The second alternative for operand 1 needs a new constraint which does
not overlap with Pg, except that it can handle 32 to generate an
optimal mov.
This patch introduces a new Ph constraint which has the [32..255]
range.
This fixes a lot of regressions when running the testsuite for an MVE
targe
Hi Jeff:
> I'd suggest deferring to gcc-17.
Yeah, I am OK with that, actually, I expected to receive this response :P
And I'm going to merge that to our downstream to make sure we have a
more stable version when GCC 17 is open.
Hi Andrew:
> That makes sense so my question about -Os is withdr
Reformatted and minor changes
On Sun, 30 Nov 2025 15:57:51 +0100
Christopher Albert wrote:
> When an array constructor has an explicit type-spec, all elements must
> be converted to that type and character elements must be
> padded/truncated to the specified length. This was working for simple
在 2025/11/29 上午4:40, Jakub Jelinek 写道:
Hi!
C++20, in particular https://wg21.link/P1120R0 paper voted into it,
deprecates various operations between enumerators from different enumeration
types etc., and as we've switched to -std=gnu++20 by default, this now
results in warnings or errors durin
On Fri, Oct 24, 2025 at 12:21 PM Andrew Pinski
wrote:
>
> diff --git a/gcc/config/i386/i386-expand.cc b/gcc/config/i386/i386-expand.cc
> index a1f1b26f78a..3196d4b600b 100644
> --- a/gcc/config/i386/i386-expand.cc
> +++ b/gcc/config/i386/i386-expand.cc
> @@ -26377,18 +26377,17 @@ ix86_gen_ccmp_n
On Sun, Nov 30, 2025 at 3:09 PM Jeff Law wrote:
>
>
>
> On 11/30/25 11:15 AM, John David Anglin wrote:
> > Okay?
> >
> > Dave
> > ---
> >
> > libcc1: Check for libgen.h header
> >
> > The basename() function is declared in libgen.h.
> >
> > 2025-11-30 John David Anglin
> >
> > libcc1/ChangeLog:
From: Owen Avery
This fixes a false positive with undefined features, improves the
implementation of Feature::as_name, and configures the testsuite to
check further stages of libcore compilation.
gcc/rust/ChangeLog:
* ast/rust-macro.h (MetaNameValueStr::get_name): New function.
On 01/12/25 04:11, Jan Hubicka wrote:
External email: Use caution opening links or attachments
+ /* For a given index, returns the symbol name. */
+ const char *get_symbol_name (int index) const;
+
+ /* For a given index, returns the filename. */
+ const char *get_filename (int index) co
On Mon, Dec 1, 2025 at 10:31 AM H.J. Lu wrote:
>
> For a basic block with only deleted instructions:
>
> (code_label 348 23 349 45 3 (nil) [0 uses])
> (note 349 348 436 45 [bb 45] NOTE_INSN_BASIC_BLOCK)
> (note 436 349 362 45 NOTE_INSN_DELETED)
>
> emit the TLS call after deleted instructions.
>
>
For a basic block with only deleted instructions:
(code_label 348 23 349 45 3 (nil) [0 uses])
(note 349 348 436 45 [bb 45] NOTE_INSN_BASIC_BLOCK)
(note 436 349 362 45 NOTE_INSN_DELETED)
emit the TLS call after deleted instructions.
gcc/
PR target/122906
* config/i386/i386-features.cc (ix86_emit
Hi Jeff,
On Sun, 2025-11-30 at 15:33 -0700, Jeff Law wrote:
>
> We don't bother with the transformation when the XOR is flipping a bit
> known to be zero (ie, a high bit of the result of the right shift or a
> low bit on the result of the left shift). For those cases we already
> figure out t
Hi!
The following testcase fails due to ODR warnings starting with the
r16-3233 change (P2115R0 PR120503 implementation).
The problem is that for C++20 we mangle differently the anonymous
union at the class scope from C++17, in C++17 the
unnamed enumeration that has an enumerator as a name for li
Regtested on x86_64-pc-linux-gnu, OK for trunk?
-- >8 --
When checking a deleted explicit specialization in a SFINAE context,
we were failing to propagate the error from mark_used. The call to
resolve_nondeduced_context in convert_to_void correctly returned
error_mark_node, but mark_single_funct
On 11/30/25 11:15 AM, John David Anglin wrote:
Okay?
Dave
---
libcc1: Check for libgen.h header
The basename() function is declared in libgen.h.
2025-11-30 John David Anglin
libcc1/ChangeLog:
* configure.ac: Check for libgen.h header.
* configure: Regenerate.
OK
je
> (1) Changes to auto-profile pass:
> The patch adds a new field timestamp to function_instance,
> and populates it in read_function_instance.
>
> It maintains a new timestamp_info_map from timestamp -> ,
> which maps timestamps sorted in ascending order to (1..N), so lowest ordered
> timestamp is
> From: Dhruv Chawla
>
> Signed-off-by: Dhruv Chawla
>
> gcc/c/ChangeLog:
>
> * Make-lang.in: Bump GCOV version to 3.
>
> gcc/cp/ChangeLog:
>
> * Make-lang.in: Bump GCOV version to 3.
>
> gcc/lto/ChangeLog:
>
> * Make-lang.in: Bump GCOV version to 3.
> - $(CREATE
> From: Dhruv Chawla
>
> Signed-off-by: Dhruv Chawla
>
> gcc/ChangeLog:
>
> * auto-profile.cc (string_table::~string_table): Update to free
> original_names_map_.
> (string_table::original_names_map_): New member.
> (string_table::clashing_names_map_): Likewise.
