Summary: | net-print/cups-1.5.3: PDF printing no longer works with app-text/poppler-0.20.2 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Michał Górny <mgorny> |
Component: | [OLD] Printing | Assignee: | Printing Team <printing> |
Status: | RESOLVED FIXED | ||
Severity: | major | CC: | aklhfex, asturm, bugs+gentoo, dabbott, derk.tebokkel, gringofent, jarausch, jlec, kde, octoploid, reavertm, travisghansen |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://bugs.freedesktop.org/show_bug.cgi?id=52215 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 426706 | ||
Attachments: |
The complete error.log message
bugfix patch for poppler poppler-0.20.2-fix-pdf-print.patch |
Description
Michał Górny
2012-07-15 09:18:12 UTC
Ewww... ok what we'd need here is a backtrace from a coredump... Alternatively, could you please test if upgrading to cups-1.6rc1 plus dependencies helps? It's not keyworded yet, but I dont see big obstacles to keywording the rc if it helps... (In reply to comment #1) > Ewww... ok what we'd need here is a backtrace from a coredump... I have the coredump already. I just need to find out how to get it from journal :D. I managed to print a PDF from firefox-13 with cups-1.6_rc1 (poppler-0.20.2), the only error line that appeared inside the log: E [15/Jul/2012:16:46:45 +0200] [Client 13] Bad request line "%16%03%01" from localhost. (In reply to comment #1) > Ewww... ok what we'd need here is a backtrace from a coredump... > > Alternatively, could you please test if upgrading to cups-1.6rc1 plus > dependencies helps? It's not keyworded yet, but I dont see big obstacles to > keywording the rc if it helps... After talking on IRC Michal sent me the PDF (contains private info, so not attached here), I'll poke at it and poppler with a sharp stick... :) It seems, that I have the same issue: The cups web interface shows: Status: stopped "/usr/libexec/cups/filter/pdftops failed" dmesg says: [617461.511246] pdftops[31278] trap divide error ip:7f43bc0c6f96 sp:7ffff1a29d10 error:0 in libpoppler.so.26.0.0[7f43bbf66000+217000] [619512.246923] pdftops[18258] trap divide error ip:7f3c08d8df96 sp:7fff369956d0 error:0 in libpoppler.so.26.0.0[7f3c08c2d000+217000] If I try to install cups-1.6_rc1 by setting: net-print/cups ~* ** =net-print/cups-filters-9999 ** in /etc/portage/package.keywords bzr downloads more than 1 GB (?) from http://bzr.linuxfoundation.org/openprinting/cups-filters: >>> Emerging (3 of 4) net-print/cups-filters-9999 >>> Unpacking source... * bzr branch start --> * repository: http://bzr.linuxfoundation.org/openprinting/cups-filters => /net/leopard/usr/portage_leopard/distfiles/bzr-src/cups-filters ... 1019181kB 432kB/s - Fetching revisions:Inserting stream:Estimate 84448/85977 Branched 6978 revisions. * exporting ... Could somebody explain, why more than 1 GB is downloaded if I find only 51 MB in /usr/portage/distfiles/bzr-src/cups-filters? root@moose:/root(55)# du -s /usr/portage/distfiles/bzr-src/cups-filters/ 51508 /usr/portage/distfiles/bzr-src/cups-filters/ 'emerge cups-filters' fails then with: CXX pdftoopvp-OPVPSplashClip.o In file included from ./filter/pdftoopvp/opvp/opvp_common.h:50:0, from filter/pdftoopvp/oprs/OPVPWrapper.h:8, from filter/pdftoopvp/oprs/OPVPSplashPath.h:5, from filter/pdftoopvp/oprs/OPVPSplash.h:20, from filter/pdftoopvp/oprs/OPRS.cxx:29: ./filter/pdftoopvp/opvp/opvp.h:159:19: warning: ISO C++ forbids zero-size array ‘data’ [-pedantic] In file included from filter/pdftoopvp/oprs/OPRS.cxx:32:0: filter/pdftoopvp/oprs/OPVPSplashXPath.h: In constructor ‘OPVPSplashXPath::OPVPSplashXPath()’: filter/pdftoopvp/oprs/OPVPSplashXPath.h:30:21: error: no matching function for call to ‘SplashXPath::SplashXPath()’ filter/pdftoopvp/oprs/OPVPSplashXPath.h:30:21: note: candidates are: /usr/include/poppler/splash/SplashXPath.h:69:3: note: SplashXPath::SplashXPath(SplashXPath*) /usr/include/poppler/splash/SplashXPath.h:69:3: note: candidate expects 1 argument, 0 provided /usr/include/poppler/splash/SplashXPath.h:52:3: note: SplashXPath::SplashXPath(SplashPath*, SplashCoord*, SplashCoord, GBool) /usr/include/poppler/splash/SplashXPath.h:52:3: note: candidate expects 4 arguments, 0 provided /usr/include/poppler/splash/SplashXPath.h:45:7: note: SplashXPath::SplashXPath(const SplashXPath&) /usr/include/poppler/splash/SplashXPath.h:45:7: note: candidate expects 1 argument, 0 provided In file included from ./filter/pdftoopvp/opvp/opvp_common.h:50:0, from filter/pdftoopvp/oprs/OPVPWrapper.h:8, from filter/pdftoopvp/oprs/OPVPSplashPath.h:5, from filter/pdftoopvp/oprs/OPVPSplashClip.h:5, from filter/pdftoopvp/oprs/OPVPSplashClip.cxx:3: ./filter/pdftoopvp/opvp/opvp.h:159:19: warning: ISO C++ forbids zero-size array ‘data’ [-pedantic] make[1]: *** [pdftoopvp-OPRS.o] Error 1 ... make[1]: *** [pdftoopvp-OPVPSplash.o] Error 1 make[1]: Leaving directory `/var/tmp/portage/net-print/cups-filters-9999/work/cups-filters-9999' make: *** [all] Error 2 * ERROR: net-print/cups-filters-9999 failed (compile phase): * emake failed Michał, what happens when you run this file through /usr/bin/pdftops (from poppler) directly? libexec/cups/filters/pdftops is merely a wrapper (execv) over this so you may be lucky to get direct coredump (unless crash happens as a result of edge-case in pdftops cmdline parameters combination). Everyone, dont bother trying cups-1.6. cups-filters (required) does not build with poppler-0.20 yet. Cheers. Then maybe something changed in cups-1.6. In 1.5 times cups pdftops filter is execv wrapper over poppler pdftops binary, therefore ther're no "doesn't build against poppler" case at all. In 1.6 I don't know. Ah yes, this pdftoopvp filter case (compilation). Anyway regarding pdftops (bug) cups version is likely irrelevant imho. pdftops (poppler one, not cupsfilter) may just crash on certain PDF files... (In reply to comment #6) > Michał, what happens when you run this file through /usr/bin/pdftops (from > poppler) directly? libexec/cups/filters/pdftops is merely a wrapper (execv) > over this so you may be lucky to get direct coredump (unless crash happens > as a result of edge-case in pdftops cmdline parameters combination). Directly it works. I guess some argv may be relevant, or cups does something ugly to input. I've opened an upstream bug here: https://bugs.freedesktop.org/show_bug.cgi?id=52215 Created attachment 318516 [details, diff]
bugfix patch for poppler
This patch fixes the problem.
Thanks to thomasf.
*** Bug 426552 has been marked as a duplicate of this bug. *** Patch works for me. Please report also on the upstream bug (see URL). Poppler is a sensitive beast which I dont really want to modify without understanding. As soon as a fix is in git master, we'll backport it here. I applied the above patch (which matches the suggested one upstream), and now I'm printing like normal. THANKS! attached patch worked for me too. thanks ! *** Bug 427108 has been marked as a duplicate of this bug. *** Created attachment 319258 [details, diff]
poppler-0.20.2-fix-pdf-print.patch
Upstream has this fixed in git, here is a copy of their patch re-rolled for poppler-0.20.2
(In reply to comment #13) > Created attachment 318516 [details, diff] [details, diff] > bugfix patch for poppler > > This patch fixes the problem. > Thanks to thomasf. Can confirm the patch works for me as well,I was prompted to look into it further as firefox was printing the pages,but konqueror wasn't. Thanks for your efforts. Geoff Fixed in poppler-0.20.2-r1, sorry this took so long. |