On 03/24/2011 08:49 PM, Joseph S. Myers wrote:
Related to that point, I notice a piece of code starting we might need to
use some other shell than /bin/sh for running subshells that tries to
determine a shell on Windows hosts. It's autoconf's job to find a
suitable shell, so if this code is
Ping for following patches
[patch libcpp]: Improve handling of DOS-filenames and -paths
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00742.html
[patch gcc c++ c-family java fortan lto]: Fix DOS-filesystem issues
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00824.html
Kai
2011/3/25 Jason Merrill ja...@redhat.com:
On 03/24/2011 08:44 PM, Kai Tietz wrote:
+static bool
+attribute_value_equal (const_tree attr1, const_tree attr2)
+{
+ if (attr1 == attr2)
+ return true;
I still think this test should be comparing TREE_VALUEs.
Jason
Ok, it makes sense to
2011/3/25 Jason Merrill ja...@redhat.com:
OK.
Jason
Committed at revision 171445.
Thanks,
Kai
On 03/21/2011 01:59 PM, Rainer Orth wrote:
diff -r 8efcb4b03d45 boehm-gc/Makefile.am
--- a/boehm-gc/Makefile.am Sun Mar 20 14:28:11 2011 +0100
+++ b/boehm-gc/Makefile.am Sun Mar 20 21:37:14 2011 +0100
@@ -4,10 +4,10 @@
## files that should be in the distribution are not mentioned in
Hi!
While working on a patch I'll post momentarily, I've noticed two
issues:
1) DW_OP_GNU_entry_value wasn't being handled by
hash_loc_operands/compare_loc_operands, which means different entry
values could be considered the same and a wrong location list
could be used; the entry value
On Fri, Mar 25, 2011 at 07:52, Nathan Froyd froy...@codesourcery.com wrote:
* fold-const.c (expr_location_or): New function.
(fold_truth_not_expr): Call it.
OK.
Diego.
On 24/03/11 13:22, Richard Earnshaw wrote:
OK
Committed, thanks.
Andrew
Joseph S. Myers jos...@codesourcery.com writes:
(powerpcle-*-solaris*): Don't disable make expect gnuserv on
target.
As I happen to see this: Solaris/PowerPC is ancient history by now
(2.5.1, 2.6, I believe). This can probably go.
Rainer
--
As discussed in the PR, I'd like to skip the 4 very expensive
gcc.dg/graphite/block-?.c tests for now.
Tested with the appropriate runtest invocation with and without
GCC_TEST_RUN_EXPENSIVE set.
Ok for mainline?
Rainer
2011-03-25 Rainer Orth r...@cebitec.uni-bielefeld.de
PR
On Fri, 25 Mar 2011, Rainer Orth wrote:
As discussed in the PR, I'd like to skip the 4 very expensive
gcc.dg/graphite/block-?.c tests for now.
Tested with the appropriate runtest invocation with and without
GCC_TEST_RUN_EXPENSIVE set.
Ok for mainline?
Ok.
Thanks,
Richard.
On Fri, 25 Mar 2011, Rainer Orth wrote:
Joseph S. Myers jos...@codesourcery.com writes:
(powerpcle-*-solaris*): Don't disable make expect gnuserv on
target.
As I happen to see this: Solaris/PowerPC is ancient history by now
(2.5.1, 2.6, I believe). This can probably go.
It
This merges the machinery for verify_stmts and verify_types_in_gimple_*.
A step in the direction to merge the various verifications we do
on gimple. A further step would move the calls to verify_expr to
appropriate places when verifying stmts instead of walking all ops.
The goal is still that
Bootstrapped etc. powerpc64-linux. OK mainline?
PR target/47487
* config/rs6000/rs6000.c (rs6000_output_function_epilogue): Support
GNU Go in traceback table.
Index: gcc/config/rs6000/rs6000.c
===
---
On Fri, Mar 25, 2011 at 09:30:22AM -0600, Jeff Law wrote:
I'm not going to chime in on this specific problem; however, it is worth
noting that many of the issues raised by the C++0x memory model also
affect the linux kernel.
But what they are seeing are certainly store data races, not load
On Fri, Mar 25, 2011 at 4:27 PM, Rainer Orth
r...@cebitec.uni-bielefeld.de wrote:
The new gcc.dg/torture/pr47917.c test currently fails the execution test
on IRIX 6.5. To get a C99-conformant snprintf, one needs to include
stdio.h with _XOPEN_SOURCE defined as 500. The following patch does
The attached patch adds the muldi3 function for bfin.
