A blind attempt (since I'm using makeinfo 4.8 where the error does not show up)
at fixing the makeinfo errors on gnat_rm.texi
Let me know if this fixes the errors.
2014-06-14 Arnaud Charlet
PR ada/61505
* gnat_rm.texi: Attempt to fix error with makeinfo 5.1
Index: gnat_rm.tex
Hi Jakub,
Looking back into the patch, I don't see any reason to delete
vxwoks-dummy.h, looks like it was my mistake - sorry about that.
So looks like your patch fixes the issue correctly.
Thanks,
Alexander
2014-06-12 21:47 GMT+04:00 Jakub Jelinek :
> Hi!
>
> Seems http://gcc.gnu.org/r197156 eff
2014-06-14 Paul Thomas
* trans-array.c (gfc_conv_array_parameter): Assign 'old_desc'
to 'new_desc' rather than some of the components.
Bootstraps and regtests OK
Committed at revision 211664.
Cheers
Paul
This patch changes the gccgo driver to pass -t to the native linker on
Solaris. This avoids warnings like
ld: warning: symbol 'go$zerovalue' has differing sizes:
(file hello.o value=0x8; file i386-pc-solaris2.11/libgo/.libs/libgo.so
value=0x800);
hello.o definition taken and upda
On Fri, 2014-06-06 at 11:37 -0400, David Edelsohn wrote:
> On Thu, Jun 5, 2014 at 3:57 PM, Peter Bergner wrote:
> > gcc/
> > PR target/61415
> > * config/rs6000/rs6000-builtin.def (BU_MISC_1): Delete.
> > (BU_MISC_2): Rename to ...
> > (BU_LDBL128_2): ... this.
> >
Hi,
this is a new collection of tests for IPA ICF pass.
Changelog:
2014-06-13 Martin Liska
Honza Hubicka
* gcc/testsuite/g++.dg/ipa/ipa-se-1.C: New test.
* gcc/testsuite/g++.dg/ipa/ipa-se-2.C: Likewise.
* gcc/testsuite/g++.dg/ipa/ipa-se-3.C: Likewise.
Hello,
after working for quite a long time, I would like to introduce new IPA pass.
Goal of the pass is to merge semantically equivalent functions and read-only
variables.
If we prove that a function A is an equivalent to a function B, depending on
circumstances, an alias, thunk or a functio
Hi,
this small patch prepares remaining infrastructure needed for the new pass.
Changelog:
2014-06-13 Martin Liska
Honza Hubicka
* ipa-utils.h (polymorphic_type_binfo_p): Function marked external
instead of static.
* ipa-devirt.c (polymorphic_type_bin
Hi,
many tests rely on a precise number of scanned functions in a dump file. If
IPA ICF decides to merge some function and(or) read-only variables, counts do
not match.
Changelog:
2014-06-13 Martin Liska
Honza Hubicka
* c-c++-common/rotate-1.c: Text
* c-c++-c
Hi, the following patch improves LRA spilling general reg pseduos
into vector regs.
Before the patch only *regular* spilled pseudos of general regs
class are assigned to vector regs on the LRA spill sub-pass. The
patch improves this by assigning vector regs to *inheritance* pseudos
on LRA as
Richard,
Something in these constraint patches broke my mips16 build (I cannot
build glibc in mips16 mode). I have cut down a test case and verified
that the problem started with this checkin:
2014-06-11 Richard Sandiford
* common.md: New file.
* doc/md.texi: Update descripti
Hi,
On 06/13/2014 08:11 PM, Jason Merrill wrote:
On 06/13/2014 01:45 PM, Paolo Carlini wrote:
But it doesn't work here, doesn't do anything fancy, it exactly prints
'v' and nothing else.
That seems to be a bug in type_to_string, which sees that "void"
starts with "v" and concludes that they
> >My concern is that peepholes are rather fragile, so imho it is not
> >inconceivable that some target will generate wrong code when you add
> >an extra (later) peephole pass. Of course, we are in stage1.
> peephole2 is significantly robust than the original peephole pass;
> largely because pee
On 13 June 2014 21:33, Jason Merrill wrote:
> On 06/13/2014 02:05 PM, Ville Voutilainen wrote:
>>
>> that example has ones that have the wrong underlying type.
>
>
> Good point. That was fixed after the DR was incorporated into the WP, which
> now says
>
>
> template struct A {
> enum E : T;
> enu
On 06/13/2014 02:05 PM, Ville Voutilainen wrote:
that example has ones that have the wrong underlying type.
