Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Paul Gevers (2018-09-15 07:41:54) > On 14-09-18 22:26, Jonas Smedegaard wrote: > > Another release of Ghostscript is now in experimental. Can someone > > please test if those autopkgtests still fail? > > 9.25~dfsg-1~exp1 passed the cups test. > > https://ci.debian.net/data/autopkgtest/testing/amd64/c/cups/994233/log.gz Great! I'll make a release for unstable now. Thanks for all the help to everyone involved! - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Hi Jonas, On 14-09-18 22:26, Jonas Smedegaard wrote: > Another release of Ghostscript is now in experimental. Can someone > please test if those autopkgtests still fail? 9.25~dfsg-1~exp1 passed the cups test. https://ci.debian.net/data/autopkgtest/testing/amd64/c/cups/994233/log.gz Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Jonas Smedegaard (2018-08-31 01:25:24) > Quoting Paul Gevers (2018-08-29 20:24:49) > > Control: tags -1 moreinfo > > > > Hi, > > > > On 29-08-18 20:20, Jonas Smedegaard wrote: > > > Thanks - that is indeed helpful, but provides only the _cups_ commands. > > > > > > Inside those are some Ghostscript command (and some data) which I would > > > need to check if/what fails with Ghostscript. > > > > Both of them are "ELF 64-bit LSB shared object" so it would help if the > > cups maintainers could help here. > > Do the freshly released experimental Ghostscript release help anything? Another release of Ghostscript is now in experimental. Can someone please test if those autopkgtests still fail? - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Hi Jonas, On 31-08-18 16:19, Jonas Smedegaard wrote: > Currently¹ I cannot (easily) setup a CUPS testing environment, so would > appreciate if someone else can confirm if the version now in testing > _also_ causes this same failure - and if so then please help ensure that > this issue does not block the security fix now in unstable to enter > testing. The cups test with only packages from testing is successfully running. It has been tried two times over the last two days, visible with the trigger "migration-reference/0" here: https://ci.debian.net/packages/c/cups/testing/amd64/ I have requested the cups test in testing with ghostscript from experimental. I hasn't finished yet, albeit it probably started soon after I requested it at 2018-08-31 19:20:50 UTC, so this isn't looking good. Feel free to check in 2:30 hours or so, it should be visible at the page linked above with the trigger "elbrus ghostscript from experimental". If that ran for 2:47 and failed it probably timed out (which you can confirm by looking at the bottom of the test log). Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Hi Jonas, On 31-08-18 01:25, Jonas Smedegaard wrote: > Do the freshly released experimental Ghostscript release help anything? I'll trigger the test shortly and report back with the results. Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Dear all, On 31-08-18 15:48, Till Kamppeter wrote: > On 31/08/18 15:36, Didier 'OdyX' Raboud wrote: >> Le vendredi, 31 août 2018, 01.25:24 h CEST Jonas Smedegaard a écrit : >>> Do the freshly released experimental Ghostscript release help anything? >> >> It doesn't seem to, unfortunately. :-( >> >> To reproduce the issue; just run this as root: >> /usr/share/cups/test-drivers I tried this on my fully up-to-date testing/buster system and it fails for a different reason: lpinfo: Bad file descriptor Not sure yet, what that means for any of this. >> Surprisingly; it will fail when testing the _second_ printer, always. >> Also, it >> doesn't seem to get fixed with the ghostscript from testing. >> >> There's something fishy here, but I can't say with certainty that it's >> ghostscript's fault :-( > > Which driver is this second printer using? As mentioned in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=907493#26 DUMMY_PRINTER_NAME=test-printer0 driver=drv:///sample.drv/dymo.ppd > Which version of cups-filters are you using? 1.21.0 has a bug in > foomatic-rip which is fixed in 1.21.1. Please update to 1.21.1 if you > have 1.21.0 currently. I had none installed. I now try with 1.20.4-1 which is in testing, where the bug was reported from and where cups fails with the ghostscript from unstable. The autopkgtest ran successfully with cups-filters/1.21.0-1 and with cups-filters/1.21.1-1 so I assume they are both fine in this respect. On my laptop, lpinfo still fails. Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Jonas Smedegaard (2018-08-31 15:43:28) > Quoting Didier 'OdyX' Raboud (2018-08-31 15:36:09) > > Le vendredi, 31 août 2018, 01.25:24 h CEST Jonas Smedegaard a écrit : > > > Do the freshly released experimental Ghostscript release help anything? > > > > It doesn't seem to, unfortunately. :-( > > > > To reproduce the issue; just run this as root: > > /usr/share/cups/test-drivers > > > > Surprisingly; it will fail when testing the _second_ printer, always. > > Also, it doesn't seem to get fixed with the ghostscript from testing. > > > > There's something fishy here, but I can't say with certainty that it's > > ghostscript's fault :-( > > Uhm, if the ghostscript in _testing_ causes that test to fail, then this > bug should *not* block the ghostscript in unstable to enter testing!!! Let me try again - I see that my previous message could easily be perceived as aggressive: Not intended at all. Sorry! Thanks, Odyx, for checking against the various versions of Ghostscript. Currently¹ I cannot (easily) setup a CUPS testing environment, so would appreciate if someone else can confirm if the version now in testing _also_ causes this same failure - and if so then please help ensure that this issue does not block the security fix now in unstable to enter testing. - Jonas ¹ I am at MMMfest, a week long festival near Paris. -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
On 31/08/18 15:36, Didier 'OdyX' Raboud wrote: Le vendredi, 31 août 2018, 01.25:24 h CEST Jonas Smedegaard a écrit : Do the freshly released experimental Ghostscript release help anything? It doesn't seem to, unfortunately. :-( To reproduce the issue; just run this as root: /usr/share/cups/test-drivers Surprisingly; it will fail when testing the _second_ printer, always. Also, it doesn't seem to get fixed with the ghostscript from testing. There's something fishy here, but I can't say with certainty that it's ghostscript's fault :-( Which driver is this second printer using? Which version of cups-filters are you using? 1.21.0 has a bug in foomatic-rip which is fixed in 1.21.1. Please update to 1.21.1 if you have 1.21.0 currently. Till
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Le vendredi, 31 août 2018, 01.25:24 h CEST Jonas Smedegaard a écrit : > Do the freshly released experimental Ghostscript release help anything? It doesn't seem to, unfortunately. :-( To reproduce the issue; just run this as root: /usr/share/cups/test-drivers Surprisingly; it will fail when testing the _second_ printer, always. Also, it doesn't seem to get fixed with the ghostscript from testing. There's something fishy here, but I can't say with certainty that it's ghostscript's fault :-( (No more time to investigate this today, sorry) -- OdyX signature.asc Description: This is a digitally signed message part.
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Didier 'OdyX' Raboud (2018-08-31 15:36:09) > Le vendredi, 31 août 2018, 01.25:24 h CEST Jonas Smedegaard a écrit : > > Do the freshly released experimental Ghostscript release help anything? > > It doesn't seem to, unfortunately. :-( > > To reproduce the issue; just run this as root: > /usr/share/cups/test-drivers > > Surprisingly; it will fail when testing the _second_ printer, always. > Also, it doesn't seem to get fixed with the ghostscript from testing. > > There's something fishy here, but I can't say with certainty that it's > ghostscript's fault :-( Uhm, if the ghostscript in _testing_ causes that test to fail, then this bug should *not* block the ghostscript in unstable to enter testing!!! - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Paul Gevers (2018-08-29 20:24:49) > Control: tags -1 moreinfo > > Hi, > > On 29-08-18 20:20, Jonas Smedegaard wrote: > > Thanks - that is indeed helpful, but provides only the _cups_ commands. > > > > Inside those are some Ghostscript command (and some data) which I would > > need to check if/what fails with Ghostscript. > > Both of them are "ELF 64-bit LSB shared object" so it would help if the > cups maintainers could help here. Do the freshly released experimental Ghostscript release help anything? - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Control: tags -1 moreinfo Hi, On 29-08-18 20:20, Jonas Smedegaard wrote: > Thanks - that is indeed helpful, but provides only the _cups_ commands. > > Inside those are some Ghostscript command (and some data) which I would > need to check if/what fails with Ghostscript. Both of them are "ELF 64-bit LSB shared object" so it would help if the cups maintainers could help here. Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Quoting Paul Gevers (2018-08-29 19:58:37) > Control: tags -1 - moreinfo > > On Wed, 29 Aug 2018 09:41:37 +0200 Jonas Smedegaard wrote: > > It would be most helpful if someone could dig out from that convoluted > > ci-in-cups test the actual ghostscript command causing cups to hang. > > Looking here: > https://sources.debian.org/src/cups/2.2.8-5/debian/tests/cups/ > > it runs: > /usr/share/cups/test-drivers > > As the log ends with: > * Driver drv:///sample.drv/dymo.ppd > - Create test printer: done. > - Print test job with /usr/share/cups/data/topsecret.pdf: > > I guess it successfully runs this command > /usr/sbin/lpadmin -p $DUMMY_PRINTER_NAME -E -m $driver -v > file:///dev/null > and fails with this command: > rid=$(/usr/bin/lp -d $DUMMY_PRINTER_NAME $file | sed -e > 's/^.*request id is \(.*\) (.*)$/\1/g') > > where > DUMMY_PRINTER_NAME=test-printer0 > driver=drv:///sample.drv/dymo.ppd > file=/usr/share/cups/data/topsecret.pdf > > Is that enough for you to continue? Thanks - that is indeed helpful, but provides only the _cups_ commands. Inside those are some Ghostscript command (and some data) which I would need to check if/what fails with Ghostscript. - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Control: tags -1 - moreinfo On Wed, 29 Aug 2018 09:41:37 +0200 Jonas Smedegaard wrote: > It would be most helpful if someone could dig out from that convoluted > ci-in-cups test the actual ghostscript command causing cups to hang. Looking here: https://sources.debian.org/src/cups/2.2.8-5/debian/tests/cups/ it runs: /usr/share/cups/test-drivers As the log ends with: * Driver drv:///sample.drv/dymo.ppd - Create test printer: done. - Print test job with /usr/share/cups/data/topsecret.pdf: I guess it successfully runs this command /usr/sbin/lpadmin -p $DUMMY_PRINTER_NAME -E -m $driver -v file:///dev/null and fails with this command: rid=$(/usr/bin/lp -d $DUMMY_PRINTER_NAME $file | sed -e 's/^.*request id is \(.*\) (.*)$/\1/g') where DUMMY_PRINTER_NAME=test-printer0 driver=drv:///sample.drv/dymo.ppd file=/usr/share/cups/data/topsecret.pdf Is that enough for you to continue? Paul signature.asc Description: OpenPGP digital signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Control: tags -1 + moreinfo Quoting Graham Inggs (2018-08-29 00:56:49) > Control: severity -1 serious > Control: found -1 ghostscript/9.22~dfsg-3 > > Hi Jonas > > I'm bumping the severity of this bug to prevent ghostscript from > migrating until the cups autopkgtest regression has been investigated. Thanks for reporting, Paul, and for blocking, Graham. It would be most helpful if someone could dig out from that convoluted ci-in-cups test the actual ghostscript command causing cups to hang. - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Control: severity -1 serious Control: found -1 ghostscript/9.22~dfsg-3 Hi Jonas I'm bumping the severity of this bug to prevent ghostscript from migrating until the cups autopkgtest regression has been investigated. Regards Graham
Bug#907493: ghostscript breaks cups autopkgtest: test times out
Source: ghostscript, cups Version: ghostscript/9.22~dfsg-3 Version: cups/2.2.8-5 X-Debbugs-CC: debian...@lists.debian.org User: debian...@lists.debian.org Usertags: breaks needs-update timeout Dear maintainers, With a recent upload of ghostscript the autopkgtest of cups started to fail in testing and unstable due to the test timing out. It now takes more than 2:47 hours to complete the test, while in the past it ran typically slightly more than 6 minutes. As ghostscript was uploaded with urgency high this regression is NOT delaying of the migration of ghostscript to testing [1] and cups will be hit by this in testing tomorrow. I have assigned the bug to both packages. Could you please investigate the situation and reassign the bug to the right package? If needed, please change the bug's severity as appropriate. More information about this bug and the reason for filing it can be found on https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation Paul [1] https://qa.debian.org/excuses.php?package=ghostscript https://ci.debian.net/data/autopkgtest/testing/amd64/c/cups/894494/log.gz signature.asc Description: OpenPGP digital signature
