Hi Segher,
on 2020/12/15 下午10:40, Segher Boessenkool wrote:
> Hi Ke Wen,
>
> On Tue, Dec 15, 2020 at 03:53:29PM +0800, Kewen.Lin wrote:
>> on 2020/12/15 上午2:51, Segher Boessenkool wrote:
>>> On Wed, Dec 02, 2020 at 05:44:24PM +0800, Kewen.Lin wrote:
--- a/gcc/config/rs6000/rs6000.c
+++
On Tue, 15 Dec 2020, Jakub Jelinek wrote:
Hi!
The following patch teaches the bswap pass to handle for small (2/4/8 byte
long) vectors a CONSTRUCTOR by determining if the bytes of the constructor
come from non-vector sources and are either nop or bswap and changing the
CONSTRUCTOR in that case
On 12/16/20 8:56 AM, Iain Sandoe via Gcc-patches wrote:
Rainer Orth wrote:
Hi Pierre-Marie,
This fixes an issue when setting the baud rate. The baud rate is set
using the cfsetospeed and cfsetispeed system calls. The code is using
speed_t for clarity. The non-blocking status is only reset wh
on 2020/12/16 下午2:35, Stott Graham via Gcc-patches wrote:
> If any ---checking options used
>
The related PR PR98311[1] has been filed by David.
BR,
Kewen
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98311
Hi,
When I was investigating unsigned int vec_init issue on Power,
I happened to find there seems something we can enhance in how
combine pass invalidate last_set (set last_set_invalid nonzero).
Currently we have the check:
if (!insn
|| (value && rsp->last_set_table_tick >= label
Tamar Christina writes:
> Hi Richard,
>
> Here's the split off complex add.
>
> Bootstrapped Regtested on aarch64-none-linux-gnu and no issues.
> Checked with armv8-a+sve2+fp16 and no issues. Note that due to a mid-end
> limitation SLP for SVE currently fails for some permutes. The tests have
>
The build suffers from the static initialization order fiasco:
==30085== Invalid read of size 4
==30085==at 0x1D451CD: hash_table_mod1 (hash-table.h:344)
==30085==by 0x1D451CD: hash_table::mem_location_hash,
vec_usage*,
simple_hashmap_traits::mem_location_hash>,
vec_usage*> >::hash_ent
Martin Liška wrote:
The build suffers from the static initialization order fiasco:
==30085== Invalid read of size 4
==30085==at 0x1D451CD: hash_table_mod1 (hash-table.h:344)
==30085==by 0x1D451CD:
hash_table::mem_location_hash,
vec_usage*,
simple_hashmap_traits::mem_location_hash
On 12/16/20 10:08 AM, Iain Sandoe wrote:
This will break bootstrap on targets without init_priority (e.g. Darwin)
All right, based on the chat on IRC, I suggest using C static constructor
with corresponding priority.
Ready to be installed?
Thanks,
Martin
>From 4ecf4f8df8d5b6b75cfe7cc5df90176e0
On 12/16/20 1:19 AM, Jeff Law wrote:
OK.  Please consider a more through description in invoke.texi
though.  What unsupported feature are we warning about and what are the
consequences of not supporting that feature.
Done that and installed the patch.
Thank you for the review,
Martin
On Wed, Dec 16, 2020 at 10:20:09AM +0100, Martin Liška wrote:
> So vec_mem_desc is not initialized before a static member in module.cc.
> We can fix it by using constructor attribute.
>
> gcc/ChangeLog:
>
> * vec.c (init_vec_mem_desc): New function.
> (vec_prefix::register_overhead):
On 12/16/20 10:26 AM, Jakub Jelinek wrote:
I think this needs to be guarded based on which compiler is used to compile
GCC.
Do you mean basing that on __GNUC__?