2011-03-25 Stuart Henderson stuart.hender...@analog.com
From Bernd Schmidt
* config/bfin/t-bfin (LIB1ASMFUNCS): Add muldi3 and umulsi3_highpart.
* config/bfin/t-bfin-elf (LIB1ASMFUNCS): Add muldi3.
*
This patch adds streamer hooks for writing and reading trees. It is
still very much WIP. In particular, the hooks for reading and writing
may change.
These callbacks are called from lto_write_tree and lto_read_tree after
the common fields in the tree have been processed. They are
It's been a while since I read Cary's proposal, and I am no longer
likely to do much work of my own in this area. So I'll just respond at
the high level.
I like very much the essential notion of the stack being of typed
entities with no specification of how the consumer actually implements
it.
Ping.
Thanks,
Martin
--
On Sat, Mar 19, 2011 at 01:48:36AM +0100, Martin Jambor wrote:
Hi,
On Fri, Mar 11, 2011 at 03:30:53PM +0100, Jan Hubicka wrote:
Index: src/gcc/cgraph.c
Seems OK, however..
To what extent is this an approval? :-)
Approval only after you convince me on the questions bellow.
Index: src/gcc/gimplify.c
===
--- src.orig/gcc/gimplify.c 2011-03-19 01:16:24.0
On Fri, Mar 25, 2011 at 05:17:07PM +0100, Jakub Jelinek wrote:
On Thu, Mar 24, 2011 at 06:08:01PM +0100, Jason Merrill wrote:
This ought to cut down on the number of cases we have to handle in
all the different places in dwarf2out that deal with this stuff.
Does that make sense to you?
The attached patch fixes a compiler warning on MinGW64
../../../../../build/gcc/src/libgfortran/io/unix.c:51:0: warning:
lseek redefined [enabled by default]
which is due to MinGW64 now supporting LFS. (Seemingly, using a similar
define as libgfortran.) The following patch should fix the
I meant to include the final patch. Here it is.
Rainer
2011-03-25 Rainer Orth r...@cebitec.uni-bielefeld.de
PR boehm-gc/11412
* configure.ac (THREADCFLAGS): New variable.
Use it instead of INCLUDES, AM_CPPFLAGS.
*-*-kfreebsd*-gnu (THREADDLLIBS):
On 03/25/2011 05:41 AM, Georg-Johann Lay wrote:
On 03/22/2011 06:48 PM, Richard Henderson wrote:
Ok. Watch out for other target problems this week.
libgcc fails to build for avr (SVN 171446)
../../../../../gcc.gnu.org/trunk/libgcc/../gcc/libgcc2.c: In function
'__negdi2':
On 03/23/2011 06:19 PM, Richard Henderson wrote:
In general, with shrink-wrapping, we can have essentially arbitrary
differences in unwind info between blocks that are sequential.
So, while that isn't the case just yet with the current shrink-wrapping
patch, it seems I will either have to make
This patch on top of
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01224.html
and
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01723.html
implements parts of Cary's typed DWARF stack proposal:
http://www.dwarfstd.org/doc/040408.1.html
I haven't looked at the patch yet, but this sounds great;
Richard Henderson r...@redhat.com writes:
This is due to a miscommunication between the middle-end and the backend
about how many arguments the setmemhi pattern takes.
(define_expand setmemhi
[(parallel [(set (match_operand:BLK 0 memory_operand )
(match_operand 2
Dear Paul,
Paul Richard Thomas wrote:
Why did Jakub's patch for PR45304 not fix this? He visited the fix in
your patch upon build_library_function_decl_1 and
gfc_get_intrinsic_lib_fndecl.
Well, my patch is in a way a follow up to Jakub's. His patch only fixed
the library call definitions and
Richard Earnshaw rearn...@arm.com writes:
gcc/
* config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Restrict FPA_REGS
case to VFPv1.
I think adding a comment that VFPv1 would require a restriction but that
that isn't supported would be useful. Then if the worst happens it will
be one
Richard Sandiford richard.sandif...@linaro.org writes:
Richard Henderson r...@redhat.com writes:
This is due to a miscommunication between the middle-end and the backend
about how many arguments the setmemhi pattern takes.
(define_expand setmemhi
[(parallel [(set (match_operand:BLK 0
Richard Sandiford richard.sandif...@linaro.org writes:
Er, = 4 even...
Or not, *sigh*. Time I went home...
Richard
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/24/11 07:09, Bernd Schmidt wrote:
This prints a bit more information in debugging dumps. Bootstrapped and
tested on i686-linux.
