This implements my idea of how POINTER_PLUS_EXPR should end up being
changed. This addresses two things, 1) reduce the use of sizetype
nodes, 2) properly handle targets where sizeof (void *) != sizeof
(sizetype). Especially 2) is important as currently we lose information
about the signedness
Why are you not changing the gcc_AC_ macro and instead introducing
duplicate code? Perhaps the right solution is to add a final argument
to the AC_RUN_IFELSE macro (don't know, I should be on holiday and
hence I do not have the source code at hand :)); in any case this is
_not_ how you add tests
On Fri, Jun 17, 2011 at 12:40 PM, Tom de Vries vr...@codesourcery.com wrote:
On 06/17/2011 12:01 AM, Jeff Law wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 06/16/11 00:39, Tom de Vries wrote:
Hi,
Consider the following example.
extern unsigned int foo (int*)
Interesting. I'd never thought about the generation/use angle to prove
a pointer was non-null. ISTM we could use that same logic to infer that
more pointers are non-null in extract_range_from_binary_expr.
Interested in tackling that improvement, obviously as an independent patch?
2011/6/17 Zdenek Dvorak rakd...@kam.mff.cuni.cz:
Interesting. I'd never thought about the generation/use angle to prove
a pointer was non-null. ISTM we could use that same logic to infer that
more pointers are non-null in extract_range_from_binary_expr.
Interested in tackling that
Hi Jack,
On 17 Jun 2011, at 03:21, Jack Howarth wrote:
The gcj compiler needs to pass -no_pie for linkage on darwin11 due
to the new -pie
default of the linker. The attached patch accomplishes this by
passing -no_pie on SYSTEMSPEC
for *-*-darwin[12]*. Since Darwin10 supports -no_pie in
Hi,
Index: tree-vrp.c
===
--- tree-vrp.c (revision 173703)
+++ tree-vrp.c (working copy)
@@ -2273,7 +2273,12 @@ extract_range_from_binary_expr (value_ra
{
/* For pointer types, we are really
On Thu, Jun 16, 2011 at 18:57, Gabriel Charette gch...@google.com wrote:
At the end of pph_out_cxx_binding I could also simply do
pph_out_uchar (stream, PPH_RECORD_END);
instead of
pph_out_cxx_binding_1 (stream, NULL, ref_p);
which has the exact same effect.
That's fine. I kind of prefer
2011/6/17 Zdenek Dvorak rakd...@kam.mff.cuni.cz:
Hi,
Index: tree-vrp.c
===
--- tree-vrp.c (revision 173703)
+++ tree-vrp.c (working copy)
@@ -2273,7 +2273,12 @@ extract_range_from_binary_expr (value_ra
{
On Fri, 17 Jun 2011, Georg-Johann Lay wrote:
I don't see what's bat with the patch, it's straight forward.
C is a high-level language, defined in terms of high-level semantics
rather than machine instructions. C code should be written where possible
using machine-independent functionality,
On Fri, Jun 17, 2011 at 01:31:14AM +0200, Jakub Jelinek wrote:
Not here, those are handled by ix86_expand_args_builtin
instead of ix86_expand_multi_arg_builtin. Furthermore, only
CODE_FOR_vcvtps2ph and CODE_FOR_vcvtps2ph256 have CONST_INT argument.
And I believe ix86_expand_args_builtin
Hi Olivier,
Rainer Orth wrote:
I've finally gotten around to this. Apart from some comment and code
cleanups along the lines of the sol2-unwind.h files, I had to minimally
adapt the N32 multithreaded code sequence for IRIX 6.5.30 that I'm
running here. While I was at it, I added N64
On 06/17/2011 12:32 AM, H.J. Lu wrote:
Jason, can you add it to git mirror?
Done.
Jason
Hi,
On Thu, 16 Jun 2011, Richard Guenther wrote:
If people want to not create useless conversions in the first place,
though, I suspect there are lots of places that create useless
conversions in the compiler.
Yeah, the above looks it comes from the frontends - gimplification
should
/libexec/gcc/x86_64-apple-darwin11.0.0/4.7.0/lto-wrapper
Target: x86_64-apple-darwin11.0.0
Configured with: ../gcc-4.7-20110617/configure --prefix=/sw
--prefix=/sw/lib/gcc4.7 --mandir=/sw/share/man --infodir=/sw/lib/gcc4.7/info
--enable-languages=c,c++,fortran,objc,obj-c++,java --with-gmp=/sw
On 06/17/2011 02:20 AM, Georg-Johann Lay wrote:
Richard Henderson schrieb:
On 06/15/2011 02:47 AM, Georg-Johann Lay wrote:
+#if defined (L_loop_ffsqi2)
+;; Helper for ffshi2, ffssi2
+;; r25:r24 = r26 + zero_extend16 (ffs8(r24))
+;; r24 must be != 0
+;; clobbers: r26
+DEFUN __loop_ffsqi2
Richard Henderson schrieb:
On 06/17/2011 02:20 AM, Georg-Johann Lay wrote:
Richard Henderson schrieb:
On 06/15/2011 02:47 AM, Georg-Johann Lay wrote:
Yes, you are right. Following patchlet ok?