Perhaps we could say that we don't support
--enable-gather-detailed-mem-stats when the compiler isn't built by GCC (or
other compil
Hi Jakub,
> On Wed, Dec 16, 2020 at 10:20:09AM +0100, Martin Liška wrote:
>> So vec_mem_desc is not initialized before a static member in module.cc.
>> We can fix it by using constructor attribute.
[...]
>> + of all static variables. */
>> +
>> +static void
>> +__attribute__((constructor (101))
Hi Iain,
> Rainer Orth wrote:
>
>> Hi Pierre-Marie,
>>
>>> This fixes an issue when setting the baud rate. The baud rate is set
>>> using the cfsetospeed and cfsetispeed system calls. The code is using
>>> speed_t for clarity. The non-blocking status is only reset when Block is
>>> True. And seri
On 12/16/20 10:38 AM, Rainer Orth wrote:
Hi Jakub,
On Wed, Dec 16, 2020 at 10:20:09AM +0100, Martin Liška wrote:
So vec_mem_desc is not initialized before a static member in module.cc.
We can fix it by using constructor attribute.
[...]
+ of all static variables. */
+
+static void
+__attr
Nathan Sidwell writes:
> Apparently 'var+=...' is not a dash thing. Fixed thusly.
>
> * config.m4: Avoid non-dash idiom
> * configure: Rebuilt.
>
> pushed (2 patches, because I didn't look carefully enough the first time)
Thanks. I think the other uses of += need the same treatme
Rainer Orth wrote:
Rainer Orth wrote:
Hi Pierre-Marie,
This fixes an issue when setting the baud rate. The baud rate is set
using the cfsetospeed and cfsetispeed system calls. The code is using
speed_t for clarity. The non-blocking status is only reset when Block is
True. And serial block
On Wed, Dec 16, 2020 at 04:19:12PM +0800, Kewen.Lin wrote:
> on 2020/12/15 下午10:40, Segher Boessenkool wrote:
> > I mean do
> >
> > rtx tmp = force_reg (inner_mode, XVECEXP (vals, 0, i));
> > op[i] = simplify_gen_subreg (SImode, tmp, inner_mode, 0);
> >
> > If that works (also
On Tue, Dec 15, 2020 at 7:11 PM Jakub Jelinek wrote:
>
> On Tue, Dec 15, 2020 at 06:10:57PM +0800, Hongtao Liu via Gcc-patches wrote:
> > --- a/gcc/config/i386/i386.c
> > +++ b/gcc/config/i386/i386.c
> > @@ -18187,21 +18187,67 @@ ix86_gimple_fold_builtin (gimple_stmt_iterator
> > *gsi)
> >
On Tue, 15 Dec 2020, Jeff Law wrote:
> > @@ -1942,7 +1942,7 @@ gen_divdf3_cc (rtx operand0 ATTRIBUTE_UN
> > gen_rtx_DIV (DFmode,
> > operand1,
> > operand2),
> > - const0_rtx)),
> > + CONST_DOUBLE_ATOF ("0", VOIDmode))),
> > gen_rtx_SET (operand0,
> > gen_rtx_DIV (D
On Mon, Dec 14, 2020 at 04:23:48PM -0600, Pat Haugen wrote:
> Prefixed instructions should not have their length explicitly set to '8'. The
> function get_attr_length() will adjust the length appropriately based on the
> value of the "prefixed" attribute.
>
> Bootstrap/regtest on powerpc64le (Po
On 12/15/20 8:42 PM, Jeff Law wrote:
Arguably not a bugfix. But I think this is reasonable, even during stage3.
Thank you.
OK by me, but give Richi and Jakub a couple days if they object due to
it not being a bugfix.
To be honest it's only removal of a dead code. Hope it's fine to instal
> Przemyslaw Wirkus writes:
> > Hi,
> >
> > Recent 'support SVE comparisons for unpacked integers' patch extends
> > operands of define_expands from SVE_FULL to SVE_ALL. This causes an
> > ICE hence this PR patch.
