* <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
@ 2004-01-18 19:31 Timothy J. Wood
2004-01-18 19:49 ` [PATCH] [COMMITTED] Fix " Andrew Pinski
2004-01-18 23:23 ` Ziemowit Laski
0 siblings, 2 replies; 6+ messages in thread
From: Timothy J. Wood @ 2004-01-18 19:31 UTC (permalink / raw)
To: gcc
Configured thusly:
../gcc/configure --prefix="$PREFIX" --enable-languages="c,c++,objc"
Compiling just:
#include <altivec.h>
with:
$PREFIX/bin/c++ -O3 -maltivec x.cpp -S -o x.s
yields zillions of errors of the form:
In file included from x.cpp:2:
/Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:124: error:
declaration of C function `float __vector__ vec_ctf(int __vector__,
char)' conflicts with
/Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:123: error:
previous declaration `float __vector__ vec_ctf(unsigned int __vector__,
char)' here
...
-tim
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH] [COMMITTED] Fix Re: <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
2004-01-18 19:31 <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0 Timothy J. Wood
@ 2004-01-18 19:49 ` Andrew Pinski
2004-01-22 6:27 ` Segher Boessenkool
2004-01-18 23:23 ` Ziemowit Laski
1 sibling, 1 reply; 6+ messages in thread
From: Andrew Pinski @ 2004-01-18 19:49 UTC (permalink / raw)
To: Timothy J. Wood; +Cc: Gcc Patch List, Andrew Pinski, gcc mailing list
The problem is that on darwin, the system headers are defined to c
headers as some of them
(mach ones) are C++ complaint.
I committed this to the 3.4 branch and the mainline as obvious.
Thanks,
Andrew Pinski
ChangeLog:
2004-01-18 Andrew Pinski <pinskia@physics.uc.edu>
* config/rs6000/altivec.h: Wrap C++ functions in extern "C++"
block.
Patch:
Index: config/rs6000/altivec.h
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/rs6000/altivec.h,v
retrieving revision 1.25
diff -u -p -r1.25 altivec.h
--- config/rs6000/altivec.h 16 Dec 2003 07:42:08 -0000 1.25
+++ config/rs6000/altivec.h 18 Jan 2004 19:48:58 -0000
@@ -118,6 +118,8 @@ extern int __altivec_link_error_invalid_
#ifdef __cplusplus
+extern "C++" {
+
/* Prototypes for builtins that take literals and must always be
inlined. */
inline vector float vec_ctf (vector unsigned int, const char)
__attribute__ ((always_inline));
@@ -6117,6 +6119,8 @@ struct __vec_step_help<vector float>
#define vec_step(t) __vec_step_help<typeof(t)>::_S_elem
+}//extern "C++"
+
#else /* not C++ */
/* "... and so I think no man in a century will suffer as greatly as
@@ -8532,6 +8536,7 @@ __ch (__bin_args_eq (vector float, (a1),
#define vec_any_numeric(a1) __builtin_altivec_vcmpeqfp_p
(__CR6_EQ_REV, (a1), (a1))
#define vec_any_out(a1, a2) __builtin_altivec_vcmpbfp_p (__CR6_EQ_REV,
(a1), (a2))
+
#endif /* __cplusplus */
On Jan 18, 2004, at 11:31, Timothy J. Wood wrote:
> Configured thusly:
>
> ../gcc/configure --prefix="$PREFIX" --enable-languages="c,c++,objc"
>
> Compiling just:
>
> #include <altivec.h>
>
> with:
>
> $PREFIX/bin/c++ -O3 -maltivec x.cpp -S -o x.s
>
> yields zillions of errors of the form:
>
> In file included from x.cpp:2:
> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
> lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:124: error:
> declaration of C function `float __vector__ vec_ctf(int __vector__,
> char)' conflicts with
> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
> lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:123: error:
> previous declaration `float __vector__ vec_ctf(unsigned int
> __vector__, char)' here
> ...
>
>
> -tim
>
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
2004-01-18 19:31 <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0 Timothy J. Wood
2004-01-18 19:49 ` [PATCH] [COMMITTED] Fix " Andrew Pinski
@ 2004-01-18 23:23 ` Ziemowit Laski
2004-01-18 23:51 ` Timothy J. Wood
1 sibling, 1 reply; 6+ messages in thread
From: Ziemowit Laski @ 2004-01-18 23:23 UTC (permalink / raw)
To: Timothy J. Wood; +Cc: gcc
Tim,
I have a patch that fixes this issue, among with many others. Alas, no
one has deigned to review it, much less approve it:
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg01111.html
--Zem
On 18 Jan 2004, at 11.31, Timothy J. Wood wrote:
>
>
>
> Configured thusly:
>
> ../gcc/configure --prefix="$PREFIX" --enable-languages="c,c++,objc"
>
> Compiling just:
>
> #include <altivec.h>
>
> with:
>
> $PREFIX/bin/c++ -O3 -maltivec x.cpp -S -o x.s
>
> yields zillions of errors of the form:
>
> In file included from x.cpp:2:
> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
> lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:124: error:
> declaration of C function `float __vector__ vec_ctf(int __vector__,
> char)' conflicts with
> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/../
> lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:123: error:
> previous declaration `float __vector__ vec_ctf(unsigned int
> __vector__, char)' here
> ...
