Re: [xslt] Solaris 10 test failure with 1.1.15

[Daniel Veillard <veillard redhat com>]

On Fri, Nov 11, 2005 at 01:25:54PM -0600, Albert Chin wrote:
> On Fri, Nov 11, 2005 at 12:11:18PM -0500, Daniel Veillard wrote:
> > On Fri, Nov 11, 2005 at 10:40:26AM -0600, Albert Chin wrote:
> > > Any ideas on this test failure (works ok on RHEL 4/AMD64):
> > 
> >    Hum, hard ... we have our own date/time routines based on the
> > XSD ones. Recompiling date.c in libexslt/ with DEBUG_EXSLT_DATE defined
> > and testing on both systems may give a hint of what is going wrong.
> 
> Ok, patch below. Does ANSI define bit types as signed/unsigned? I
> think they're unsigned by default. I'd suggest making all bittypes
> signed or unsigned explicitly.
> 
> -- 
> albert chin (china thewrittenword com)
> 
> -- snip snip
> Index: libexslt/date.c
> ===================================================================
> --- libexslt/date.c.orig	2004-11-08 09:35:03.000000000 -0600
> +++ libexslt/date.c	2005-11-11 13:16:57.659095000 -0600
> @@ -89,7 +85,7 @@
>      unsigned int	min	:6;	/* 0 <=  min    <= 59	*/
>      double		sec;
>      unsigned int	tz_flag	:1;	/* is tzo explicitely set? */
> -    int			tzo	:11;	/* -1440 <= tzo <= 1440 */
> +    signed int		tzo	:11;	/* -1440 <= tzo <= 1440 */
>  };
>  
>  /* Duration value */

  Albert,

First sorry for the delay, I know I have a bit of backlog with your mails, but
I don't have the time right now to go over them, but I'm not forgetting you,
I will try to process them within one or two days.
And yes you're right we should flag them as signed to be 100 sure, this should
probably be changed in the schemastype module of libxml2, at least a double
check should be done.

  thanks !

Daniel

-- 
Daniel Veillard      | Red Hat http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/