> >
> > This patch adds this relaxation for:
> > + reduc__scal_ and
> > + arch64_pre
Hello all,
On Wed, Dec 16, 2020 at 9:31 AM Martin Liška wrote:
> I've just create PR for it:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98312
Thank you. I can reproduce the issue: at this point I suspect that my
testing hasn’t detected this because of some inconsistency with my
incremental b
Hi Harald,
ICE-on-invalid: testcase by Arseny, draft patch by Steve,
slightly polished and regtested by me on x86_64-pc-linux-gnu.
OK for master.
Thanks for the patch!
Best regards
Thomas
"Maciej W. Rozycki" writes:
> On Tue, 15 Dec 2020, Jeff Law wrote:
>
>> > @@ -1942,7 +1942,7 @@ gen_divdf3_cc (rtx operand0 ATTRIBUTE_UN
>> >gen_rtx_DIV (DFmode,
>> >operand1,
>> >operand2),
>> > - const0_rtx)),
>> > + CONST_DOUBLE_ATOF ("0", VOIDmode))),
>> >gen_rtx_SET
Przemyslaw Wirkus writes:
> > This is a bug in the vectoriser: the vectoriser shouldn't generate
> > IFN_REDUC_MAX calls that the target doesn't support.
> >
> > I think the problem comes from using the wrong interface to get the index
> > type for a COND_REDUCTION. vectorizable_reduction has:
>
On Tue, Dec 15, 2020 at 4:44 PM Jeff Law wrote:
>
>
>
> On 12/15/20 10:30 AM, H.J. Lu wrote:
> > When definitions marked with used attribute and unmarked definitions are
> > placed in the section with the same name, switch to a new section if the
> > SECTION_RETAIN bit doesn't match.
> >
> > gcc/
Before CWG DR 1955 the controlling expression for an #elif must be
syntactically correct, meaning this won't compile with C++11 compilers
such as gcc 4.8:
The solution is to define __has_include(X) as 0 for compilers that don't
support it.
The second problem is that when is found, it is used
wit
Jonathan was kind enough to test this fix. It doesn't address all the
solaris issues, but does get us further.
Solaris' sys/socket uses the poisoned bcopy identifier, so we must
#include it before poisoning. The inclusion happens in cody.hh, so we
preemptively copy a bit of cody's inclusion lo
This makes the code more explicit for the sake of static analyzers.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-valuer.adb (Scan_Decimal_Digits): Tweak overflow test.
(Scan_Integral_Digits): Likewise.diff --git a/gcc/ada/libgnat/s-valuer.adb b/gcc/ada/l
This adds some OS constants that are needed to control the serial port
on GNU/Linux.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* s-oscons-tmplt.c: Add some OS constants.diff --git a/gcc/ada/s-oscons-tmplt.c b/gcc/ada/s-oscons-tmplt.c
--- a/gcc/ada/s-oscons-tmplt.c
+++ b/
Unit Ada.Text_IO.Fixed_IO is now generic, which makes it necessary to
exclude the generic body from SPARK explicitly because it uses type
Long_Long_Float which is not supported in GNATprove. This ensures that
the unit can be instantiated from SPARK code.
Tested on x86_64-pc-linux-gnu, committed on
Reject contracts Contract_Cases, Test_Case and Subprogram_Variant whose
expression is either "null", "(null record)" or has extra parentheses.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch13.adb (Analyze_Aspect_Specifications): Add a codefix
for extra parent
GNAT.Expect.Non_Blocking_Spawn executes memory allocation/deallocation
after calling fork on the child process side. The libgmem library could
write to the same gmem.out file in the parent and child processes
simultaneously before the child process would load and initialize the
new executable file.
There is a typo in the formula to compute Maxpow exposed in the comment.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-powtab.ads (Maxpow): Use explicit formula in comment.diff --git a/gcc/ada/libgnat/s-powtab.ads b/gcc/ada/libgnat/s-powtab.ads
--- a/gcc/ada/libg
The example for Test_Case pragma was rejected by the compiler; now
fixed.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* doc/gnat_rm/implementation_defined_pragmas.rst
(Test_Case): Change integer to float literals.