Good point. That was fixed after the DR was incorporated into the WP,
which now says
template struct A {
enum E : T;
enum class S : T;
};
template<> enum A::E : int { eint }; // OK
tem
On Thu, Jun 12, 2014 at 07:47:49PM +0200, Jakub Jelinek wrote:
> Seems http://gcc.gnu.org/r197156 effectively reverted
> the PR45078 fix for arm*-linux* (where unfortunately tm_file
> is always overridden).
>
> Was the removal of vxworks-dummy.h from that line intentional
> or just some mistake?
>
On 06/13/2014 01:45 PM, Paolo Carlini wrote:
But it doesn't work here, doesn't do anything fancy, it exactly prints
'v' and nothing else.
That seems to be a bug in type_to_string, which sees that "void" starts
with "v" and concludes that they are the same. Oops.
Jason
On 13 June 2014 19:56, Jason Merrill wrote:
> This needs a test that we complain about a specialization of an unscoped
> enum. Perhaps just the test from 14.7p6??
Well, a specialization of an unscoped enum with the proper underlying type
is fine, and that example has ones that have the wrong un
Hi,
On 06/13/2014 07:39 PM, Jason Merrill wrote:
On 06/13/2014 01:02 PM, Paolo Carlini wrote:
Don't strip_typedefs; the error message should name the typedef.
But then, are we going to just print 'v'?!? We want "typedef-name for
type 'v'"?!? We already have the caret exactly below the 'v'... P
On 06/13/2014 01:02 PM, Paolo Carlini wrote:
Don't strip_typedefs; the error message should name the typedef.
But then, are we going to just print 'v'?!? We want "typedef-name for
type 'v'"?!? We already have the caret exactly below the 'v'... Please
clarify which exact message you want to see.
> On Fri, 13 Jun 2014, Jan Hubicka wrote:
>
> > >
> > > When you extract the address and use it. For example when you
> > > do auto-parallelization and outline a part of your function it
> > > passes arrays as addresses.
> > >
> > > Or if you start to introduce address induction variables like
On 06/13/2014 06:59 PM, Paolo Carlini wrote:
Hi,
On 06/13/2014 06:49 PM, Jason Merrill wrote:
On 06/13/2014 12:42 PM, Paolo Carlini wrote:
+ "invalid use of typedef-name for type %qT in "
+ "parameter declaration", strip_typedefs (type));
Don't strip_typedefs; the
Hi,
On 06/13/2014 06:49 PM, Jason Merrill wrote:
On 06/13/2014 12:42 PM, Paolo Carlini wrote:
+ "invalid use of typedef-name for type %qT in "
+ "parameter declaration", strip_typedefs (type));
Don't strip_typedefs; the error message should name the typedef.
But then
2014-06-13 17:58 GMT+02:00 Jeff Law :
> On 06/13/14 09:56, Richard Henderson wrote:
>>
>> On 06/13/2014 08:36 AM, Jeff Law wrote:
>>>
>>> So you may have answered this already, but why can't this be a combiner
>>> pattern?
>>
>>
>> Until pass_duplicate_computed_gotos, we (intentionally) have a sing
This needs a test that we complain about a specialization of an unscoped
enum. Perhaps just the test from 14.7p6??
template struct A {
enum E: T;
enum class S: T;
};
template<> enum A::E: int { eint }; // OK
template<> enum class A::S: int { sint }; // OK
template en
On 06/13/2014 12:42 PM, Paolo Carlini wrote:
+ "invalid use of typedef-name for type %qT in "
+ "parameter declaration", strip_typedefs (type));
Don't strip_typedefs; the error message should name the typedef.
Jason
On 2014-06-11, 1:17 PM, Chung-Lin Tang wrote:
> On 2014/6/11 下午 06:32, James Greenhalgh wrote:
>>
>> Hi,
>>
>> A recent change somewhere exposed a latent bug between LRA and the definition
>> of the movsi_compare0 pattern.
>>
>> This pattern ties the source and destination register of a set togethe
Hi,
On 06/13/2014 06:09 PM, Jason Merrill wrote:
On 06/13/2014 06:58 AM, Paolo Carlini wrote:
+error_at (DECL_SOURCE_LOCATION (decl),
+ "invalid use of typedef-name for type "
+ "% in parameter declaration");
+ else if (cv_qualified_p (type))
+erro
On 06/11/14 15:45, Ilya Enkovich wrote:
On 11 Jun 22:47, Uros Bizjak wrote:
This should read:
/* { dg-do compile { target { ia32 } } } */
/* { dg-options "-O2 -march=corei7 -mfpmath=387" } */
The x86 part is OK with this change.
