AT_* flags

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

AT_* flags

dtbartle
Hello,

I noticed that the AT_* flags (e.g. AT_FDCWD) are defined in
io/fcntl.h. Shouldn't they be defined in the platform-dependent
bits/fcntl.h? Solaris (and probably other platforms) define different
values for these flags.

-- David
Reply | Threaded
Open this post in threaded view
|

Re: AT_* flags

Ulrich Drepper
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

David Bartley wrote:
> I noticed that the AT_* flags (e.g. AT_FDCWD) are defined in
> io/fcntl.h. Shouldn't they be defined in the platform-dependent
> bits/fcntl.h? Solaris (and probably other platforms) define different
> values for these flags.

And why should this matter?  The only other supported kernel is Hurd and
they can and should use the same value.

- --
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEARECAAYFAkhSf0kACgkQ2ijCOnn/RHSr5wCgnX6Q2nIJtsmMUu+f1Jnn8246
8TkAnikXNGbTIWFBqKQgitCzDcm6I84A
=DWTr
-----END PGP SIGNATURE-----
Reply | Threaded
Open this post in threaded view
|

Re: AT_* flags

dtbartle
On Fri, Jun 13, 2008 at 10:08 AM, Ulrich Drepper <[hidden email]> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> David Bartley wrote:
>> I noticed that the AT_* flags (e.g. AT_FDCWD) are defined in
>> io/fcntl.h. Shouldn't they be defined in the platform-dependent
>> bits/fcntl.h? Solaris (and probably other platforms) define different
>> values for these flags.
>
> And why should this matter?  The only other supported kernel is Hurd and
> they can and should use the same value.

Glibc has been ported to some of the *bsd systems and might eventually
be ported to other platforms. Wouldn't it generally be a good idea to
make glibc as portable/platform-independent as is reasonable?

-- David

>
> - --
> ➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (GNU/Linux)
>
> iEYEARECAAYFAkhSf0kACgkQ2ijCOnn/RHSr5wCgnX6Q2nIJtsmMUu+f1Jnn8246
> 8TkAnikXNGbTIWFBqKQgitCzDcm6I84A
> =DWTr
> -----END PGP SIGNATURE-----
>
>
Reply | Threaded
Open this post in threaded view
|

Re: AT_* flags

jbailey
On Fri, Jun 13, 2008 at 9:34 AM, David Bartley
<[hidden email]> wrote:

> On Fri, Jun 13, 2008 at 10:08 AM, Ulrich Drepper <[hidden email]> wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> David Bartley wrote:
>>> I noticed that the AT_* flags (e.g. AT_FDCWD) are defined in
>>> io/fcntl.h. Shouldn't they be defined in the platform-dependent
>>> bits/fcntl.h? Solaris (and probably other platforms) define different
>>> values for these flags.
>>
>> And why should this matter?  The only other supported kernel is Hurd and
>> they can and should use the same value.
>
> Glibc has been ported to some of the *bsd systems and might eventually
> be ported to other platforms. Wouldn't it generally be a good idea to
> make glibc as portable/platform-independent as is reasonable?

I'm pretty certain I'm the last person who touched the Solaris port of
glibc, and that wasn't in this millennium.  Let whichever porter picks
it up next deal with it.

--
Jeff Bailey - http://www.raspberryginger.com/jbailey/
"Gay marriage will encourage people to be gay, in the same way that
hanging around tall people will make you tall."
 - Peter Gallagher
Reply | Threaded
Open this post in threaded view
|

Re: AT_* flags

Kristian Van Der Vliet
On Fri, 2008-06-13 at 13:45 -0700, Jeff Bailey wrote:

> On Fri, Jun 13, 2008 at 9:34 AM, David Bartley
> <[hidden email]> wrote:
> > On Fri, Jun 13, 2008 at 10:08 AM, Ulrich Drepper <[hidden email]> wrote:
> >> David Bartley wrote:
> >>> I noticed that the AT_* flags (e.g. AT_FDCWD) are defined in
> >>> io/fcntl.h. Shouldn't they be defined in the platform-dependent
> >>> bits/fcntl.h? Solaris (and probably other platforms) define different
> >>> values for these flags.
> >>
> >> And why should this matter?  The only other supported kernel is Hurd and
> >> they can and should use the same value.
> >
> > Glibc has been ported to some of the *bsd systems and might eventually
> > be ported to other platforms. Wouldn't it generally be a good idea to
> > make glibc as portable/platform-independent as is reasonable?
>
> I'm pretty certain I'm the last person who touched the Solaris port of
> glibc, and that wasn't in this millennium.  Let whichever porter picks
> it up next deal with it.

I imagine David is talking about the more general portability of Glibc
as a whole, not specifically it's portability to Solaris. This is a
concern of mine, as I maintain the out-of-tree port of Glibc for
Syllable, although the specific issue raised here isn't one of great
concern to me personally.

--
Vanders
http://www.syllable.org


Reply | Threaded
Open this post in threaded view
|

Re: AT_* flags

Roland McGrath
In reply to this post by dtbartle
When there are ports that actually work that need different values, then
some things can move to bits/*.h headers as needed.  We won't do it just
for its own sake before it has a practical effect to enable something real
that has good reason to require it.