* gnat_rm.texi: Regenerate.diff --git a/gcc
When System.Memory implementation is using Debug_Pools, Common ATCB's
uninitialized Global_Task_Lock_Nesting component may be used.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnarl/s-tporft.adb (Register_Foreign_Thread): Set
Global_Task_Lock_Nesting before us
Simplify membership test on N_Procedure_Call_Statement and
N_Function_Call using the collective N_Subprogram_Call subtype. Cleanup
only; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch6.adb, exp_util.adb, sem_ch4.adb, sem_disp.adb,
sem
The final computation now needs to be protected against artificial
underflow when the value is very small.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-valrea.adb (Maxexp32): New constant array.
(Maxexp64): Likewise.
(Maxexp80): Likewise.
Local variables in parsing of formal object declarations take only
positive values, so their types can be narrowed to Pos. Cleanup only;
semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* par-ch12.adb (P_Formal_Object_Declarations): Refine types to
It is not leaking on Linux because the memory heap is reinitialized when
the new executable is loaded into the child process after calling fork.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/g-expect.adb (Non_Blocking_Spawn): Deallocate elements
on Arg_List
Most of the errors of the form "(Ada 83) ..." don't have a colon; fix
inconsistencies.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* par-ch3.adb (P_Modular_Type_Definition): Remove colon from
error message.
* sem_ch11.adb (Check_Duplication): Likewise.
Fix mismatch in iterating over a range for Ada 2012 and referencing an
array for Ada 95. This didn't affect the behaviour, because the
referenced boolean flags in both arrays were the same for the iterated
subrange.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* impunit.adb
Simplify membership test on N_Delay_Relative_Statement and
N_Delay_Until_Statement using the collective N_Delay_Statement subtype.
Cleanup only; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch9.adb, sem_warn.adb: Simplify membership test.diff
Rename it to IS_Scope so that calling scope (node) under gdb is
unambiguous, so ease debugging sessions.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* ali.ads, ali.adb, bindo-writers.adb, lib-writ.adb (Scope):
Renamed to IS_Scope.diff --git a/gcc/ada/ali.adb b/gcc/
This adds complete postconditions and contract cases to subprograms in
Ada.Strings.Fixed.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/a-strfix.ads: Add postconditions and contract cases to
subprograms.
patch.diff.gz
Description: application/gzip
A loop parameter specification over a container is rewrtten as an
iterator specification, so that expansion can use the Iterator
opwrations declared for the container type. If an Ada2020 interator
filter appears on the loop_parameter_specification it must be transfered
to the iterator, and preanaly
Implement symbolic tracebacks on armhf-linux. It's an ELF target so
just a matter of some bookkeeping changes.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-objrea.ads (Object_Arch): Add ARM enum
* libgnat/s-objrea.adb (Initialize): Add EM_ARM case.
On 10/1/19 10:58 AM, Alexandre Oliva wrote:
+ if (opts->x_flag_asynchronous_unwind_tables == 2)
+ opts->x_flag_unwind_tables
+ = opts->x_flag_asynchronous_unwind_tables = 1;
Hello.
Spotted that when I tried to make x_flag_asynchronous_unwind_tables a boolean
flag.
Anyway t
David reported an AIX problem, which I think this addresses. I've not
looked at pr98314 yet but there's certainly overlap
This fixes installers that don't understand -p.
c++tools/
(install): Do not use -p, use mkinstalldirs.