>
> + /* For a given index, returns the symbol name. */
> + const char *get_symbol_name (int index) const;
> +
> + /* For a given index, returns the filename. */
> + const char *get_filename (int index) const;
> +
> + /* For a given symbol name index, returns the filename index. */
> + int g
In this PR we're getting code like this out of the gimple optimizers:
_1 = a_4(D) << 63;
_2 = a_4(D) >> 1;
_3 = _2 ^ 1;
_5 = _1 | _3;
Note the XOR in that sequence. It spoils our ability to recognize the
rotation. As a result we get code like this for rv64gcb:
srli
Hi Chris!
Am 30.11.25 um 15:57 schrieb Christopher Albert:
When an array constructor has an explicit type-spec, all elements must
be converted to that type and character elements must be
padded/truncated to the specified length. This was working for simple
cases but failing when:
1. Elements w
Ping, any interest in this patch? Thanks!
https://gcc.gnu.org/pipermail/gcc-patches/2025-November/700815.html
On Sun, Nov 16, 2025 at 12:33 AM Lewis Hyatt wrote:
>
> Hello-
>
> This patch makes it easier to disable a specific language, which is a need
> that comes up from time to time. I have tes
This commit fixes a few problems in algol68/a68-diagnostics.cc,
reported by David Malcolm in-list.
gcc/ChangeLog
* algol68/a68-diagnostics.cc (diagnostic): Copypasto "a meek"
should be "a firm". Support printing line number for programs
longer than 9 lines. Use obstack_a
Hi David.
> On Sun, 2025-11-30 at 02:06 +0100, Jose E. Marchesi wrote:
>> This commit adds the diagnostics infrastructure for the Algol 68
>> front-end.
>
> Hi Jose, congrats on getting all this merged.
Thanks! :)
> A few comments inline below:
>
>>
>> Signed-off-by: Jose E. Marchesi
>> Co-a
> On Sun, Nov 30, 2025 at 01:57:38PM +0100, Jose E. Marchesi wrote:
>> A few Fortran tests that pass both Fortran and C programs in the same
>> compilation check for warnings like:
>>
>> { dg-warning "command-line option '-fcheck=all' is valid for Fortran but
>> not for C" "" { target *-*-* }
Okay?
Dave
---
libcc1: Check for libgen.h header
The basename() function is declared in libgen.h.
2025-11-30 John David Anglin
libcc1/ChangeLog:
* configure.ac: Check for libgen.h header.
* configure: Regenerate.
diff --git a/libcc1/configure.ac b/libcc1/configure.ac
inde
Hi Mike,
This should be a standalone patch, not "1 of 2". The patch to use vector
pair load/store should be a separate patch.
On 25/11/25 1:50 pm, Michael Meissner wrote:
> This patch adds support for a new PowerPC cpu (future).
>
> The intention of this patch is to provide a testing platform
On 11/29/25 7:04 PM, Jakub Jelinek wrote:
On Sat, Nov 29, 2025 at 06:37:03PM +0530, Jason Merrill wrote:
Unfortunately it is not. ATTR_ID here is just the attribute name, so
lookup_attribute_spec attempts to find it in the gnu namespace rather than
in the "internal " namespace in which it is pr
On 11/30/25 6:02 AM, Egas Ribeiro wrote:
On 2025-11-29 15:30, Jason Merrill wrote:
Once we enter cp_parser_placeholder_type_specifier with
parser->auto_is_implicit_function_template_parm_p we know we're dealing with
an abbreviated function template, no need to wait until we actually see the
'au
On Sun, 2025-11-30 at 02:06 +0100, Jose E. Marchesi wrote:
> This commit adds the diagnostics infrastructure for the Algol 68
> front-end.
Hi Jose, congrats on getting all this merged.
A few comments inline below:
>
> Signed-off-by: Jose E. Marchesi
> Co-authored-by: Marcel van der Veer
>
>
When an array constructor has an explicit type-spec, all elements must
be converted to that type and character elements must be
padded/truncated to the specified length. This was working for simple
cases but failing when:
1. Elements were parenthesized: [integer :: ([1.0])]
2. Constructors were n
On Sun, Nov 30, 2025 at 01:57:38PM +0100, Jose E. Marchesi wrote:
> A few Fortran tests that pass both Fortran and C programs in the same
> compilation check for warnings like:
>
> { dg-warning "command-line option '-fcheck=all' is valid for Fortran but
> not for C" "" { target *-*-* } 0 }
>
>
Thanks all.
If most of us agree on that, I will commit it and continue the rest part soon.
Pan
-Original Message-
From: Robin Dapp
Sent: Monday, November 24, 2025 11:14 PM
To: Jeff Law ; Li, Pan2 ; Robin Dapp
; [email protected]
Cc: [email protected]; [email protected]; Ch
A few Fortran tests that pass both Fortran and C programs in the same
compilation check for warnings like:
{ dg-warning "command-line option '-fcheck=all' is valid for Fortran but not
for C" "" { target *-*-* } 0 }
The problem is that the warning may actually indicate the option is
valid for
On Fri, Nov 28, 2025 at 10:30 AM Mathias Krause wrote:
>
> When generating 16-bit code via -m16, the NOP mcount code generation
> emits a 5-byte NOP. However, that is neither a valid i8086 instruction
> (long NOPs are PentiumPro+), nor would it get decoded as a 5-byte
> instruction. It's a 4-byte
On Linux/x86_64,
9f1ae12f0ee0d3cebb7a5249b7f741c51c5934a0 is the first bad commit
commit 9f1ae12f0ee0d3cebb7a5249b7f741c51c5934a0
Author: Jose E. Marchesi
Date: Sat Nov 29 16:57:51 2025 +0100
a68: build system changes
caused
FAIL: gfortran.dg/c-interop/allocate-errors.f90 -O0 at line
36 matches
Mail list logo
