[cairo] postscript genarated by 1.5.20 blocks Dell5100cn

[Carl Worth]

On Thu, 10 Apr 2008 18:55:52 +0200, "Jiri Cerny" wrote:
> I recently updated to rawhide with new cairo 1.5.20. Now, when I try
> to print a pdf file it simply blocks our Dell5100cn printer for more
> than 60 minutes (may be more, I switched off the printer after this
> time. It is not possible to cancel the job).  It happens with at least
> 3 different PDF's so it does not seem to be file dependent.

Hi Jiri,

Thank you very much for reporting this bug.

It may not be file-dependent, but we could definitely use your help
debugging this problem, (particularly since none of us have direct
access to such a printer that I know of).

So one thing that would be helpful first of all is if you could
generate the smallest possible PostScript file that causes a problem
with your printer. Then if you can send that to us, we can start
debugging further.

> I did some bisecting and what I get is following:
...
>     Make PS backend use cairo-pdf-operators.c

Thanks for bisecting down to that.

What this commit does is define some, (very small), PostScript
functions at the beginning of the PostScript output. These PostScript
functions are then called later in the file for creating paths,
stroking, and positioning glyphs.

We designed these functions so that they are named identically to (and
act compatibly with) operators within the PDF specification. This
allows the PostScript and PDF backends within cairo to emit identical
output for large portions output-file, (reducing code and bug
duplication within cairo).

> The postscripts files that are generated during the bisecting are
> always viewable by gv (Some of them have some moved glyphs), so there
> is probably some bad interaction with the printer.
>
> It might be related to bug 15348 ?

Yes, in both cases it's likely that there are printer bugs in dealing
with the PostScript file, (and it's also possible that cairo's
PostScript output is in error). And as discussed in that bug report,
we're going to need more information on exactly what it is that the
printer doesn't like. So once we have a minimal file, (as can be
produced by cairo), that causes problems with the printer, the next
step would be to trim that file down further by hand until the problem
is as isolated as possible.

Then the problem can be reported to the vendor of the printer, and we
can also see what we can do to workaround the problem in cairo.

> I can post the files and do some other testing, but better tomorow.

Yes, please, as described above.

Thanks so much for your report,

-Carl
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20080410/04ae3fbe/attachment.pgp