(clean): Fix typo.
pushing to trunk
--
Natha
On Wed, Dec 16, 2020 at 1:18 PM Pierre-Marie de Rodat
wrote:
> Thank you. I can reproduce the issue: at this point I suspect that my
> testing hasn’t detected this because of some inconsistency with my
> incremental builds. The actual correction is a few patches later in my
> porting queue: I’m ab
Hi Pierre-Marie,
> On Wed, Dec 16, 2020 at 1:18 PM Pierre-Marie de Rodat
> wrote:
>> Thank you. I can reproduce the issue: at this point I suspect that my
>> testing hasn’t detected this because of some inconsistency with my
>> incremental builds. The actual correction is a few patches later in m
Libstdc++ has been using __STRICT_ANSI__ for years, e.g. to decide
whether or not std::is_integral<__int128> should be true.
gcc/ChangeLog:
* doc/cpp.texi (Common Predefined Macros): Mention that
__STRICT_ANSI__ is used by libstdc++ too.
OK for trunk?
I'd also like to tell users
On Wed, 16 Dec 2020, Richard Sandiford wrote:
> >> Presumably you can't use CONST0_RTX (mode) here? Assuming that's the
> >> case, then this is fine.
> >
> > Well, `mode' is VOID for simplicity and to match what the middle end
> > presents as an operand to the COMPARE operation, as also shown w
On Wed, Dec 16, 2020 at 2:18 PM Martin Liška wrote:
>
> On 10/1/19 10:58 AM, Alexandre Oliva wrote:
> > + if (opts->x_flag_asynchronous_unwind_tables == 2)
> > + opts->x_flag_unwind_tables
> > + = opts->x_flag_asynchronous_unwind_tables = 1;
>
> Hello.
>
> Spotted that when I tried
Hi!
As mentioned in the PR, shrink-wrapping disqualifies for prologue
placement basic blocks that have EDGE_CROSSING incoming edge.
I don't see why that is necessary, those edges seem to be redirected
just fine, both on x86_64 and powerpc64. In the former case, they
are usually conditional jumps
On Wed, Dec 16, 2020 at 5:05 AM Jakub Jelinek wrote:
>
> On Wed, Dec 02, 2020 at 05:15:21AM -0800, H.J. Lu via Gcc-patches wrote:
> > gcc/
> >
> > PR middle-end/93195
> > PR middle-end/93197
> > * configure.ac (HAVE_GAS_SECTION_LINK_ORDER): New. Define 1 if
> > the assembl
This adds a test to compare the performance of std::atomic_flag with
similar operations on std::atomic_uchar and std::atomic_int.
libstdc++-v3/ChangeLog:
PR libstdc++/46447
* testsuite/performance/29_atomics/atomic_flag.cc: New test.
Tested powerpc64le-linux. Committed to trunk.
This fixes a bug caused by a mismatch between the macros defined by
when GCC is built and the macros defined by when
users include . If the user code is compiled with
_XOPEN_SOURCE defined to 500 or 600, Darwin suppresses the
ENOTRECOVERABLE and EOWNERDEAD macros, which are not defined by SUSv3
(
Currently the , and
headers can be included in C++98 and C++11 modes,
but gives errors. With this change they can be included, but define
nothing.
libstdc++-v3/ChangeLog:
PR libstdc++/98319
* include/experimental/random: Only define contents for C++14
and later.
On Wed, Dec 16, 2020 at 05:36:04AM -0800, H.J. Lu wrote:
> gas is correct.
>
> > comdat or for retain), then we should do something like the following
> > untested change, but if it is gas bug, it should be fixed there.
> >
> > 2020-12-15 Jakub Jelinek
> >
> > * varasm.c (default_elf_as
module.cc has a static initializer that ends up in a circular
dependency when detailed mem stats are enabled. This removes the need
for that initializer to be dynamic, and we punt to the lazy
initializing we already had inside the object in question anyway. At
the cost of an additional indirect
On 12/15/20 3:00 AM, Nikhil Benesch wrote:
If this patch looks good, I'll submit it upstream tomorrow.
Assuming no news is good news, I sent
https://go-review.googlesource.com/c/gofrontend/+/278672.
On 12/16/20 5:02 AM, Richard Sandiford wrote:
Nathan Sidwell writes:
Apparently 'var+=...' is not a dash thing. Fixed thusly.
