On Tue, 3 Apr 2018, Martin Liška wrote:
> I would like to document features I prepared in time frame of GCC 8.
A few observations in addition what Martin (S.) provided (Thanks!):
Index: htdocs/gcc-8/changes.html
===
+ GCOV tool can
On Apr 3, 2018, Alexandre Oliva wrote:
> On Apr 2, 2018, Jason Merrill wrote:
>> On Sat, Mar 31, 2018 at 7:12 AM, Alexandre Oliva wrote:
>>> struct a {
>>> static int const z, i = __builtin_offsetof(struct b { int j; }, j);
>>> b c;
>>> };
>>> int const a::z = __builtin_offsetof(struct d { in
On Apr 3, 2018, Jason Merrill wrote:
> On Tue, Apr 3, 2018 at 11:47 AM, Jason Merrill wrote:
>> On Tue, Apr 3, 2018 at 3:44 AM, Alexandre Oliva wrote:
>>> + if ((complain & tf_conv))
>>> +return false;
>> I think we want to return true.
Yeah, it doesn't work at all returning false.
>
On Apr 3, 2018, Jason Merrill wrote:
> On Tue, Apr 3, 2018 at 3:54 AM, Alexandre Oliva wrote:
>> On Apr 2, 2018, Jason Merrill wrote:
>>
>>> On Sat, Mar 31, 2018 at 4:23 AM, Alexandre Oliva wrote:
On Mar 30, 2018, Jason Merrill wrote:
template
void foo(T t) {
typename
PR libstdc++/85183
* include/std/variant (_Move_assign_base::operator=): Fix incorrect
value categories.
* testsuite/20_util/variant/85183.cc: New.
Tested powerp64le-linux, committed to trunk.
commit eda33cf5c2399e0f25fae17df46efa0b534c778a
Author: Jonathan Wake
Hi again ;)
On 03/04/2018 22:54, Martin Sebor wrote:
+more consistently. Mutually excclusive attribute specifications are
exclusive
Paolo.
Hi,
On 03/04/2018 22:54, Martin Sebor wrote:
+mutually exclusive attribute specifications and hande such conflicts
handle
Thanks,
Paolo.
Oh my,
On 03/04/2018 22:54, Martin Sebor wrote:
+declaration or on distinct declarations of the same entitiy. For
entity
It's getting late here in Italy, I would recommend a systematic check of
the typos.
Paolo.
Hi!
The following testcase is miscompiled, because for vinserti32x4 the expander
creates an incorrect VEC_MERGE selector; it matches what the define_insn
later does, so often it works properly, but when the middle-end sees the
RTL, it can try to simplify it (in this case fwprop1, but in theory als
Am 03.04.2018 um 17:21 schrieb Dominique d'Humières:
Hi Dominique,
The new patch regtest fine now. However as said on IRC this looks as a kludge
made necessary by a questionable (invalid) test.
What I want to avoid is to have first an error and then a warning
for the same thing. If we say th
The attached changes add documentation of some of the options
I worked on for GCC 8.
The links to the GCC 8 manual don't work because there is no
gcc-8 documentation directory. I have checked them by hand
by substituting the GCC 7 directory. (I wonder: would it be
possible to populate the GCC 8
On Mon, 2 Apr 2018, Translation Project Robot wrote:
> 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.
On 04/03/2018 10:26 AM, dave.pa...@oracle.com wrote:
This patch fixes handlng of -Werror=return-type. Currently, even with
the flag specified, return type errors remain warnings.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55976
Function c_finish_return (), when calling pedwarn, should specifi
My recent patch to make sure that we've given an error for failed
template argument conversion caught this latent bug: when we evaluate
__builtin_constant_p within a constexpr function, we don't fold it
until later, at constexpr evaluation time when we have values for the
arguments. Which makes se
Here, because the class has a list constructor, we avoid the shortcut
in build_special_member_call, so we need to use the new backup path
for C++17 copy elision in build_over_call, which means teaching
conv_binds_ref_to_prvalue about this case.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit
On 4/3/18 1:40 PM, H.J. Lu wrote:
> On Tue, Apr 3, 2018 at 11:36 AM, Peter Bergner wrote:
>> gcc/testsuite/
>> PR rtl-optimization/84878
>> * gcc.dg/pr84878.c: New test.