Uros.
Thanks for comment! Here is a fixed version.
Ilya
--
On 06/13/2014 06:58 AM, Paolo Carlini wrote:
+ error_at (DECL_SOURCE_LOCATION (decl),
+ "invalid use of typedef-name for type "
+ "% in parameter declaration");
+ else if (cv_qualified_p (type))
+ error_at (DECL_SOURCE_LOCATION (
On 06/13/14 09:56, Richard Henderson wrote:
On 06/13/2014 08:36 AM, Jeff Law wrote:
So you may have answered this already, but why can't this be a combiner pattern?
Until pass_duplicate_computed_gotos, we (intentionally) have a single indirect
branch in the entire function. This vastly reduce
On 06/13/2014 08:36 AM, Jeff Law wrote:
> So you may have answered this already, but why can't this be a combiner
> pattern?
Until pass_duplicate_computed_gotos, we (intentionally) have a single indirect
branch in the entire function. This vastly reduces the size of the CFG.
Peep2 is currently
On 06/13/14 03:56, Kyrill Tkachov wrote:
Hi all,
I noticed a memory corruption bug while adding some scheduler bypasses
in the arm backend.
genattrtab would segfault while processing the bypasses. Valgrind
confirmed this.
The problem is that when processing the bypassed reservations,
make_autom
On 06/12/14 17:38, Ilya Enkovich wrote:
It looks ok to me - did you test with all languages? In particular did
you test Ada?
I configure compiler with no language disabling and then run bootstrap
and make check. Does it mean all languages are covered? Will make more
testing if required.
In add
On 06/12/14 10:21, Kai Tietz wrote:
Hello,
I updated i386.md part of the patch. Initial patch included handling
of blockage, which is obviously superflous. Additionally I merged
32-bit and 64-bit peephole2 versions by using mode-specifier W.
ChangeLog
2014-06-12 Kai Tietz
* config/i3
On 06/12/14 15:06, Segher Boessenkool wrote:
Will that work on other targets?
Well, this is the only point I am a bit concerned too. In general I
wouldn't expect here any issues to run peephole after scheduling, as
peephole doesn't do anything a new run of ira/lra would require.
My concern
This patch to libgo stop a backtrace at a few recognized functions.
On x86_64 Solaris the makecontext function does not properly indicate
that it is at the top of the stack. Attempting to unwind the stack past
a call to makecontext tends to crash. This patch changes libgo to look
for certain fun
This patch to libgo: always sets gcnext_sp to a pointer-aligned address.
The gcnext_sp field is only used on systems that do not use split
stacks. It marks the bottom of the stack for the garbage collector.
This change makes sure that the stack bottom is always aligned to a
pointer value.
Previo
Hi Guys,
I am applying the patch below to fix a small problem with the RX port
- it was using non-log based alignment values for jumps, loops and
labels when user specified alignment was enabled.
Cheers
Nick
gcc/ChangeLog
2014-06-13 Nick Clifton
* config/rx/rx.h (JUMP_ALIGN):
Hi,
On Thu, 12 Jun 2014 12:49:13, Richard Biener wrote:
>
> On Thu, Jun 12, 2014 at 10:36 AM, Eric Botcazou wrote:
>>> Btw, I wonder if we can simply mark the MEMs generated from spill code
>>> with MEM_NOTRAP_P so we can remove the special casing of
>>> frame-pointer-based addresses from add whi
On 06/10/2014 08:34 AM, Jan Hubicka wrote:
Hi,
ipa-reference is somewhat stupid and builds its data sets for all variables
including
addressable and public one just to prune them out after all bitmaps are
constructed.
This used to make sense when the profile generation happened at compile time,
On 15/05/14 11:38 +0100, Jonathan Wakely wrote:
On 15/05/14 07:36 +0200, Daniel Krügler wrote:
Looking at the definition of the new alias
__cv_tuple_size
you might want to apply LWG 2313
http://cplusplus.github.io/LWG/lwg-defects.html#2313
and simplify its definition even further.
I forgot
On 15/05/14 13:30 +0100, Jonathan Wakely wrote:
We are missing specializations for wchar_t, char16_t and char32_t.
The preprocessor condition I've added at the top of is
to avoid needing to #include , which drops names in the
global namespace. GCC defines __UINT_LEAST16_TYPE__ and
__UINT_LEAST3
I accidentally committed r211502 with the body of a new test duplicated twice.