* config.m4: Avoid non-dash idiom
* configure: Rebuilt.
pushed (2 patches, because I didn't look carefully enough the first time)
Thanks. I think
On 12/16/20 8:05 AM, Jonathan Wakely wrote:
Before CWG DR 1955 the controlling expression for an #elif must be
syntactically correct, meaning this won't compile with C++11 compilers
such as gcc 4.8:
The solution is to define __has_include(X) as 0 for compilers that don't
support it.
The second
On 12/16/20 6:47 AM, Jakub Jelinek wrote:
> On Wed, Dec 16, 2020 at 05:36:04AM -0800, H.J. Lu wrote:
>> gas is correct.
>>
>>> comdat or for retain), then we should do something like the following
>>> untested change, but if it is gas bug, it should be fixed there.
>>>
>>> 2020-12-15 Jakub Jeli
Thanks Jakub for pointing at libcpp.
The -enable-checking configure code in libcody didn't play well with
us. This just uses libcpp's configurey for that piece.
libcody/
* configure.ac: Use libcpp's enable-checking code.
* configure: Rebuilt.
pushing to trunk
--
Nathan
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v3/ChangeLog:
PR libstdc++/96083
* include/ext/throw_allocator.h: Use __has_builtin to check for
__builtin_sprintf support, and use std::sprtinf if necessary.
Tested powerpc64le-linux. Committ
Hi Jakub,
when running sollve_vv tests (which we're trying to make work out-of-the-box),
some of the 5.0 tests which were testing map clause ordering also required
supporting mapping the same variable multiple times.
This was not allowed before, so besides the small changes needed in the
front-en
Hi Jakub,
we have some other sollve_vv tests for OpenMP 5.0, which tests the occurrence of
'delete' map kind on the target directive.
Looking through the specification, as early as 4.5, it seems that there's no
wording that release/delete map kinds are explicitly prohibited on structured
block
d
On 12/16/20 3:38 AM, Kewen.Lin via Gcc-patches wrote:
on 2020/12/16 下午2:35, Stott Graham via Gcc-patches wrote:
If any ---checking options used
The related PR PR98311[1] has been filed by David.
BR,
Kewen
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98311
now fixed
--
Nathan Sidwell
Hi!
On Wed, Dec 16, 2020 at 02:35:32PM +0100, Jakub Jelinek wrote:
> As mentioned in the PR, shrink-wrapping disqualifies for prologue
> placement basic blocks that have EDGE_CROSSING incoming edge.
> I don't see why that is necessary, those edges seem to be redirected
> just fine, both on x86_64
On Wed, Dec 16, 2020 at 09:28:56AM -0600, Segher Boessenkool wrote:
> This used to not work, as mentioned in the original patch submission:
> https://patchwork.ozlabs.org/project/gcc/patch/52f14532eb742ac8d878a185a46a88da7b0326eb.1442588483.git.seg...@kernel.crashing.org/
> I wonder what changed?
On Wed, Dec 16, 2020 at 09:22:49AM -0500, Nathan Sidwell wrote:
> Thanks Jakub for pointing at libcpp.
>
> The -enable-checking configure code in libcody didn't play well with
> us. This just uses libcpp's configurey for that piece.
