped entry.
>
> Ok with fixing that.
> Richard.
Fixed and installed as r238300.
M.
The following adds the ability to transform
if (x != 0)
x = x / 10;
to
x = x / 10;
as requested by PR. Plus it adds some more ops where such transform
is possible.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2016-07-13 Richard Biener
Hi!
On Tue, 12 Jul 2016 13:03:43 -0700, Cesar Philippidis
wrote:
> This patch contains both Jakub's OpenMP and my OpenACC fixes for
> PR71704. For reference, the discussion for the original patches can be
> found here
> Ah, ok. I dove into the code and that indeed seems to be the case.
>
> Why is that function ignoring volatiles but not volatile asms or
> old-style asms? I think the only control-flow transfer asms are
> allowed to make are asm gotos?
You mean
if (gasm *asm_stmt = dyn_cast (t))
if
> As tree merging really replaces trees it has to error on the side of not
> merging while canonical type merging has to error on the side of "merging"
> to make types alias.
OK, then the former won't be sufficient for Ada, there are known cases where
producers and clients of a package cannot
Hi,
This patch requires int32plus and ptr32plus for a couple of tests,
tweaks Wduplicated-cond-3.c to use a smaller constant that fits in
16 bits, and marks one test as too big
for avr.
Committed to trunk.
Regards
Senthil
2016-07-13 Senthil Kumar Selvaraj
Ping x3
On 06/21/2016 02:18 PM, Chung-Lin Tang wrote:
Ping x2
On 2016/6/7 08:03 PM, Chung-Lin Tang wrote:
Ping.
On 2016/5/11 02:57 AM, Bernhard Reutner-Fischer wrote:
On May 9, 2016 4:26:50 PM GMT+02:00, Chung-Lin Tang
wrote:
Hi, this patch resolves an ICE for
On Wed, Jul 13, 2016 at 02:57:01PM +0930, Alan Modra wrote:
> This is what I've bootstrapped and regression tested on
> powerpc64le-linux. I'm using Peter's testcases from this thread
> rather than the one in the original patch submission, because that one
> relies on -O0 not reducing the
On Wed, 13 Jul 2016, Jan Hubicka wrote:
> >
> > I'd rather not expose/change need_fake_edge_p as that has a very
> > specific purpose.
> >
> > Why don't you simply add a call to is_ctrl_altering_stmt on the
> > last stmt of the block in loop_only_exit_p? It's a waste of
> > time doing stuff on
Also fix confusion between pointer and _CharT*, so that allocators with
fancy pointers work correctly.
The _M_data() function returns pointer, but we were using it where
_CharT* was required. This introduces a new _M_c_str() function which
returns a _CharT* instead, and uses that for c_str() and
This patch contains some unrelated tweaks
- Supplying no-ldregs variant for andqi3, iorqi3 where a const_int mask affects
only 1 bit
- Some patterns that match situations with zero_extend that can be performed
with less instructions / register pressure.
- comparing HI against -1
Ok for
On Tuesday 12 July 2016 15:57:41 Kyrill Tkachov wrote:
> On 12/07/16 11:26, Thomas Preudhomme wrote:
> > Hi Kyrill,
>
> Hi Thomas,
>
> > On Friday 20 May 2016 14:22:48 Kyrill Tkachov wrote:
> >> Hi Thomas,
> >>
> >> On 17/05/16 11:14, Thomas Preudhomme wrote:
> >>> Ping?
> >>>
> >>> ***
On 13/07/16 13:05 +0300, Ville Voutilainen wrote:
Ha, that was indeed in just one place.
See below.
I made the above changes and also made converting assignment operators
SFINAE. That SFINAE
seems consistent with how constructors and relops work. And yes, there
are still some members like
Hi!
As discussed before, "offloading compilation is slow; I suppose because
of having to invoke several tools (LTO streaming -> mkoffload -> offload
compilers, assemblers, linkers -> combine the resulting images; but I
have not done a detailed analysis on that)". For this reason it is
beneficial
On Tue, Jul 12, 2016 at 4:54 PM, Martin Liška wrote:
> On 07/12/2016 03:15 PM, Richard Biener wrote:
>> The scan for 1 *n_ after FRE looks still useful. Btw, the testcase
>> doesn't fail for me,
>> we _do_ hoist the division in PRE, just not with -m32 anymore. Can
>> you confirm
On 13 July 2016 at 13:05, Ville Voutilainen wrote:
>> Dunno why this has _Tp const& rather than const _Tp&, could you fix it
>> while you're in the file anyway? It's a bit confusing to have one
>> place using a different style.
>
> Ha, that was indeed in just one
On 13 July 2016 at 01:31, Jonathan Wakely wrote:
> On 11/07/16 23:41 +0300, Ville Voutilainen wrote:
>>
>> @@ -785,41 +785,60 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
>> }
>> };
>>
>> + template
>> +using __optional_relop_t =
>> +
Hi Thomas,
On 07/07/16 15:32, Thomas Preudhomme wrote:
Hi Kyrill,
Please find an updated version in attachment. Please note I made quite a few
other changes as well. The most important one was to add new ARMv8-M Baseline
only alternatives to the two movt insns in order to have non predicable
> Hi,
>
> On Fri, Jun 24, 2016 at 01:41:05PM -0700, Evgeny Stupachenko wrote:
> > Hi,
> >
> > Fix ICE when IPA-cp and target_clones are applied to the same function.
> > Is the patch ok for trunk?
>
> I can't approve anything but since I wrote most of IPA-CP, it may
> count that I am fine with
>
> I'd rather not expose/change need_fake_edge_p as that has a very
> specific purpose.
>
> Why don't you simply add a call to is_ctrl_altering_stmt on the
> last stmt of the block in loop_only_exit_p? It's a waste of
> time doing stuff on every stmt that can only make a difference
> on the
Hi!
Working on something else regarding the C/C++ OpenACC routine directive,
I couldn't but untangle that arcane location_t handling, currently using
a dummy OMP_CLAUSE_SEQ. Along the way, I also updated some comments, and
simplified some code. OK for trunk? (Another C/C++ OpenACC routine
On Tue, 2016-07-12 at 16:35 -0600, Martin Sebor wrote:
> On 07/12/2016 10:12 AM, Manuel López-Ibáñez wrote:
> > On 12/07/16 16:59, Martin Sebor wrote:
> > > You're probably right. I suspect I have a tendency to overuse
> > > the quotes (e.g, the -Wplacement-new warning also quotes the
> > >
On Tue, Jul 12, 2016 at 04:35:50PM -0600, Martin Sebor wrote:
> Based on the gcc.pot file it does look like quoted numbers are
> far less common than unquoted ones (just 10 messages where they
> are quoted vs 528 unquoted).
>
> I've added this as a guideline to the Wiki and assuming no one
>
On Wed, May 18, 2016 at 7:43 PM, Jeff Law wrote:
> On 05/17/2016 06:28 AM, Richard Biener wrote:
>>
>>
>> The following patch addresses PR71104 which shows verify-SSA ICEs
>> after gimplify-into-SSA. The issue is that for returns-twice calls
>> we gimplify register uses in the
On Wed, 13 Jul 2016, Uros Bizjak wrote:
> On Sun, Jul 10, 2016 at 10:12 AM, Uros Bizjak wrote:
> > On Wed, Jul 6, 2016 at 3:18 PM, Richard Biener wrote:
> >
> >>> > 2016-07-04 Richard Biener
> >>> >
> >>> > PR
99 matches
Mail list logo