Johann
* config/avr/libgcc.S (__ctzsi2, __ctzhi2):
Map zero to 255.
You'd also
On Jun 16, 2011, at 7:21 PM, Jack Howarth wrote:
The gcj compiler needs to pass -no_pie for linkage on darwin11 due to the
new -pie
default of the linker. The attached patch accomplishes this by passing
-no_pie on SYSTEMSPEC
for *-*-darwin[12]*. Since Darwin10 supports -no_pie in its
I've done a lot of g++-only testsuite runs lately and I noticed that it
didn't parallelize all that well. The patch below adds a couple more
.exp files to the parallel infrastructure. dg-torture.exp is the big
one; it takes about as much time as old-deja.exp.
Other valid candidates are lto.exp
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 06/17/11 08:12, Zdenek Dvorak wrote:
if we do, we definitely should not -- the only point of such a construction
is
to bypass the pointer arithmetics restrictions,
Ok, we don't. Where does the C standard say that NULL + 0 is undefined?
I
On Fri, Jun 17, 2011 at 10:45 AM, Fang, Changpeng
changpeng.f...@amd.com wrote:
Why not just move AVX256_SPLIT_UNALIGNED_STORE
and AVX256_SPLIT_UNALIGNED_LOAD to ix86_tune_indices?
I would like to keep the -m option so that at least we can explicitly turn
off the splittings when regressions are
OK.
Jason
Joseph S. Myers schrieb:
On Fri, 17 Jun 2011, Georg-Johann Lay wrote:
I don't see what's bat with the patch, it's straight forward.
C is a high-level language, defined in terms of high-level
semantics rather than machine instructions. C code should be
written where possible using
Georg-Johann Lay schrieb:
To come back to the original topic, here is a tentative patch for
better popcount and parity:
* config/avr/t-avr (LIB1ASMFUNCS): Rename _loop_ffsqi2 to
_ffsqi2_nz.
* confif/avr/libgcc.S: Ditto. Rename __loop_ffsqi2 to __ffsqi2_nz.
Hi all,
I have just committed to trunk a simple patch for an OOP regression
with MOVE_ALLOC (see comment #9-#11 in the PR):
http://gcc.gnu.org/viewcvs?view=revisionrevision=175151
I will apply this to the 4.6 branch soon.
Cheers,
Janus
This patch seems to break ia64 and some other targets. I have updated
Bug 49429 with a test case triage. It looks like for some EXPR,
may_be_aliased returns incorrect value and hence can_escape
incorrectly concludes the variable can't escape resulting in removal
of useful stores.
(So
On Fri, Jun 17, 2011 at 3:18 PM, Fang, Changpeng changpeng.f...@amd.com wrote:
Hi,
I added AVX256_SPLIT_UNALIGNED_STORE to ix86_tune_indices
and put m_COREI7, m_BDVER1 and m_GENERIC as the targets that
enable it.
Is this OK?
Can you do something similar to how
Read the comment in the diff for all the details.
I found this while working on my current patch, adding some assertion at the
end of the function would create a new build error and moving it around in the
function would stop showing the error.
I discussed this with Collin and Jeff and it
There were two pph_start_record functions, one for the in part, one for the out
part. I renamed them to have different names to avoid confusion for the reader
(and it also makes it easier for tags).
2011-06-17 Gabriel Charette gch...@google.com
* gcc/cp/pph-streamer-in.c
On 06/17/2011 12:02 PM, Janne Blomqvist wrote:
PING.
Also, here's a slightly simplfied testcase:
Yes, OK
Jerry
The attached patch moves the linkage on $(LDLIBICONV) from libgcj.spec.in to
a more correct
direct linkage on libgcj which contains the unresolved symbols. This change
also solves problems
when a multilib only has libiconv for one of the two architectures (ie building
gcc trunk on
x86_64
Another special case for the rule that an rvalue reference to function
is an lvalue.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit df1f90245e5436ceadf6b3df5cc6f7ef7e0536c0
Author: Jason Merrill ja...@redhat.com
Date: Fri Jun 17 16:45:44 2011 -0400
PR c++/49458
* call.c
32 matches
Mail list logo