>
> libcody/
> * configure.ac:
On 12/15/20 10:29 PM, Jeff Law wrote:
On 11/11/20 6:09 PM, Martin Sebor via Gcc-patches wrote:
The attached patch builds on top of the series I posted last
week(*) to improve the detection of out of bounds pointers
and C++ references, as well as a subset of invalid pointer
relational and subtr
On 12/16/20 8:00 AM, Jonathan Wakely via Gcc-patches wrote:
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v3/ChangeLog:
PR libstdc++/96083
* include/ext/throw_allocator.h: Use __has_builtin to check for
__builtin_sprintf support, and use
On 16/12/20 09:23 -0700, Martin Sebor via Libstdc++ wrote:
On 12/16/20 8:00 AM, Jonathan Wakely via Gcc-patches wrote:
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v3/ChangeLog:
PR libstdc++/96083
* include/ext/throw_allocator.h: Use __has_bui
On 16/12/20 17:02 +, Jonathan Wakely wrote:
On 16/12/20 09:23 -0700, Martin Sebor via Libstdc++ wrote:
On 12/16/20 8:00 AM, Jonathan Wakely via Gcc-patches wrote:
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v3/ChangeLog:
PR libstdc++/96083
On 12/16/20 10:40 AM, Jakub Jelinek wrote:
On Wed, Dec 16, 2020 at 09:22:49AM -0500, Nathan Sidwell wrote:
Thanks Jakub for pointing at libcpp.
The -enable-checking configure code in libcody didn't play well with
us. This just uses libcpp's configurey for that piece.
libcody/
On Wed, 2 Dec 2020 at 04:31, Daniel Engel wrote:
>
> Hi Christophe,
>
> On Thu, Nov 26, 2020, at 1:14 AM, Christophe Lyon wrote:
> > Hi,
> >
> > On Fri, 13 Nov 2020 at 00:03, Daniel Engel wrote:
> > >
> > > Hi,
> > >
> > > This patch adds an efficient assembly-language implementation of IEEE-
> >
On Wed, Dec 16, 2020 at 04:36:37PM +0100, Jakub Jelinek wrote:
> On Wed, Dec 16, 2020 at 09:28:56AM -0600, Segher Boessenkool wrote:
> > This used to not work, as mentioned in the original patch submission:
> > https://patchwork.ozlabs.org/project/gcc/patch/52f14532eb742ac8d878a185a46a88da7b0326eb.
On Wed, Dec 16, 2020 at 11:17:29AM -0600, Segher Boessenkool wrote:
> It was just a bootstrap+regression check, so no new testcase was needed.
> I don't remember what target, but powerpc (32+64, BE) probably.
I'll bootstrap/regtest the patch on powerpc64 and powerpc64le then.
>
> > The testcase i
On Wed, Dec 16, 2020 at 06:31:35PM +0100, Jakub Jelinek wrote:
> On Wed, Dec 16, 2020 at 11:17:29AM -0600, Segher Boessenkool wrote:
> > It was just a bootstrap+regression check, so no new testcase was needed.
> > I don't remember what target, but powerpc (32+64, BE) probably.
>
> I'll bootstrap/r
offsetof is underspecified. GCC happened to accept an unneeded
explicit scoping, clang does not.
gcc/cp/
* module.cc (dumper::push): Clangify offsetof use.
pushed
--
Nathan Sidwell
diff --git i/gcc/cp/module.cc w/gcc/cp/module.cc
index d2806966e43..e9ea18636b4 100644
--- i/gcc
On 12/16/20 10:04 AM, Jonathan Wakely wrote:
On 16/12/20 17:02 +, Jonathan Wakely wrote:
On 16/12/20 09:23 -0700, Martin Sebor via Libstdc++ wrote:
On 12/16/20 8:00 AM, Jonathan Wakely via Gcc-patches wrote:
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v
On понеделник, 14 декември 2020 г. 20:53:36 EET Dimitar Dimitrov wrote:
> On петък, 11 декември 2020 г. 19:00:35 EET abebeos wrote:
> > After "digesting" a bit more your review, I need to thank you for opening
> > my eyes re "cherrypick" suggestion and... the missing g++ summaries. I
> > need
> > t
Hi!
Jonathan mentioned on IRC that ISO/IEC 14882:2020 has been published
yesterday (and indeed it appears on www.iso.org for sale).
I think we should reflect that in our documentation and in cxx-status.html,
patches attached.
I understand we want to keep C++20 support experimental even in GCC 11,
Since "6fbec038f7a Use SHF_GNU_RETAIN to preserve symbol definitions",
GCC could create sections with the 'R' flag, which GAS would map to
SHF_GNU_RETAIN.