>
> Wrong test filename.
Ooops, thanks for spotting that! Will fix.
Peter
On Mon, Mar 26, 2018 at 4:10 AM, H.J. Lu wrote:
> On Wed, Mar 14, 2018 at 4:41 AM, H.J. Lu wrote:
>> On Wed, Feb 21, 2018 at 3:02 AM, H.J. Lu wrote:
>>> On Wed, Oct 18, 2017 at 5:25 PM, H.J. Lu wrote:
config/plugins.m4 has
if test "$plugins" = "yes"; then
AC_SEARCH_LIBS
On Tue, Apr 3, 2018 at 11:36 AM, Peter Bergner wrote:
> On 4/2/18 9:21 AM, Alexander Monakov wrote:
>> On Tue, 27 Mar 2018, Richard Biener wrote:
>>> If they only appear in the exit/entry block ignoring them should be safe.
>>>
>>> But who knows...
>>
>> Roman and I discussed a related problem a f
On 4/2/18 9:21 AM, Alexander Monakov wrote:
> On Tue, 27 Mar 2018, Richard Biener wrote:
>> If they only appear in the exit/entry block ignoring them should be safe.
>>
>> But who knows...
>
> Roman and I discussed a related problem a few weeks ago, so here's my 2c.
> As I don't have any special D
>This patch adds scan-ltrans-tree-dump.
Please check all error calls to talk about the correct function -- at least
scan-ltrans-tree-dump-times is wrong.
thanks,
>
>Bootstrapped and reg-tested on x86_64.
>
>OK for stage4/stage1 trunk?
>
>Thanks,
>- Tom
In this testcase, we have multiple declarations of malloc. The one in
the "system header" has a non-throwing exception specification, while
the one in user code has none. By default we allow this mismatch when
the old declaration is in a system header, but we mean to prefer the
type of the user d
Now that we partially instantiate generic lambdas, we run into an
issue with C++17 constexpr if: in the testcase, the branches of the
constexpr if become non-dependent when partially instantiated, but the
condition remains dependent. In that situation, we mustn't substitute
into the branches, so w
Hi!
The following testcase ICEs in fixed_parameter_pack_p_1, because parm is
error_mark_node and we assert it is one of the 3 TREE_CODEs we handle.
Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
trunk?
2018-04-03 Jakub Jelinek
PR c++/85147
* pt.c
On Wed, Mar 21, 2018 at 4:14 PM, Marek Polacek wrote:
> On Wed, Mar 21, 2018 at 02:55:36PM -0400, Jason Merrill wrote:
>> On Wed, Mar 21, 2018 at 2:37 PM, Marek Polacek wrote:
>> > On Thu, Mar 15, 2018 at 08:55:59AM -0400, Jason Merrill wrote:
>> >> On Thu, Mar 15, 2018 at 7:49 AM, Marek Polacek
On Tue, Apr 3, 2018 at 12:56 PM, Jason Merrill wrote:
> On Mon, Mar 26, 2018 at 4:01 PM, Jason Merrill wrote:
>>
>> On Mon, Mar 26, 2018 at 2:55 PM, Andreas Schwab
>> wrote:
>> > On Mär 26 2018, Jakub Jelinek wrote:
>> >> On Mon, Mar 26, 2018 at 08:33:41PM +0200, Andreas Schwab wrote:
>> >>> On
On 04/03/2018 05:16 AM, Martin Liška wrote:
Hello.
I would like to document features I prepared in time frame of GCC 8.
Martin
Just a few minor nits:
@@ -113,6 +113,93 @@
possible for the user to have a finer-grained control over the loop
unrolling optimization.
+
+GCOV
On April 3, 2018 6:07:06 PM GMT+02:00, Jakub Jelinek wrote:
>Hi!
