Réf. : Re: Réf. : Re: [cups.general]

[Antoine PICOLET]

Antoine PICOLET wrote:
> 
> Here are the traces (in debug level) in my error log.
> They show that the file (an OS/2 BMP image) passes through the imagetops
> filter and then to the socket backend (note the problem occured also 
with
> the lpd backend); then, the imagetops filter crashes because it cannot
> deal with the file format, but the backend exit value is 0; and, 
finally,
> the CUPS scheduler performs as if the backend did crash and the filter
> not, so it disables the printer but keeps the corrupted job...
> Note that I use syslog for logging, so the lines may be in 
non-chronologic
> order :
> 
> ProcessIPPRequest(0x40303008[6]): operation_id = 0002
> print_job(0x40303008[6], ipp://localhost:631/printers/printer1)
> print_job: auto-typing file...
> print_job: request file type is image/x-bitmap.
> check_quotas(0x40303008[6], 0x81a4fe8[printer1])
> check_quotas: requesting-user-name = 'root'
> print_job: requesting-user-name = 'root'
> Adding start banner page "none" to job 113.
> copy_banner(0x40303008[6], 0x8139b18[113], none)
> add_file(con=0x40303008[6], job=113, filetype=image/x-bitmap,
> compression=0)
> Adding end banner page "none" to job 113.
> copy_banner(0x40303008[6], 0x8139b18[113], none)
> Job 113 queued on 'printer1' by 'root'.
> Job 113 hold_until = 0
> SaveJob: Closing file 8...
> StartJob(113, 0x81a4fe8)
> StartJob() id = 113, file = 0/1
> job-sheets=none,none
> banner_page = 0
> StartJob: argv =
> 
"printer1","113","root","test-bmp-os2.bmp","1","","/var/spool/cups/d00113-001"
> StartJob: envp[0]="PATH=/usr/lib/cups/filter:/bin:/usr/bin"
> StartJob: envp[1]="SOFTWARE=CUPS/1.1"
> StartJob: envp[2]="USER=root"
> StartJob: envp[3]="CHARSET=iso-8859-1"
> StartJob: envp[4]="LANG=en"
> StartJob: envp[5]="TZ=UTC"
> StartJob: envp[6]="PPD=/etc/cups/ppd/printer1.ppd"
> StartJob: envp[7]="CUPS_SERVERROOT=/etc/cups"
> StartJob: envp[8]="RIP_MAX_CACHE=8m"
> StartJob: envp[9]="TMPDIR=/var/spool/cups/tmp"
> StartJob: envp[10]="CONTENT_TYPE=image/x-bitmap"
> StartJob: envp[11]="DEVICE_URI=socket://i915.prn.msy.sagem:9100"
> StartJob: envp[12]="PRINTER=printer1"
> StartJob: envp[13]="CUPS_DATADIR=/usr/share/cups"
> StartJob: envp[14]="CUPS_FONTPATH=/usr/share/cups/fonts"
> StartJob: Allocating status buffer...
> StartJob: statusfds = [ 8 9 ]
> StartJob: filterfds[1] = [ 10 -1 ]
> StartJob: filter = "/usr/lib/cups/filter/imagetops"
> StartJob: filterfds[0] = [ 11 12 ]
> start_process("/usr/lib/cups/filter/imagetops", 0xbffef530, 0xbffee8a0,
> 10, 12, 9)
> PID 17886 stopped with status 1!
> Started filter /usr/lib/cups/filter/imagetops (PID -17886) for job 113.
> StartJob: backend = "/usr/lib/cups/backend/socket"
> StartJob: filterfds[1] = [ -1 10 ]
> start_process("/usr/lib/cups/backend/socket", 0xbffef530, 0xbffee8a0, 
11,
> 10, 9)
> Started backend /usr/lib/cups/backend/socket (PID 17887) for job 113.
> StartJob: Adding fd 8 to InputSet...
> add_job_state_reasons(0x40303008[6], 113)
> ProcessIPPRequest: 6 status_code=0
> ProcessIPPRequest: Adding fd 6 to OutputSet...
> [Job 113] printer1 113 root test-bmp-os2.bmp 1 
/var/spool/cups/d00113-001
> [Job 113] ImageOpen("/var/spool/cups/d00113-001", 1, 1, 100, 0, (nil))
> [Job 113] offset = 26
> [Job 113] Bad BMP dimensions 102500389x1572865x8251
> [Job 113] Unable to open image file for printing!
> PID 17887 exited with no errors.
> [Job 113] Attempting to connect to host i915.prn.msy.sagem on port 9100
> [Job 113] Connected to host, sending print job...