OK.
jeff
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora -
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/24/11 07:14, Bernd Schmidt wrote:
We have a mechanism to prevent the scheduler from touching certain
blocks; this is used by modulo scheduling. sched-ebb does not honor this
flag currently; this patch fixes it. Bootstrapped and tested on
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/24/11 07:19, Bernd Schmidt wrote:
We can currently select an insn to be scheduled, only to find out that
it's not actually valid at the current time, either due to state
conflicts or being an asm with something else already scheduled in the
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/25/11 02:48, Kai Tietz wrote:
Ping for following patches
[patch libcpp]: Improve handling of DOS-filenames and -paths
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00742.html
[patch gcc c++ c-family java fortan lto]: Fix DOS-filesystem
On 03/25/2011 10:51 AM, Richard Sandiford wrote:
Thanks. I think it needs to be s/!= 4/= 6/ though, so that
match_scratches still work when 6 operands really are passed in.
For the record, I audited all setmem and movmem patterns.
There are is only one that uses 6 operands: i386.
There are
2011/3/25 Jeff Law l...@redhat.com:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 03/25/11 02:48, Kai Tietz wrote:
Ping for following patches
[patch libcpp]: Improve handling of DOS-filenames and -paths
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00742.html
[patch gcc c++ c-family
This patch to the Go frontend avoids a double error when there is a
negative shift count. We used to give a negative shift count error
followed by an overflow error. This just gives the former.
Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu.
Committed to mainline.
Ian
diff -r
On Fri, Mar 25, 2011 at 11:09 AM, Alan Modra amo...@gmail.com wrote:
Much of the following patch is based on Michael Meissner's support for
vector reloads. The predicates.md change teaches the predicate used
by the Y constraint to check cmodel medium addresses in case such
addresses should
Hi!
I've been removing mempcpy uses from libquadmath and replacing them with
memcpy + increment, but apparently missed removing one letter in one case.
Without a mempcpy prototype on targets that don't have mempcpy apparently
gcc just warns, but optimizes it into memcpy anyway, so no big harm was
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
var-tracking (and other modules) have a habit of assigning values to
objects, then never using the values. We still need the RHS side effect
in these cases, so the assignment was deleted leaving just the RHS as a
statement.
Fixed thusly for
Add pph_stream_write_tree and pph_stream_read_tree callbacks.
Make them handle DECL_SAVED_TREE. This produces several new
failures since it exposes more unhandled trees.
Committed on pph branch. Fixes for the introduced failures coming up
in the next patch.
cp/ChangeLog.pph
2011-03-25 Diego
Handle STATEMENT_LIST and add hooks for writing non-pointer values
into bitpacks.
This patch fixes most of the failures in pph.exp (only 14 remain). We
now handle STATEMENT_LIST properly.
The patch also introduces hooks for writing/reading non pointer values
from trees.
Tested on x86_64.
David Edelsohn wrote:
On Fri, Mar 25, 2011 at 11:09 AM, Alan Modra amo...@gmail.com wrote:
Much of the following patch is based on Michael Meissner's support for
vector reloads. =A0The predicates.md change teaches the predicate used
by the Y constraint to check cmodel medium addresses in
On 25 March 2011 22:38, Jonathan Wakely wrote:
The tests are invalid. There were preconditions added to match_results
member functions recently, which we now check in debug mode, and those
tests violate them.
I'll deal with it, thanks for testing it.
2011-03-25 Jonathan Wakely
2011-03-25 Jonathan Wakely jwakely@gmail.com
* include/std/future (future::share): Add.
(packaged_task::result_type): Remove as per LWG 2030.
(packaged_task::packaged_task): Remove redundant constructors, as per
LWG 1514.
*
On Fri, Mar 25, 2011 at 09:36:04PM +0100, Ulrich Weigand wrote:
Looks good to me, except ...
+ mem = change_address (mem, VOIDmode, scratch_or_premodify);
Maybe replace_equiv_address instead, to avoid losing the memory
attribute information (alignment, alias set, ...)?
Yes, in fact
On 03/25/2011 01:19 PM, Jeff Law wrote:
* var-tracking.c (canonicalize_values_mark): Delete unused
lhs assignment.
(canonicalize_values_star, set_variable_part): Likewise.
(clobber_variable_part, delete_variable_part): Likewise.
Ok.
r~
Hi,
[sorry for breaking the threading I've deleted the mails I'm answering
already]
In any case, citing from
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01250.html
Here's the patch I came up with. It is on top of the previous one, so
if we want to backport to 4.6 later, both are
On 03/25/2011 10:34 AM, Bernd Schmidt wrote:
I don't know much about the unwinding code. I'm currently thinking about
writing out a cfi_remember_state at the start of the function, restoring
that clean state when necessary at the start of a new block and emitting
the necessary directives to
51 matches
Mail list logo