>
>As the following testcases show, fold_builtin_expect replaces
>__builtin_expect (x && y, z) with
>__builtin_expect (x, z) && __builtin_expect (y, z)
>and similarly for || instead of &&. It does so without any kind of
>copying of
On April 3, 2018 6:07:17 PM GMT+02:00, Jakub Jelinek wrote:
>Hi!
>
>The bb_uses and bb_defs bitmaps are never used if !*split_p, the
>comments
>even describe why:
> /* For the new created basic block, there is no dataflow info at all.
> So skip the following dataflow update and check. */
>On
This patch fixes handlng of -Werror=return-type. Currently, even with
the flag specified, return type errors remain warnings.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55976
Function c_finish_return (), when calling pedwarn, should specifiy
OPT_Wreturn_type as the diagnostic index if warn_r
On Tue, Apr 3, 2018 at 12:06 PM, Jakub Jelinek wrote:
> From N2965 it is unclear to me what should the trait do on classes with
> incomplete types, but given that __bases already returns an empty pack,
> this patch does the same for __direct_bases.
>
I think it should be ill-formed, i.e. call co
OK.
On Tue, Apr 3, 2018 at 12:06 PM, Jakub Jelinek wrote:
> Hi!
>
> The following testcase ICEs in fixed_parameter_pack_p_1, because parm is
> error_mark_node and we assert it is one of the 3 TREE_CODEs we handle.
>
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
>
OK.
On Tue, Apr 3, 2018 at 12:06 PM, Jakub Jelinek wrote:
> Hi!
>
> If at least one attribute is errorneous, we get error_mark_node instead
> of the attributes list.
>
> The following patch fixes error recovery for that. Bootstrapped/regtested
> on x86_64-linux and i686-linux, ok for trunk?
>
>
OK.
On Tue, Apr 3, 2018 at 12:07 PM, Jakub Jelinek wrote:
> Hi!
>
> When we clear DECL_DECLARED_CONSTEXPR_P because the decl doesn't have
> literal type, we returned immediately, which unfortunately breaks e.g.
> OpenMP gimplification, as required DECL_EXPR is not emitted for VLA vars.
>
> This
On 03.04.2018 19:02, Alexander Monakov wrote:
> On Tue, 3 Apr 2018, Andrey Belevantsev wrote:
>
>> Hello,
>>
>> This issue ended up being fixed the way different from described in the PR.
>> We do not want to walk away from the invariant "zero SCHED_TIMES -- insn
>> is not scheduled" even for boo
Hi!
As mentioned in the PR, GCC (and clang) predefines
{__BYTE_ORDER__,__ORDER_{LITTLE,BIG,PDP}_ENDIAN__}
macros, and {,sys/,machine/}endian.h headers predefine
{,__}{BYTE_ORDER,{LITTLE,BIG,PDP}_ENDIAN}
macros (depending on which target and feature test macros).
elf.c in GCC 8 used __BYTE_ORDER, w
Hi!
The bb_uses and bb_defs bitmaps are never used if !*split_p, the comments
even describe why:
/* For the new created basic block, there is no dataflow info at all.
So skip the following dataflow update and check. */
On the new bb in that case DF_LR_BB_INFO is NULL, so &(NULL->field) is
Hi!
For V64QImode, elt can be 0 to 63, so 1 << elt invokes UB in the compiler
and we get wrong masks.
Bootstrapped/regtested on x86_64-linux and i686-linux, committed to trunk as
obvious.
2018-04-03 Jakub Jelinek
PR target/85169
* config/i386/i386.c (ix86_expand_vector_set):
Hi!
As the following testcases show, fold_builtin_expect replaces
__builtin_expect (x && y, z) with
__builtin_expect (x, z) && __builtin_expect (y, z)
and similarly for || instead of &&. It does so without any kind of
copying of the z argument, which means if it is e.g. statement expression
witho
Hi!
If at least one attribute is errorneous, we get error_mark_node instead
of the attributes list.
