Using a foomatic-db driver (Epson-EPL-6100L-epl6100l.ppd) and an external tool (epsoneplijs-0.4.0), printing works perfectly with both cups-1.1.20-r3 and cups-1.1.20-r2, as long as you use ghostscript-7.07.1-r1 (or prior). ghostscript-7.07.1-r7 fails. It appears, that the command line gs -q -dBATCH -dSAFER -dNOPAUSE -sProcessColorModel=DeviceGray -dBitsPerSample=1 -sDEVICE=ijs -sIjsServer=/usr/local/bin/ijs_server_epsonepl -dIjsUseOutputFD -sDeviceManufacturer=Epson -sDeviceModel=EPL6100L -sIjsParams="" -sOutputFile=xxx.raw xxx.ps (or alike) makes ghostscript dump core. gdb says this happened due to signal 11 (segmentation fault), and: (gdb) backtrace #0 0x081f0d69 in cups_sync_output () Which is not exactly rich output. Please give instructions if further debugging needed.
Just chiming-in to confirm this. With ghostscript-7.07.1-r7 and cups-1.1.20-r3, all my print jobs fail with state "aborted", downgrading to ghostscript-7.07.1-r1 restores normal functionality.
On my system, its because the ebuild has the search paths set wrong for ghostscript. If there is a need to run -r7, then you can apply the hack on bug #67913
Ghostscript throws core; it doesn't display a message that it couldn't find something. Even with the symbolic link created, it still throws core.
can you please test every gs version from 7.07.1-r1 to 7.07.1-r7 so we can exactly see which version introduced the problem
The problems occurs first with -r2.
Just to note that this is a very serious bug, especially because it doesn't produce any sane output as to where the problem resides. And since I had problems already with a shaky parport config, I presumed that it had to do with the hardware connection. >7.07.1-r1 should be hard masked. It has nothing to do with cups. I use a special ppd for a Kyocera Mita FS 1010 Postscript printer and gstogs dies silently here as well.
I have changed the ebuild so it produces an unstripped executable (there might have been an easier way I didn't know about). The fault happens in gsijs_fill_rectangle() which is forked at that time. Apparently, this function was added with the patch gs7.07.1-krgb.patch introduced with -r2. Hope this might help.
Not sure if this is related (haven't tracked down any core image since its 7am and I don't wanna be hacking right now), but ghostscript 7.07.1-r7 screws my printing up big time as well.. Trying 7.07.1-r1.
Downgrading made it work again. I think something is up with ghostscript7.07.1-r7. I have an HP deskjet 970Cxi, if it helps anything.
On my system (a PPC G3) ghostscript v7.07.1-r1 works well. I've tried also ghostscript v7.07.1-r7 without krgb patch, but it doesn't seem to work.
Why isnt this masked? Until this bug is corrected users updating gentoo will find themselves without cups and waste hours, as I did trying to reinstall this and that reconfigure cups, check everything over, search google and the forums and , maybe , hit on a thread that tells them about this bug. Then on reading the bug I find this bug and the short-term remidy has been know for over six weeks! Well before I did the upgrade. I need never had been affected by this issue. Isnt this just the sort of thing portage is supposed to handle? Please get this masked a.s.a.p, to save other duplicating this time-wasting. Thanks.
I masked ppc a known to work version just in case the printing team takes action masking the package.
-r7 has important changes, i can't just mask it. please try to find out which patches cause these problems
What more can I do concerning debugging? Can somebody please assist?
you can remove all patches from the ebuild, and then activate one after the other to see which one causes the problems
@Heinrich Wendel: Excuse me: Have you read my recent comments?
sure, you said the the krgb patch causes the problems, but iago couldn't confirm that
Sigh. Of course I can split the patch files and look for those parts which have caused the problem. In this case, I need to understand 1) what has been changed and 2) why it has been changed. This is not trivial, because the documentation from Gentoo side is practically non-existant. Sigh again. Maybe I will spend my time on debugging ghostscript. Can somebody please tell me how to debug a multi-threaded application? And what does this gsijs_fill_rectangle() do? Does anybody know?
Hello. I've been having the segfault problem too. Removing the krgb patch didn't solve it, but removing gs7.07.1-fontconfig-rh.patch did. I did as Heinrich asked and disabled all patches, then re-enabled them in turn. Note that a few patches were needed just for a successful build. Anyway, it works for me with all the patches included except gs7.07.1-fontconfig-rh.patch, so that one seems to be the problem. [I'm using an Epson Stylus 740 printer - it seems to be using the stcolor driver; the gimp-print drivers don't seem to be available to me, but that's another story (probably something stupid). Printing from KDE 3.3.1 seems to be ok.]
*** Bug 72603 has been marked as a duplicate of this bug. ***
removed both patches in -r8
*** Bug 74278 has been marked as a duplicate of this bug. ***
*** Bug 72821 has been marked as a duplicate of this bug. ***
