Eric Blake | 4 Apr 03:29 2007
Picon

Re: stdint.h bug


According to Christopher Faylor on 4/3/2007 1:13 PM:
> On Tue, Apr 03, 2007 at 06:15:14PM +0000, Eric Blake wrote:
>> Cygwin defines wint_t as unsigned int (valid, per POSIX), but then defines
>>
>> #ifndef WINT_MIN
>> #define WINT_MIN (-2147483647 - 1)
>> #define WINT_MAX (2147483647)
>> #endif
>>
>> which is invalid given the underlying type of wint_t.  Can we get this fixed 
>> (either make wint_t a signed type, or change WINT_MIN and WINT_MAX)?
> 
> Patch?

Well, I was hoping for some feedback as to whether changing wint_t to be
signed was preferable over changing WINT_MIN.  But further research shows
Linux also uses an unsigned type for wint_t, so:

2007-04-03  Eric Blake  <ebb9 <at> byu.net>

	* include/stdint.h (WINT_MIN, WINT_MAX): Fix definition.

--
Don't work too hard, make some time for fun as well!

Eric Blake             ebb9 <at> byu.net
Index: include/stdint.h
(Continue reading)

Corinna Vinschen | 4 Apr 09:38 2007

Re: stdint.h bug

On Apr  3 19:29, Eric Blake wrote:
> 	* include/stdint.h (WINT_MIN, WINT_MAX): Fix definition.

Thanks, applied.

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Ryan C. Gordon | 4 Apr 10:26 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...


> mnt_type is always "system" or "user" ... this patch changes this to 
> make an earnest effort to match what a GNU/Linux system would report, 
> and moves the system/user string to mnt_opts.

I sent in the copyright assignment paperwork for this around two weeks 
ago...just wanted to follow up to see if that was ever received, and if 
so, if this patch can be committed or needs further work.

Thanks,
--ryan.

Corinna Vinschen | 4 Apr 10:49 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...

On Apr  4 04:26, Ryan C. Gordon wrote:
> 
> >mnt_type is always "system" or "user" ... this patch changes this to 
> >make an earnest effort to match what a GNU/Linux system would report, 
> >and moves the system/user string to mnt_opts.
> 
> I sent in the copyright assignment paperwork for this around two weeks 
> ago...just wanted to follow up to see if that was ever received, and if 
> so, if this patch can be committed or needs further work.

Sorry, I didn't get the note from our dept so far.  I'll investigate...

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Corinna Vinschen | 4 Apr 18:03 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...

Ryan,

On Apr  4 10:49, Corinna Vinschen wrote:
> On Apr  4 04:26, Ryan C. Gordon wrote:
> > 
> > >mnt_type is always "system" or "user" ... this patch changes this to 
> > >make an earnest effort to match what a GNU/Linux system would report, 
> > >and moves the system/user string to mnt_opts.
> > 
> > I sent in the copyright assignment paperwork for this around two weeks 
> > ago...just wanted to follow up to see if that was ever received, and if 
> > so, if this patch can be committed or needs further work.
> 
> Sorry, I didn't get the note from our dept so far.  I'll investigate...

your assignment arrived and has been signed.

Chris, are you going to review the patch?

Thanks,
Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Christopher Faylor | 4 Apr 20:18 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...

On Wed, Apr 04, 2007 at 06:03:09PM +0200, Corinna Vinschen wrote:
>Ryan,
>
>On Apr  4 10:49, Corinna Vinschen wrote:
>> On Apr  4 04:26, Ryan C. Gordon wrote:
>> > 
>> > >mnt_type is always "system" or "user" ... this patch changes this to 
>> > >make an earnest effort to match what a GNU/Linux system would report, 
>> > >and moves the system/user string to mnt_opts.
>> > 
>> > I sent in the copyright assignment paperwork for this around two weeks 
>> > ago...just wanted to follow up to see if that was ever received, and if 
>> > so, if this patch can be committed or needs further work.
>> 
>> Sorry, I didn't get the note from our dept so far.  I'll investigate...
>
>your assignment arrived and has been signed.
>
>Chris, are you going to review the patch?

Yep.  I'll do it in the next couple of days.

cgf

Christopher Faylor | 4 Apr 20:19 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...

On Wed, Apr 04, 2007 at 02:18:33PM -0400, Christopher Faylor wrote:
>On Wed, Apr 04, 2007 at 06:03:09PM +0200, Corinna Vinschen wrote:
>>Ryan,
>>
>>On Apr  4 10:49, Corinna Vinschen wrote:
>>> On Apr  4 04:26, Ryan C. Gordon wrote:
>>> > 
>>> > >mnt_type is always "system" or "user" ... this patch changes this to 
>>> > >make an earnest effort to match what a GNU/Linux system would report, 
>>> > >and moves the system/user string to mnt_opts.
>>> > 
>>> > I sent in the copyright assignment paperwork for this around two weeks 
>>> > ago...just wanted to follow up to see if that was ever received, and if 
>>> > so, if this patch can be committed or needs further work.
>>> 
>>> Sorry, I didn't get the note from our dept so far.  I'll investigate...
>>
>>your assignment arrived and has been signed.
>>
>>Chris, are you going to review the patch?
>
>Yep.  I'll do it in the next couple of days.

Actually, I'd appreciate it if the patch could be resubmitted against
current CVS.

cgf

Eric Blake | 6 Apr 03:36 2007
Picon

Re: stdint.h bug


According to Corinna Vinschen on 4/4/2007 1:38 AM:
> On Apr  3 19:29, Eric Blake wrote:
>> 	* include/stdint.h (WINT_MIN, WINT_MAX): Fix definition.
> 
> Thanks, applied.

A two-line patch, and I _still_ managed to botch it.  POSIX requires that
WINT_MIN be unsigned if (wint_t)0 promotes to an unsigned type.

Or in other words, (-1 < WINT_MIN) == ((wint_t) -1 < 0) must be true.

2007-04-05  Eric Blake  <ebb9 <at> byu.net>

	* include/stdint.h (WINT_MIN): Fix sign.

--
Don't work too hard, make some time for fun as well!

Eric Blake             ebb9 <at> byu.net
Index: include/stdint.h
===================================================================
RCS file: /cvs/src/src/winsup/cygwin/include/stdint.h,v
retrieving revision 1.8
diff -u -p -r1.8 stdint.h
--- include/stdint.h	4 Apr 2007 07:37:53 -0000	1.8
+++ include/stdint.h	6 Apr 2007 01:31:37 -0000
 <at>  <at>  -158,7 +158,7  <at>  <at>  typedef unsigned long long uintmax_t;
(Continue reading)

Corinna Vinschen | 6 Apr 10:25 2007

Re: stdint.h bug

On Apr  5 19:36, Eric Blake wrote:
> 	* include/stdint.h (WINT_MIN): Fix sign.

Thanks,  applied.

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Ryan C. Gordon | 6 Apr 16:49 2007

Re: [PATCH] getmntent()->mnt_type values that match Linux...


> Actually, I'd appreciate it if the patch could be resubmitted against
> current CVS.

I'll do this shortly.

--ryan.


Gmane