On 4/18/11, Matthew Dempsky wrote:
> On Sun, Apr 17, 2011 at 5:29 PM, Philip Guenther wrote:
>> I haven't worked through all the bit-twiddling and asm, but what I
>> have looks good and the abundance of regress tests that keep showing
>> up in "cvs up -d" inspires confidence. With that x87-->__x
On Sun, Apr 17, 2011 at 5:29 PM, Philip Guenther wrote:
> I haven't worked through all the bit-twiddling and asm, but what I
> have looks good and the abundance of regress tests that keep showing
> up in "cvs up -d" inspires confidence. With that x87-->__x87 nit, ok
> guenther@
ok matthew@ as we
I only see one remaining namespace glitch:
On Sun, Apr 17, 2011 at 4:11 AM, Martynas Venckus
wrote:
...
> +/*
> + * fenv_t represents the entire floating-point environment
> + */
> +typedef struct {
> + struct {
> + unsigned int __control; /* Control word register */
>
The diff I sent before has been corrupted by Gmail (wrapped diff lines).
I'm re-attaching it below.
Index: include/Makefile
===
RCS file: /cvs/src/include/Makefile,v
retrieving revision 1.157
diff -u -r1.157 Makefile
--- include/Make
On 3/22/11, Mark Kettenis wrote:
>> Date: Sun, 20 Mar 2011 03:08:41 +0200
>> From: Martynas Venckus
>>
>> Index: include/fenv.h
>> ===
>> RCS file: include/fenv.h
>> diff -N include/fenv.h
>> --- /dev/null1 Jan 1970 00:00:00
> Date: Sun, 20 Mar 2011 03:08:41 +0200
> From: Martynas Venckus
>
> Index: include/fenv.h
> ===
> RCS file: include/fenv.h
> diff -N include/fenv.h
> --- /dev/null 1 Jan 1970 00:00:00 -
> +++ include/fenv.h18 Mar 2011 20:38:
-0800 (PST)
> Received: from mdempsky.mtv.corp.google.com (mdempsky.mtv.corp.google.com
> [172.17.81.82])
> by mx.google.com with ESMTPS id n15sm530297fam.36.2011.03.08.11.50.51
> (version=SSLv3 cipher=OTHER); Tue, 08 Mar 2011 11:50:53 -0800 (PST)
> Date: Tue, 8 Mar 2011 11:50:4
> Date: Sun, 13 Mar 2011 20:22:23 -0700
> From: Philip Guenther
>
> These headers violate POSIX namespace rules in various ways:
> - pulling in , , and
> defines names that aren't permitted.
>(seems unnecessary too; just use unsigned short/int/long long
> instead of uint16_t/uint32_t/uint64
On Tue, Mar 8, 2011 at 11:50 AM, Matthew Dempsky wrote:
> Currently, the diff only supports amd64, i386, and sparc64 (the only
> arches that NetBSD supports fenv.h on),
I'd also appreciate help on adding support for the other platforms.
Thanks!
On Sun, Mar 13, 2011 at 8:22 PM, Philip Guenther wrote:
> Hmm, any test programs that are smaller and preferably not need X11?
> My little V100 takes forever with a build as is...
I don't know of any at the moment. I'll try to look for something.
>> You should be using __sparc64__ instead of __
On Sun, Mar 13, 2011 at 9:14 AM, Brad wrote:
> On 08/03/11 2:50 PM, Matthew Dempsky wrote:
>> The diff below adds support for C99's to libm. It's based on
>> NetBSD's implementation with minimal changes to work on OpenBSD.
>>
>> Currently, the diff only supports amd64, i386, and sparc64 (the onl
On 08/03/11 2:50 PM, Matthew Dempsky wrote:
The diff below adds support for C99's to libm. It's based on
NetBSD's implementation with minimal changes to work on OpenBSD.
Currently, the diff only supports amd64, i386, and sparc64 (the only
arches that NetBSD supports fenv.h on), but I've only t
12 matches
Mail list logo