Removed in r211635.
My apologies!
--AlanIndex: gcc/testsuite/gcc.dg/vect/vect-singleton_1.c
===
--- gcc/testsuite/gcc.dg/vect/vect-singleton_1.c (revisi
Hi Cesar!
On Wed, 4 Jun 2014 20:42:16 -0700, Cesar Philippidis
wrote:
> The way that the loop clause
> is handle in fortran is that all loops get lowered with the collapse
> clause set. By default, for non-concurrent loops, collapse is set to 1.
> And when collapse == 1, nothing special happens
Hi Ilmir!
On Wed, 4 Jun 2014 17:53:38 +0400, Ilmir Usmanov wrote:
> you should update tree-pretty-print.c (I forgot this):
> @@ -675,13 +675,13 @@ dump_omp_clause (pretty_printer *buffer, tree
> clause, int spc, int flags)
>
> case OMP_CLAUSE_WORKER:
> pp_string (buffer, "worker("
Hi!
On Wed, 4 Jun 2014 20:58:06 -0700, Cesar Philippidis
wrote:
> On 06/04/2014 06:53 AM, Ilmir Usmanov wrote:
> >> @@ -1817,13 +1818,9 @@ scan_sharing_clauses (tree clauses, omp_context
> >> *ctx)
> >> case OMP_CLAUSE_PRIVATE:
> >> case OMP_CLAUSE_FIRSTPRIVATE:
> >> case OMP_C
How do people feel about this change, to add a #warning to the
header saying it's not implemented?
This might stop people reporting bugs that it doesn't work (we know!)
or complaining that the code shouldn't be shipped if it doesn't work.
Now that we have an implementation in the 4.9.0 release
The following leverages the "extra" work done by FRE/PRE now
(propagating copies and constants) and removes the copy-prop
pass run during early optimizations (no passes after it expose
copy propagation opportunities). It also moves the 3rd
copy-prop pass from right after FRE (where it is now poin
Hi again ;)
On 06/13/2014 12:20 PM, Paolo Carlini wrote:
... nope, sorry, the latter doesn't pass testing, accepts:
(void = 0)
and void2.C regresses. Thus I propose something conservative, like the
below.
.. or tackle directly void2.C and use:
if (type == void_type_node
This patch detects cases where we can tell at compile time that an
address clause value is compatible with the alignment of the object
so that we do not need to issue a warning.
The following is compiled with -gnatwa -gnatld7 -gnatj55
1. pragma Restrictions (No_Exception_Propagation);
2
... nope, sorry, the latter doesn't pass testing, accepts:
(void = 0)
and void2.C regresses. Thus I propose something conservative, like the
below.
Thanks,
Paolo.
///
Index: cp/decl.c
===
--- cp/decl.c (rev
In Float_Check_Overflow mode, Succ applied to type'Last or Pred applied
to type'First generates a constraint error since the argument is out of
range. This was not previously changed, the following test:
1. with Ada.Exceptions; use Ada.Exceptions;
2. with Text_IO; use Text_IO;
3. pr
Global variable Root_Environment was used in the Project Manager,
but was not initialized by GNATCOLL and GPS. This patch eliminates
the direct use of Root_Environment.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-06-13 Vincent Celier
* makeutl.ads (Compute_Builder_Switches)
This patch fixes the handling of attribute reference 'Old in the presence
of Assertion_Policy (Checked) pragma, when a unit is compiled without the
-gnata flag.
Compiling and executing the following:
gnatmake -q assertion_policy_test.adb
assertion_policy_test
Must yield:
+ Ass
This patches fixes the use of custom parsers when trying to expand
command line arguments like "*.adb".
When run from the test directory, the following program should output
"next source >>> test_cmd_line1.adb".
with Ada.Text_IO; use Ada.Text_IO;
with GNAT.Command_Line; use GNAT.Comma
We now check for null statement sequences, and for extended return
statements. In addition, the message generated now includes a
description of the non-permitted construct as shown in this
test program (compiled with -gnatj60 -gnatl)
1. pragma Ada_2012;
2. pragma Compiler_Unit_Warning;
On Wed, 14 May 2014, Richard Biener wrote:
>
> This makes FRE/PRE substitute values into all uses instead of leaving
> copies and initializations from constants in the IL which requires
> a copyprop pass to clean up things (which we usually place directly
> after FRE/PRE).