The following patch fixes error recovery for that. Bootstrapped/regtested
on x86_64-linux and i686-linux, ok for trunk?
2018-04-03 Jakub Jelinek
PR c++/85140
* name-lookup.c (
Hi!
When we clear DECL_DECLARED_CONSTEXPR_P because the decl doesn't have
literal type, we returned immediately, which unfortunately breaks e.g.
OpenMP gimplification, as required DECL_EXPR is not emitted for VLA vars.
This patch instead of returning immediately just clears it and so that
we don'
Hi!
>From N2965 it is unclear to me what should the trait do on classes with
incomplete types, but given that __bases already returns an empty pack,
this patch does the same for __direct_bases.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2018-04-03 Jakub Jelinek
Hello,
In these PRs we deal with the dependencies we are forced to make between a
debug insn and its previous insn (unless bb head). In the latter case, if
such an insn has been moved, we fixup its movement so it aligns with the
sel-sched invariants. We also carefully adjust seqnos in the case w
On Tue, 3 Apr 2018, Andrey Belevantsev wrote:
> Hello,
>
> This issue ended up being fixed the way different from described in the PR.
> We do not want to walk away from the invariant "zero SCHED_TIMES -- insn
> is not scheduled" even for bookkeeping copies (testing showed it trips over
> assert
Hello,
This issue ended up being fixed the way different from described in the PR.
We do not want to walk away from the invariant "zero SCHED_TIMES -- insn
is not scheduled" even for bookkeeping copies (testing showed it trips over
asserts designed to catch this). Rather we choose merging exprs
Hi,
Option -ftree-loop-distribution is improved and enabled by default at -O3 for
GCC8.
This patch describes the change, is it OK?
Thanks,
binIndex: htdocs/gcc-8/changes.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-8/changes.html,
On Tue, Apr 3, 2018 at 11:47 AM, Jason Merrill wrote:
> On Tue, Apr 3, 2018 at 3:44 AM, Alexandre Oliva wrote:
>> On Apr 2, 2018, Jason Merrill wrote:
>>
>>> On Sat, Mar 31, 2018 at 2:24 AM, Alexandre Oliva wrote:
On Mar 30, 2018, Jason Merrill wrote:
> I don't think we need thi
Hello,
This issues is about the correct order in which we need to call the
routines that fix up the control flow for us.
Best,
Andrey
2018-04-03 Andrey Belevantsev
PR rtl-optimization/83962
* sel-sched-ir.c (tidy_control_flow): Correct the order in which we call
tidy_fallthr
On Tue, Apr 3, 2018 at 3:44 AM, Alexandre Oliva wrote:
> On Apr 2, 2018, Jason Merrill wrote:
>
>> On Sat, Mar 31, 2018 at 2:24 AM, Alexandre Oliva wrote:
>>> On Mar 30, 2018, Jason Merrill wrote:
>>>
I don't think we need this; if arg is overloaded, we take the
type_unknown_p early
On Tue, Apr 3, 2018 at 3:54 AM, Alexandre Oliva wrote:
> On Apr 2, 2018, Jason Merrill wrote:
>
>> On Sat, Mar 31, 2018 at 4:23 AM, Alexandre Oliva wrote:
>>> On Mar 30, 2018, Jason Merrill wrote:
>>> template
>>> void foo(T t) {
>>> typename T::template C u = t;
>>> T::template C (t);
>>
Hello,
This issue is when we cannot correctly make insn tick data for the
unscheduled code (but processed by the modulo scheduler). Fixed by closely
following usual scheduling process as described in the PR.
Best,
Andrey
2018-04-03 Andrey Belevantsev
PR rtl-optimization/83530
OK.
On Tue, Apr 3, 2018 at 5:02 AM, Paolo Carlini wrote:
> Hi,
>
> the reason why we ICE here is very simple: we pass an error_mark_node
> generated in cp_parser_template_parameter_list to rewrite_template_parm
> which ICEs (via get_template_parm_index). I believe it makes sense to
> robustify th
After middle-end changes combine now gets fed different input, from
which it makes different (but just as efficient) code. So remove the
test for particular asm output.
