Re: [PATCH] PR fortran/95880 - [9/10/11 Regression] ICE in gfc_add_type, at fortran/symbol.c:2030

Hi Harald, A rather straightforward issue (mis-)referencing the proper symbol name in an error message. OK for master / backport? OK, but don't forget the test case :-) Because the if takes up quite a few lines, it is also an option to write something like gfc_error ("Symbol %qs at %L canno

[pushed] coroutines, testsuite: Update log messages. [NFC, PR95519]

Hi. this corrects some pastos in the user error messages for the testcases, in PR95519. NFC. tested on x86_64-darwin, applied to master. thanks Iain This does not affect the test functionality, but only user- facing debug messages when the tests are run outside the test-suite. gcc/testsuite/C

[PATCH] libstdc++: Add a __nonnnull__ attribute to std::string's _CharT* constructor

2020-06-28 Ville Voutilainen Add a __nonnnull__ attribute to std::string's _CharT* constructor * include/bits/basic_string.h (string(_CharT*, const _Alloc&)): Add a __nonnull__ attribute. * testsuite/21_strings/basic_string/cons/char/nonnull.cc: New. * testsuite/21_strings/b

Re: [PATCH] libstdc++: Add a __nonnnull__ attribute to std::string's _CharT* constructor

On Sun, 28 Jun 2020 at 13:56, Ville Voutilainen wrote: > > 2020-06-28 Ville Voutilainen > > Add a __nonnnull__ attribute to std::string's _CharT* constructor > * include/bits/basic_string.h (string(_CharT*, const _Alloc&)): > Add a __nonnull__ attribute. > * testsuite/21_strings

Aw: Re: [PATCH] PR fortran/95880 - [9/10/11 Regression] ICE in gfc_add_type, at fortran/symbol.c:2030

Hi Thomas, > OK, but don't forget the test case :-) the testcase was in the patch. I just do not add it to the commit message, because that's now done automagically by the new scripts. > Because the if takes up quite a few lines, it is also an option to > write something like > > gfc_error ("Sy

[PATCH PR95855]A missing ifcvt optimization to generate fcsel

Hi, This is a simple fix for pr95855. With this fix, pass_split_paths can recognize the if-conversion opportunity of the testcase and doesn't duplicate the corresponding block. Added one testcase for this. Bootstrap and tested on both aarch64 and x86 Linux platform, no new regres

[pushed] coroutines: Handle namespaces while scanning local vars [PR95711].

Hi, We need to skip past namespace decls when scanning the bind expression var lists checking for local vars. The PR notes that this omission results in an ice-on-valid. tested on x86_64-linux, darwin, powerpc64-linux, applied to master as obvious (and will backport to 10.2). thanks Iain gcc/

[PATCH] x86: Enable FMA in rsqrt2 expander

Enable FMA in rsqrt2 expander and fold rsqrtv16sf2 expander into rsqrt2 expander which expands to UNSPEC_RSQRT28 for TARGET_AVX512ER. Although it doesn't show performance change in our workloads, FMA can improve other workloads. gcc/ PR target/88713 * config/i386/i386-expand.c (ix

[PATCH, committed] PR fortran/95340 - [10/11 Regression] ICE in gfc_match_select_rank, at fortran/match.c:6690

Committed as obvious. NULL pointer dereference originally found by Steve. My patch is shorter. :-) Regtested on x86_64-pc-linux-gnu. Thanks, Harald PR fortran/95340 - ICE in gfc_match_select_rank, at fortran/match.c:6690 Do not dereference NULL pointer when querying array shape of possibly i

[committed] d: Merge upstream dmd 8508c4e68.

Hi, This patch merges the D front-end implementation with upstream dmd 8508c4e68. Fixes a performance bug where 'static foreach' would take an exponentially long time to expand during CTFE. In the following example: static foreach (i; 0..3) {} Compilation time had been reduced from aro

[PATCH] PR fortran/71706 - [8/9/10/11 Regression] [Coarray] ICE on using sync images with integer(kind<>4), with -fcoarray=lib -fcheck=bounds

Here's a fix to bounds-checking code that manifests itself essentially with checking enabled. Once found and understood, the fix is trivial: just properly convert the argument kind of SYNC IMAGES for checking. Regtested on x86_64-pc-linux-gnu. OK for master / backports where appropriate? Thanks

Re: [PING][PATCH] underline null argument in -Wnonnull (PR c++/86568)

On 6/26/20 3:58 PM, Jeff Law wrote: On Mon, 2020-06-22 at 12:02 -0600, Martin Sebor wrote: Ping: https://gcc.gnu.org/pipermail/gcc-patches/2020-June/547415.html Jason already approved the C++ changes (with a couple of minor tweaks). I'm still looking for an approval of the corresponding middle

Re: [PATCH] libstdc++: Add a __nonnnull__ attribute to std::string's _CharT* constructor

On 28/06/20 14:01 +0300, Ville Voutilainen via Libstdc++ wrote: On Sun, 28 Jun 2020 at 13:56, Ville Voutilainen wrote: 2020-06-28 Ville Voutilainen Add a __nonnnull__ attribute to std::string's _CharT* constructor * include/bits/basic_string.h (string(_CharT*, const _Alloc&)):

Re: [PATCH] libstdc++: std::variant doesn't like types with a defaulted virtual destructor [PR95915]

On 27/06/20 18:20 +0300, Ville Voutilainen via Libstdc++ wrote: On Sat, 27 Jun 2020 at 17:53, Ville Voutilainen wrote: On Fri, 26 Jun 2020 at 21:20, Jonathan Wakely wrote: > For these three tests I think this would be slightly better: > > // { dg-additional-options "-Wno-deprecated" { target

Re: [PATCH] libstdc++: Add a __nonnnull__ attribute to std::string's _CharT* constructor

On Mon, 29 Jun 2020 at 00:16, Jonathan Wakely wrote: > >Hmm, let's use dg-additional-options here too, and axe the pointless > >-std=gnu++11. > > I agree the -std=gnu++11 isn't needed, but thre doesn't seem to be any > advantage to dg-additional-options here. The reason I suggested it for > th oth

Re: [PATCH] handle MEM_REF with void* arguments (PR c++/95768)

On 6/23/20 1:12 AM, Richard Biener wrote: On Tue, Jun 23, 2020 at 12:22 AM Martin Sebor via Gcc-patches wrote: On 6/22/20 12:55 PM, Jason Merrill wrote: On 6/22/20 1:25 PM, Martin Sebor wrote: The attached fix parallels the one for the equivalent C bug 95580 where the pretty printers don't c