[Pkg-cups-devel] Bug#322226: #322226 is two different problems that manifest similarly

[Will Aoki]

I've been investigating this problem off and on as time permits, and
I've concluded that there are actually two different problems with
subtley different sources. This accounts for the differences that Reuben
Thomas reports and for the problems I see when printing to hpijs and
ljet4 queues. The first problem is the regression of STR#207 which I
originally reported in 2005, and this problem can be fixed by using the
upstream pdftops instead of the filter script introduced in 1.1.22-6.
The alternate pdftops filter introduced in 1.3.5-2 does not appear to
fix this problem.

(The landscape -> portrait orientation problem caused by the pdftops
filter results in the output shown on pages 4, 6 and 8 of the
'cups-test-results.pdf' file I attached to message #40 of December 31,
2007.)

The second problem, the one which causes ljet4 and hpijs output to
break, is either a problem in Ghostscript or in the way CUPS or Foomatic
drive Ghostscript.  The output varies a bit depending on whether it's
sent through gs-esp or gs-gpl, as demonstrated in the attached PDF.
Using gs-esp produces portrait output that's been greatly enlarged, and
gs-gpl produces portrait output up in a corner of the page. The attached
file demonstrates this problem with output from a printer being driven
from CUPS via Foomatic and HPIJS.

The second problem does not affect output to Postscript printers.


Once I know more about the problem, I'll split this bug and assign the
cloned copy as appropriate.

-- 
William Aoki     KD7YAF    waoki at umnh.utah.edu    5-1924
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gs-problem-demo.pdf
Type: application/pdf
Size: 81218 bytes
Desc: not available
Url : http://lists.alioth.debian.org/pipermail/pkg-cups-devel/attachments/20080403/dab2926a/attachment-0001.pdf