Tested, committing.
Segher
2018-04-03 Segher Boessenkool
gcc/testsuite/
PR target/85126
* gcc.target/pow
Hello,
I will post patches to various recent selective scheduling PRs as replies
to this mail. The patches have been tested on x86-64 (default languages)
and ia64 (c.c++), in case of ppc issues I've checked on the ppc cross
compiler. I will also run the ia64 boostrap with enabled sel-sched but i
Hi Thomas,
> Le 31 mars 2018 à 13:57, Thomas König a écrit :
>
> Hi Dominique,
>
> These have been resolved now - I have removed the invalid code
> from substr_6.f90 (PR85130), and the error is now suppressed
> in the attached patch.
>
> Re-regression-tested. OK for trunk?
>
> Regards
The ne
On 03/02/2018 05:55 PM, Cesar Philippidis wrote:
* config/nvptx/nvptx.c (oacc_bcast_partition): Declare.
One last thing: this variable needs to be reset to zero for every function.
Without this reset, we can generated different code for a function
depending on whether there's another
On Mon, Apr 2, 2018 at 6:34 PM, Martin Sebor wrote:
> Attached is a changed patch that uses "cannot have side effects"
> instead of "can have no side effects."
> +effects it does not make sense for such a function to return @code{void}.
> +Declaring such functions is diagnosed.
Let's make the la
[ was: [og7] vector_length extension part 2: Generalize state
propagation and synchronization ]
On 03/02/2018 05:55 PM, Cesar Philippidis wrote:
- if (oacc_bcast_size < data.offset)
- oacc_bcast_size = data.offset;
The current state of nvptx.c uses this construct a lot, which is ha
On 04/03/2018 03:27 AM, Pedro Alves wrote:
On 04/02/2018 11:34 PM, Martin Sebor wrote:
On 04/02/2018 12:09 AM, Sandra Loosemore wrote:
On 03/27/2018 03:21 PM, Pedro Alves wrote:
On 03/27/2018 09:19 PM, Martin Sebor wrote:
On 03/27/2018 01:38 PM, Pedro Alves wrote:
On 03/27/2018 07:18 PM, Mar
Finishing testing with the full suite on Linux-PPC64, tested partially
on Linux-x64.
2018-04-03 Ville Voutilainen
gcc/cp
PR c++/65923
* parser.c (cp_parser_unqualified_id): Add a new parameter
and check it for the literal diagnostic.
(cp_parser_using_declaration): Adjust.
On 03/04/18 14:40, Christophe Lyon wrote:
Hi,
I've noticed that armv8_2-fp16-move-1.c fails on arm-linux-gnueabi and
arm-none-eabi targets because:
dg-add-options arm_v8_2a_fp16_scalar adds -mfloat-abi=softfp
but we also have
dg-options "-O2 -mfloat-abi=hard
which means the testcase is compiled
Hi,
I've noticed that armv8_2-fp16-move-1.c fails on arm-linux-gnueabi and
arm-none-eabi targets because:
dg-add-options arm_v8_2a_fp16_scalar adds -mfloat-abi=softfp
but we also have
dg-options "-O2 -mfloat-abi=hard
which means the testcase is compiled with both -mfloat-abi=hard
-mfloat-abi=softf
On Tue, Apr 03, 2018 at 02:31:56PM +0200, Martin Liška wrote:
> Hi.
>
> This is second part of the UBSAN. Richard Sandiford help me to wrap the offset
> into poly_uint64. The offset is neither logically signed nor logically
> unsigned.
