David Dawes wrote:
> 
> On Thu, Feb 19, 2004 at 11:20:50PM +0100, Roland Mainz wrote:
> >David Dawes wrote:
> >> >David Dawes wrote:
> >> >> CVSROOT:        /home/x-cvs
> >> >> Module name:    xc
> >> >> Changes by:     [EMAIL PROTECTED]       04/02/11 13:11:26
> >> >>
> >> >> Log message:
> >> >>    799. Some more font path checks.
> >> >>
> >> >> Modified files:
> >> >>       xc/lib/font/fontfile/:
> >> >>         dirfile.c encparse.c fontfile.c
> >> >>       xc/programs/Xserver/hw/xfree86/:
> >> >>         CHANGELOG
> >> >>
> >> >>   Revision      Changes    Path
> >> >>   3.18          +17 -1     xc/lib/font/fontfile/dirfile.c
> >> >>   1.20          +7 -2      xc/lib/font/fontfile/encparse.c
> >> >>   3.22          +30 -11    xc/lib/font/fontfile/fontfile.c
> >> >>   3.3139        +2 -1      xc/programs/Xserver/hw/xfree86/CHANGELOG
> >> >
> >> >David:
> >> >Somehow these changes broke Xprt's handing of printer builtin fonts
> >> >(e.g. font paths prefixed with "PRINTER:" which are "enabled"
> >> >dynamically on per-model-config basis).
> >>
> >> Can you isolate which of the changes causes the problem by adding them
> >> in one at a time?
> >
> >Yes, it seems that my original observation was wrong. Not the
> >printer-builtin fonts are affected but parts of the font path are
> >dropped.
> >The following statement in xc/lib/font/fontfile/dirfile.c causes the
> >failure:
> >(from
> >http://xprint.freedesktop.org/cgi-bin/bugzilla/attachment.cgi?id=95&action=view)
> >-- snip --
> >+       }
> >+       if (!found_font) {
> >+           FontFileFreeDir (dir);
> >+           fclose(file);
> >+           return BadFontPath;
> >        }
> >-- snip --
> >It seems that the change now makes it mandatory that the Xserver allows
> >to drop invalid font path elements... ;-/
> 
> I guess that's this change:
> 
>   70. Changed behavior of fontfile: don't drop the entire directory if some
>       fonts cannot be rendered (Egbert Eich).
> 
> What exactly is it doing that breaks the printer fonts?

It seems that my original observation about missing printer buildin
fonts was wrong. The server refused to start and I thought that this was
caused by problems with the printer buildin fonts (note that I have
modified the font code in my own tree - if someone tries to start Xprt
with an invalid (or missing) configuration the server now refuses to
start (this is mainly to get rid of the myths about Xprint caused by
server misconfiguration)). After some more debugging I realised that the
dropping of font path elements isn't supported by my codebase (since I
am working on a X11R6.6-based tree which misses the code which can drop
parts of the font path without aborting the server startup) so the
server fails to start. My fault... ;-(

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) [EMAIL PROTECTED]
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 2426 901568 FAX +49 2426 901569
 (;O/ \/ \O;)
_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel

Reply via email to