PhotoSmart C8180 hangs when printing duplex

One strange thing is also that with this 1-page job it says "processing page 3" in the Properties dialog of system-config-printer, after the printer has pulled back the paper.

One small correction: When the printer ejects the paper its display switches to "Print canceled" already. The job still stays in "Processing" state.

Here is the error_log. hpcups closes without errors. What hangs is the "hp" CUPS backend.

till@till-laptop:~/printing/cpd/common-printing-dialog/gtk-dialog$ ps auxwww | grep hp
[...]
lp 3869 0.9 0.0 42752 1952 ? Sl 01:34 0:02 hp:/usb/Photosmart_C8100_series?serial=MY7B5F70P704YH 1962 till Test Page 1 job-uuid=urn:uuid:4ce68171-3a17-3117-47b1-350f5d900cf4
[...]

When the printer ejects the page after some minutes (around 5 minutes) the ḧp"backend keeps running (infinite loop?). error_log of this point attached.

In the PPD file cupsEvenDuplex is set True. This means CUPS rasterizer is should send a even number of pages when duplex is set. This means 1-page print job should send a extra blank page. This feature worked on my Fedora 9 box using hplip 3.9.6b.

The additional blank page is correctly created, by the pdftopdf filter.

I have done

cupsfilter -m application/vnd.cups-pdf -p /etc/cups/ppd/Photosmart-C8100-series.ppd -o PageSize=A4Duplex -o Duplex=DuplexNoTumble /usr/share/system-config-printer/testpage-a4.ps > out.pdf

to see whether the extra blank page is created. It really gets created.

Then I have run

cupsfilter -m application/vnd.cups-raster -p /etc/cups/ppd/Photosmart-C8100-series.ppd -o PageSize=A4Duplex -o Duplex=DuplexNoTumble /usr/share/system-config-printer/testpage-a4.ps > out.raster

to see whether the CUPS Raster which is fed into hpcups is OK. I checked it with rasterview (http://www.easysw.com/~mike/rasterview/index.html) and it is perfect.

I have captured the driver output via a queue clone printing into a file:

cupsctl FileDevice=yes
lpadmin -p test -E -v file:/tmp/printout -P /etc/cups/ppd/Photosmart-C8100-series.ppd
lpr -P test -o PageSize=A4Duplex -o Duplex=DuplexNoTumble /usr/share/system-config-printer/testpage-a4.ps

The resulting file looks like that it is complete, but I could not see whether hpcups perhaps skipped the second page because it is blank. It should not skip it. This would be a bug.

Then I have sent the result to the printer not using HPLIP:

nc -w1 192.168.2.101 9100 < /tmp/printout

Same behavior as when printing through the regular queue, only difference that "nc" exits after the data transfer. It waits only 1 second for answers from the printer and the printer did not answer in this time frame. the "hp" CUPS backend perhaps waits without timeout.

The printer hangs with message "Now printing..." on its screen. After some minutes the page is ejected and the display shows "Print canceled...". Now the printer needs to be power-cycled.

Note that my original queue with which I discovered the bug is USB-based:

till@till-laptop:~/printing/cpd/common-printing-dialog/gtk-dialog$ lpstat -v Photosmart-C8100-series
device for Photosmart-C8100-series: hp:/usb/Photosmart_C8100_series?serial=MY7B5F70P704YH
till@till-laptop:~/printing/cpd/common-printing-dialog/gtk-dialog$

The connection type does not seem to matter as sending the job with the "nc" command shows the same problem.

Another remark: Documents with an even number of pages print perfectly.

If found out something more:

It seems that hpcups (or the code which hpcups and hpijs have in common) also adds a blank page to the end of the job. Icame to this conclusion because:

1. When I print a test page (1-page job) with system-config-printer the "Properties" dialog shows the printer state during the job execution and it says "Printing page 3" in the end. It should print only two pages, the test page itself znd 1 extra blank page.

2. When I edit the PPD file to have "*cupsEvenDuplex: False" and restart CUPS everything works fine.

Possibilities to fix this bug:

1. Let the driver not add a blank page as "*cupsEvenDuplex: True" adds a page already. Be careful that after fixing this both hpcups and hpijs work correctly.

2. Set "*cupsEvenDuplex: False" in the PPD as the driver already adds the needed blank page. This method cannot break hpijs. But you should still also test hpijs to have also the correct choice for "*cupsEvenDuplex:" in the hpijs PPDs.

Shoul be easy to fix in 3.9.8.

The PS C8100 series has a firmware issue with blank pages. The printer will hang. This was fixed in the 3.9.8 public release by removeing cupsEvenDuplex.

OK, works for me now with the released 3.9.8. Thank you.

I've been testing with HPLIP 3.9.12 and have found it necessary to modify cupsEvenDuplex to False in order to correctly print single page documents to duplex-enabled printers including an HP P3005x and an HP CM1312NFI. I've done this by modifying the hpcups.drv file as it seems to be common to printers which use the hpcups ljcolor or ljmono driver.