>
> Patch can bootstrap on x86_64-linux-gnu and survives reg
On Tue, Apr 03, 2018 at 02:30:23PM +0200, Richard Biener wrote:
> On Tue, 3 Apr 2018, Alan Modra wrote:
>
> > On Tue, Apr 03, 2018 at 01:01:23PM +0200, Richard Biener wrote:
> > > On Fri, 30 Mar 2018, Peter Bergner wrote:
> > >
> > > > On 3/29/18 9:35 AM, Alan Modra wrote:
> > > > > On Thu, Nov 1
On Tue, Apr 3, 2018 at 11:39 AM, Martin Jambor wrote:
> Hi,
>
> PR 84947 shows that when we LTO calls with type-mismatches, we can end
> up doing undefined shifts because we try to work with precision of types
> which do not have any.
>
> Fixed basically in the same way as Martin proposed in Bugzi
Hi Gerald,
Yes the Arm trademark is now "Arm" so that should be used when referring to
anything related to the trademark.
(see https://www.arm.com/company/policies/trademarks/guidelines-trademarks)
I left the ARM there in reference to the port in GCC as that seems to be what
we've always called
Hi all,
> I believe this patch is wrong. Please revert. See the PR84762 testcase.
>
I'm happy to revert it given the regression and GCC 8 release being imminent,
but looking at the code again it seems to me that the original code may also not
be fully correct? Particularly I'm wondering what ha
Hi.
This is second part of the UBSAN. Richard Sandiford help me to wrap the offset
into poly_uint64. The offset is neither logically signed nor logically unsigned.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Martin
gcc/ChangeLog:
2018-04-03 Ma
On Tue, 3 Apr 2018, Alan Modra wrote:
> On Tue, Apr 03, 2018 at 01:01:23PM +0200, Richard Biener wrote:
> > On Fri, 30 Mar 2018, Peter Bergner wrote:
> >
> > > On 3/29/18 9:35 AM, Alan Modra wrote:
> > > > On Thu, Nov 16, 2017 at 03:27:01PM +, Tamar Christina wrote:
> > > >> --- a/gcc/expr.c
Hi.
This helps the warning with --save-temps. Doing that one needs to preserve
comments
in preprocessed source file.
Ready for trunk?
Martin
gcc/ChangeLog:
2018-04-03 Martin Liska
PR preprocessor/78497
* gcc.c: Add -C when using -Wimplicit-fallthrough and --save-temps.
gcc
On Tue, Apr 03, 2018 at 01:01:23PM +0200, Richard Biener wrote:
> On Fri, 30 Mar 2018, Peter Bergner wrote:
>
> > On 3/29/18 9:35 AM, Alan Modra wrote:
> > > On Thu, Nov 16, 2017 at 03:27:01PM +, Tamar Christina wrote:
> > >> --- a/gcc/expr.c
> > >> +++ b/gcc/expr.c
> > >> @@ -2769,7 +2769,9 @
It used "vector" but that is not defined. Let's use __vector instead.
Committed.
Segher
2018-04-03 Segher Boessenkool
gcc/testsuite/
* gcc.target/powerpc/pr82015.c: Use __vector instead of vector.
---
gcc/testsuite/gcc.target/powerpc/pr82015.c | 4 ++--
1 file changed, 2 inserti
On Tue, Apr 3, 2018 at 12:48 AM, Jim Wilson wrote:
> On Fri, Mar 23, 2018 at 5:33 AM, Richard Biener
> wrote:
>> I'm leaving the "simple" combiner patch to review by others
>> but for RISC-V you could simply #define SHIFT_COUNT_TRUNCATED to zero
>> to fix the issue. Then add patterns if it turns
Committed.
Richard.
2018-04-03 Richard Biener
* sese.h (recompute_all_dominators): Remove.
Index: gcc/sese.h
===
--- gcc/sese.h (revision 259024)
+++ gcc/sese.h (working copy)
@@ -228,19 +228,6 @@ if_region_get_condit
Hello.
I would like to document features I prepared in time frame of GCC 8.