SHF_GNU_RETAIN support was not available in gold until Binutils commit
ff4bc37d77, on 2020-12-14. Until the support was added, invalid binaries
> On Dec 16, 2020, at 6:35 AM, Maciej W. Rozycki wrote:
>
> ...
> NB the PDP-11 pieces affected here and tripping this assertion are I
> believe dead code, as these insns are only ever produced by splitters and
> do not appear to be referred by their names.
Right; I gave them names for do
Hi Nikhil,
> On 12/15/20 3:00 AM, Nikhil Benesch wrote:
>> If this patch looks good, I'll submit it upstream tomorrow.
>
> Assuming no news is good news, I sent
> https://go-review.googlesource.com/c/gofrontend/+/278672.
sorry for the delay, but unfortunately news is not so good: I get
runtime_s
On 16/12/20 19:36 +0100, Jakub Jelinek wrote:
Hi!
Jonathan mentioned on IRC that ISO/IEC 14882:2020 has been published
yesterday (and indeed it appears on www.iso.org for sale).
I think we should reflect that in our documentation and in cxx-status.html,
patches attached.
I understand we want to
Recent changes to use __int128 as an integer-like type in and
to optimize std::uniform_int_distribution mean that the library relies
on __int128 more heavily than in the past.
The library expects that if __int128 is supported then either
__GLIBCXX_TYPE_INT_N_0 is defined (and we treat is like the
On 16/12/20 17:02 +, Jonathan Wakely wrote:
On 16/12/20 09:23 -0700, Martin Sebor via Libstdc++ wrote:
On 12/16/20 8:00 AM, Jonathan Wakely via Gcc-patches wrote:
Clang doesn't support __builtin_sprintf, so use std::sprintf instead.
libstdc++-v3/ChangeLog:
PR libstdc++/96083
I'd missed an install-strip rule in c++tools. Here it is, cribbed
from gcc/ subdir.
c++tools/
* Makefile.in (INSTALL): Replace with ...
(INSTALL_PROGRAM): ... this.
(INSTALL_STRIP_PROGRAM): New.
(install-strip): New target.
(install): Use INSTALL_
On Wed, Dec 16, 2020 at 11:49 AM Nathan Sidwell wrote:
>
>
> I'd missed an install-strip rule in c++tools. Here it is, cribbed
> from gcc/ subdir.
>
> c++tools/
Missing PR other/98328 here.
> * Makefile.in (INSTALL): Replace with ...
> (INSTALL_PROGRAM): ... this.
>
I think this is nonsense code, we seem to be naming an instantiation
of a template template parm. But this fixes the ICE. Perhaps we
should diagnose the issue earlier?
gcc/cp/
* parser.c (cp_parser_elaborated_type_specifier): Test
BOUND_TEMPLATE_TEMPLATE_PARM before che
Rather than early-include sys/socket.h, let's allow the includer to
tell cody no networking.
libcody/
* cody.hh: Allow user to set CODY_NETWORKING.
gcc/cp/
* mapper-resolver.cc: Remove early include of
sys/socket.h. Specify no CODY_NETWORKING instead.
On 12/16/20 2:20 PM, Rainer Orth wrote:
Hi Nikhil,
On 12/15/20 3:00 AM, Nikhil Benesch wrote:
If this patch looks good, I'll submit it upstream tomorrow.
Assuming no news is good news, I sent
https://go-review.googlesource.com/c/gofrontend/+/278672.
sorry for the delay, but unfortunately n
Hi All,
This adds implementation for the optabs for complex additions. With this the
following C code:
void f90 (float complex a[restrict N], float complex b[restrict N],
float complex c[restrict N])
{
for (int i=0; i < N; i++)
c[i] = a[i] + (b[i] * I);
}
generates
1 - 100 of 141 matches
Mail list logo