>
> This should open t
The functionality is now provided in a target-independent manner
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-06-13 Geert Bosch
* gnat_rm.texi, s-tasinf-solaris.ads, sem_prag.adb, gnat_ugn.texi,
s-tasinf-mingw.ads, s-tasinf.ads, s-tasinf-linux.ads,
s-tasinf-v
Seems sth easily forgotten (4.8.3 and 4.7.4 are examples of that).
We should make sure to at least update displayed copyright years.
Committed.
Richard.
2014-06-13 Richard Biener
* releasing.html: Add step to verify displayed copyright years.
Index: releasing.html
=
Hi all,
I noticed a memory corruption bug while adding some scheduler bypasses
in the arm backend.
genattrtab would segfault while processing the bypasses. Valgrind
confirmed this.
The problem is that when processing the bypassed reservations,
make_automaton_pairs allocates memory in proport
Hi,
On 06/12/2014 11:20 PM, Jason Merrill wrote:
On 06/12/2014 03:14 PM, Paolo Carlini wrote:
... in terms of code proper, the below is much better, IMHO. Assuming,
as I understand, we have no reason to call the rather heavy same_type_p
when we already know that VOID_TYPE_P (type) is true...
This patch fixes an obscure bug that causes the secondary stack to be used
before it is initialized in certain cases. This can only happen if (1) the
-gnatE switch is used to disable the static elaboration mode, (2) the -p
switch is passed to gnatbind to tell it to choose a pessimistic (worst-case)
This patch removes a spurious error on a unit to which the Preelaborate
pragma applies. The error appeared on a unit that holds an instantiation of
a package containing a type declaration with an array component whose default
value is given by an actual in the instance, but the error may occur in o
Aspects Import and Export were not treated like the equivalent pragmas
wrt issuing warnings on missing initialization before use. Now fixed.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-06-13 Yannick Moy
* sem_ch13.adb (Analyze_Aspect_Specifications/Aspect_Import,
As
The Refined_Post aspect defined in SPARK 2014 should be considered like
a postcondition wrt issuing warnings on variable references.
Tested on x86_64-pc-linux-gnu, committed on trunk
2014-06-13 Yannick Moy
* sem_warn.adb (Check_Unset_References): Take
case of Refined_Post into
This patch enables the recognition/processing of pragma Restrictions
(No_Dependence => unit) in system.ads, allowing more flexibility in
configuring specialized versions of System.
Given a system.ads that contains the line
pragma Restrictions (No_Dependence => Ada.Text_IO);
Compiling the followi
On Fri, 13 Jun 2014, Jan Hubicka wrote:
> >
> > When you extract the address and use it. For example when you
> > do auto-parallelization and outline a part of your function it
> > passes arrays as addresses.
> >
> > Or if you start to introduce address induction variables like
> > the vectoriz
On Fri, Jun 13, 2014 at 1:38 AM, Ilya Enkovich wrote:
> 2014-06-12 11:55 GMT+04:00 Richard Biener :
>> On Wed, Jun 11, 2014 at 6:08 PM, Ilya Enkovich
>> wrote:
>>> Hi,
>>>
>>> This patch fixes problem with size emitted for static structures with
>>> flexible array. I found a couple of trackers
On Fri, Jun 13, 2014 at 12:14 AM, Jan Hubicka wrote:
> Hi,
> while updating vect_can_force_dr_alignment_p for section API I noticed the
> predicate is bit confused about when it can update the alignment.
>
> We need to check that decl_binds_to_current_def_p and in case we compile
> a partition als
On Wed, Jun 11, 2014 at 06:17:47PM +0100, Chung-Lin Tang wrote:
> On 2014/6/11 下午 06:32, James Greenhalgh wrote:
> >
> > Hi,
> >
> > A recent change somewhere exposed a latent bug between LRA and the
> > definition
> > of the movsi_compare0 pattern.
> >
> > This pattern ties the source and dest
On 12 June 2014 18:04, Yufeng Zhang wrote:
> Hi,
>
> The patch fixes a bug in the AArch64 backend in calculating the beginning
> address of the unnamed incoming arguments on the stack, i.e. the initial
> value of __va_list->__stack. aarch64_layout_arg incorrectly calculates the
> size of named ar
2014-06-12 23:46 GMT+02:00 David Wohlferd :
>
> On 6/12/2014 9:21 AM, Kai Tietz wrote:
>>
>> with addition of adding a second peephole2 pass after sched2 pass, I
>> was able to get some improvement for PR target/39284. I think by this
>> addition we can close bug as fixed.
>> Additionally addition
73 matches
Mail list logo