Martin
Index: htdocs/gcc-8/changes.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-8/changes.html,v
retrieving revision 1.51
diff --unified -r1.51 changes.htm
On 04/02/2018 06:18 PM, Jason Merrill wrote:
About d in the example, I'm not sure how you mean the global namespace
is the current scope; we should be pushed into a when parsing the
initializer for a::z.
this may be related to the brokenness I encountered in 85046. We push
member structs int
On Fri, 30 Mar 2018, Peter Bergner wrote:
> On 3/29/18 9:35 AM, Alan Modra wrote:
> > On Thu, Nov 16, 2017 at 03:27:01PM +, Tamar Christina wrote:
> >> --- a/gcc/expr.c
> >> +++ b/gcc/expr.c
> >> @@ -2769,7 +2769,9 @@ copy_blkmode_to_reg (machine_mode mode, tree src)
> >>
> >>n_regs = (b
Committed.
Richard.
2018-04-03 Richard Biener
PR testsuite/85154
* gcc.dg/vect/vect-95.c: Remove scan for alignment peeling.
Index: gcc/testsuite/gcc.dg/vect/vect-95.c
===
--- gcc/testsuite/gcc.dg/vect/vect-95.c
Hi Bill,
On Sun, Apr 01, 2018 at 08:24:32PM -0500, Bill Schmidt wrote:
> * gcc.target/powerpc/undef-bool.C: New file.
> * gcc.target/powerpc/undef-bool.c: New file.
I'm not sure two files with the same basename will not cause problems.
In either case it is confusing. Please rename on
Hi,
PR 84947 shows that when we LTO calls with type-mismatches, we can end
up doing undefined shifts because we try to work with precision of types
which do not have any.
Fixed basically in the same way as Martin proposed in Bugzilla, the patch
below also updates the comment and dump message to r
On 04/02/2018 11:34 PM, Martin Sebor wrote:
> On 04/02/2018 12:09 AM, Sandra Loosemore wrote:
>> On 03/27/2018 03:21 PM, Pedro Alves wrote:
>>> On 03/27/2018 09:19 PM, Martin Sebor wrote:
On 03/27/2018 01:38 PM, Pedro Alves wrote:
> On 03/27/2018 07:18 PM, Martin Sebor wrote:
>> +Becau
Hi,
the reason why we ICE here is very simple: we pass an error_mark_node
generated in cp_parser_template_parameter_list to rewrite_template_parm
which ICEs (via get_template_parm_index). I believe it makes sense to
robustify the latter like all the other functions involved in
build_deduction
On 03/25/2018 04:30 PM, Thomas Koenig wrote:
[This is take two, the first one was rejected due to size].
Hello world,
the does what the ChangeLog and the Subject say. Regression-tested
on x86_64-pc-linux-gnu.
FTR, this caused PR85166 - "[nvptx, libgfortran] Libgomp fortran tests
using stop
On Apr 2, 2018, Jason Merrill wrote:
> On Sat, Mar 31, 2018 at 7:12 AM, Alexandre Oliva wrote:
>> struct a {
>> static int const z, i = __builtin_offsetof(struct b { int j; }, j);
>> b c;
>> };
>> int const a::z = __builtin_offsetof(struct d { int k; }, k);
>> d e;
>> Since d is visible, I
On Apr 2, 2018, Jason Merrill wrote:
> On Sat, Mar 31, 2018 at 4:23 AM, Alexandre Oliva wrote:
>> On Mar 30, 2018, Jason Merrill wrote:
>> template
>> void foo(T t) {
>> typename T::template C u = t;
>> T::template C (t);
>> T::template C::f (t, u);
>> }
> We should be able to distingu
On Apr 2, 2018, Jason Merrill wrote:
> On Sat, Mar 31, 2018 at 2:24 AM, Alexandre Oliva wrote:
>> On Mar 30, 2018, Jason Merrill wrote:
>>
>>> I don't think we need this; if arg is overloaded, we take the
>>> type_unknown_p early exit, so the code lower down is always dealing
>>> with a singl
Hello!
When x87 rounding bits are changed in x87 control word, the
calculation clears precision bits in the new control word in certain
cases. While precision bits are not used for frndint and fist[p]
instructions, let's play safe and change only rounding bits. Also
note, that for !TARGET_PARTIAL_
86 matches
Mail list logo