> [Job 113] Print file sent, waiting for printer to finish...
> [Job 113] Ready to print.
> UpdateJob: job 113, file 0 is complete.
> UpdateJob: Removing fd 8 from InputSet...
> StopJob: id = 113, force = 0
> Saving printers.conf...
> StopJob: printer state is 5
> StopJob: Freeing status buffer...
> SaveJob: Closing file 8...
> WriteClient: Removing fd 6 from OutputSet...
> ReadClient() 6, used=0
> CloseClient() 6
> CloseClient: Removing fd 6 from InputSet and OutputSet...
> 
> Helge Blischke <h.blischke at srz.de>
> 
> Envoyé par : cups-bounces at easysw.com
> 18/03/2005 13:07
> Veuillez répondre à "Mirror of cups.general Newsgroup"
> Remis le : 18/03/2005 13:10
> 
> 
>         Pour :  cups at easysw.com
>         cc :    (ccc : Antoine PICOLET/DRD/SAGEM)
>         Objet : Re: [cups.general] Printer queue gets closed when print 
file is
> 
> Antoine PICOLET wrote:
> >
> > Hello,
> >
> > I am currently using CUPS 1.1.19, on a Linux SuSE Linux Entreprise
> Server
> > 8.0. I do use a lot of different printers, generally using lpd and
> socket
> > backends.
> >
> > I supppose that, when I send to print a file which will make a filter
> fail
> > (typically, an OS/2 BMP - by the way, is this normal ?, or some PDF
> > files), the job is removed by the scheduler, and gets an 'aborted'
> status,
> > and the printer queue itself is left unchanged. However, in my case, 
the
> > printer queue gets closed, and the job remains; as soon as I reopen 
the
> > queue (by 'enable'), the job tries to print again and the queue
> > immediately gets closed again. I have to remove the job before I can
> > reopen the queue.
> > Note that this does not happen with files in an unsupported format
> > (application/msword...) since they are immediately refused by the
> server.
> >
> > I know I am using an old version of CUPS. I tried version 1.1.23, and
> the
> > problem does not seem to occur anymore, however it did not occur each
> time
> > with 1.1.19, and,I cannot find it in the list of bugs corrected 
between
> > 1.1.19 and 1.1.23. Has this behavior been specifically corrected ? 
Does
> > anyone have seen this kind of problem (in any version of CUPS), or 
know
> > what kind of misconfiguration might be causing it ?
> >
> > Thank you in advance .
> >
> > Antoine
> 
> You should look into the error_log and see which filter(s)/backend(s)
> are
> said to have exited with a status > 0.
> 
> Helge
> 
>

> Well, if you carefully look at your log, you'll find that the imagetops
> filter
> has been started and the PId 17886 assigned.
> And the log tells you that the PID 17886 (i.e. just this filter) stopped
> with status 1. Thus, the scheduler puts the queue into the stop state.

> Helge


Well, this is exactly what I am saying... the logs say that the backend is 
OK, 
and the filter is not because the print data is corrupted. In this case, I 

believe CUPS is not supposed to put the queue in stop state - which it 
would 
have had to do if the backend had had an error - but rather abort the job
itself, giving it an "aborted" status ! 
Doing it the way it does (ie, stopping the print queue when a corrupted 
job arrives) means that anybody can close a print queue by sending some 
weird 
data, and that it then takes an administrator to cancel the job and reopen
the print queue.




-- 
Helge Blischke
Softwareentwicklung
SRZ Berlin | Firmengruppe besscom
http://www.srz.de
_______________________________________________
cups mailing list
cups at easysw.com
http://lists.easysw.com/mailman/listinfo/cups