>
>
> -tim
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
2004-01-18 23:23 ` Ziemowit Laski
@ 2004-01-18 23:51 ` Timothy J. Wood
0 siblings, 0 replies; 6+ messages in thread
From: Timothy J. Wood @ 2004-01-18 23:51 UTC (permalink / raw)
To: Ziemowit Laski; +Cc: gcc
Andrew Pinski's patch
<http://gcc.gnu.org/ml/gcc-patches/2004-01/msg01808.html> fixed
particular the problem I was having. I'm not qualified to say whether
your patch might fix other issues I'll find later, though :)
-tim
On Jan 18, 2004, at 3:22 PM, Ziemowit Laski wrote:
> Tim,
>
> I have a patch that fixes this issue, among with many others. Alas,
> no one has deigned to review it, much less approve it:
>
> http://gcc.gnu.org/ml/gcc-patches/2004-01/msg01111.html
>
> --Zem
>
> On 18 Jan 2004, at 11.31, Timothy J. Wood wrote:
>
>>
>>
>>
>> Configured thusly:
>>
>> ../gcc/configure --prefix="$PREFIX" --enable-languages="c,c++,objc"
>>
>> Compiling just:
>>
>> #include <altivec.h>
>>
>> with:
>>
>> $PREFIX/bin/c++ -O3 -maltivec x.cpp -S -o x.s
>>
>> yields zillions of errors of the form:
>>
>> In file included from x.cpp:2:
>> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/
>> ../lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:124:
>> error: declaration of C function `float __vector__ vec_ctf(int
>> __vector__, char)' conflicts with
>> /Volumes/Space/Users/bungi/Source-head/GNU/gcc/gcc-3.4/install/bin/
>> ../lib/gcc/powerpc-apple-darwin7.2.0/3.4.0/include/altivec.h:123:
>> error: previous declaration `float __vector__ vec_ctf(unsigned int
>> __vector__, char)' here
>> ...
>>
>>
>> -tim
>>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [COMMITTED] Fix Re: <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
2004-01-18 19:49 ` [PATCH] [COMMITTED] Fix " Andrew Pinski
@ 2004-01-22 6:27 ` Segher Boessenkool
2004-01-22 8:50 ` Segher Boessenkool
0 siblings, 1 reply; 6+ messages in thread
From: Segher Boessenkool @ 2004-01-22 6:27 UTC (permalink / raw)
To: Andrew Pinski; +Cc: Timothy J. Wood, Gcc Patch List, gcc mailing list
On 18-jan-04, at 20:49, Andrew Pinski wrote:
> * config/rs6000/altivec.h: Wrap C++ functions in extern "C++"
> block.
I never heard about extern "C++" -- what is it, where is it documented?
> +}//extern "C++"
This is wrong; please use a /* */ comment (yes I use std=gnu99 for all
my
own code, but not everyone does) (and coding conventions require some
more spaces, too ;-) )
Segher
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] [COMMITTED] Fix Re: <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0
2004-01-22 6:27 ` Segher Boessenkool
@ 2004-01-22 8:50 ` Segher Boessenkool
0 siblings, 0 replies; 6+ messages in thread
From: Segher Boessenkool @ 2004-01-22 8:50 UTC (permalink / raw)
To: Andrew Pinski; +Cc: Gcc Patch List, Timothy J. Wood, gcc mailing list
> I never heard about extern "C++" -- what is it, where is it documented?
Wow, it turns out, I don't know C++ at all.
>> +}//extern "C++"
>
> This is wrong; please use a /* */ comment (yes I use std=gnu99 for all
> my
> own code, but not everyone does) (and coding conventions require some
> more spaces, too ;-) )
Never mind, this was in #ifdef __cplusplus protected code, as Andrew
pointed
out to me privately. Which leaves only the spaces issue ;-)
Segher
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-01-22 8:49 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-01-18 19:31 <altivec.h> not happy with 3.4 on powerpc-apple-darwin7.2.0 Timothy J. Wood
2004-01-18 19:49 ` [PATCH] [COMMITTED] Fix " Andrew Pinski
2004-01-22 6:27 ` Segher Boessenkool
2004-01-22 8:50 ` Segher Boessenkool
2004-01-18 23:23 ` Ziemowit Laski
2004-01-18 23:51 ` Timothy J. Wood
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).