[Foomatic] Bugus print with Lexmark E210/300dpi

Till Kamppeter till.kamppeter at gmx.net
Mon Apr 5 18:14:12 PDT 2004


Marcelo Ricardo Leitner wrote:
> On Mon Mar 01, 2004 at 13:25:00 -0500, Grant Taylor wrote:
> 
>>Marcelo Ricardo Leitner <mrl at conectiva.com.br> writes:
>>
>>
>>>I'm using this printer with cups+foomatic and it is configurated directly
>>>through foomatic-configure. What happens is that I can't print with 300dpi
>>>resolution, but I can do it via Windows. The other resolutions seems to be
>>>ok.
>>>
>>>When printing at 300dpi I got a "quaked" print. I have one test page scanned
>>>over here http://bugzilla.conectiva.com.br/attachment.cgi?id=882&action=view
>>>and the page that should have been printed
>>>http://bugzilla.conectiva.com.br/attachment.cgi?id=906&action=view
>>
>>>Any ideas of what could be happening?
>>
>>Hmm.  This *might* be a bug caused by my compression bug workaround.
>>
>>The printer (in "smartgdi" mode, which is the only documented mode for
>>Samsung non-PCL printers) accepts a series of bands of compressed
>>bitmap data.  Each band must be 64KB or smaller.
>>
>>For bands which are larger than 64KB, there is a problem, as we have
>>only the one known compression method.  So my workaround "fuzzes" the
>>original print data and recompresses.  Basically it subtracts
>>horizontal resolution iteratively until the band "fits" when
>>compressed.  This is visible (mainly in dithered areas) as half-inch
>>sized horizontal distortions on prints.
>>
>>That said, your image doesn't really look like the distortion my hack
>>introduces.  And many if not all parts of that test page should print
>>without triggering my hack; usually only data like full-page-width f-s
>>dithered areas don't compress adequately.  Ghostscript's internal
>>halftoning is coarser than a straight 300dpi f-s dither, and rarely
>>causes this issue.
>>
>>I would suspect some other sort of problem, perhaps a resolution
>>problem of some sort; I have never used the driver much at 300dpi.
>>
>>Which exact driver version are you using?  The only way to really tell
>>is probably to obtain the matching sources for your install
>>ghostscript and inspect the gsdevgdi.c source directly.  Unless a
>>Connectiva person can volunteer what they ship...
>>
>>We should also verify that the driver is the problem and not foomatic
>>(which has an increasing number of scarey driver output postprocessing
>>features).  Can you print a 300dpi page successfully by invoking
>>ghostscript directly from the command line into your printer's /dev
>>node?
> 
> ---end quoted text---
> 
> Sorry for the time this has been left alone, I was waiting the result test.
> 
> As described at www.linuxprinting.org, the user could make a successful
> direct printing at 300dpi by using the foomatic-rip command. The report, in
> portuguese, can be found at http://bugzilla.conectiva.com.br/show_bug.cgi?id=9353.
> 
> The user was using cups at 1.1.20, foomatic at 3.0.1 and espgs 7.07.1.
> There are some other patches applied to the packages. They should not be
> related to this but if you want I can give you the sources.
> 
> Thanks for helping,

As Grant already mentioned, there is a problem with the compression of 
the bitmap data. Probably the original CUPS test page, the Conectiva 
CUPS test page, and the text file with the driver documentation give 
different sizes for the compressed bands and so the bug does not show up 
always. Grant here is a scanned image of a bad 300-dpi-printout of the 
Conectiva test page:

http://bugzilla.conectiva.com.br/attachment.cgi?id=882&action=view

and here is the original PostScript file

http://bugzilla.conectiva.com.br/attachment.cgi?id=906&action=view

Can you have a look at this?

    Till




More information about the Printing-foomatic mailing list