Bug#900485: [Debian-med-packaging] Bug#900485: python-pbcore: FTBFS and Debci failure with NumPy 1.14
Hi, Andreas On June 25, 2018 3:32:00 AM EDT, Andreas Tille wrote: >Hi Graham, > >On Sun, Jun 24, 2018 at 08:48:32PM +0200, Graham Inggs wrote: >> Control: tags -1 + patch > >thanks for the patch. Since we are targeting in Git at version 1.4.2 I >adapted the patch to this version and it works for this. I also fixed >another issue in the test suite[1]. Unfortunately there are remaining >errors in the test suite which are all connected to a possible issue >with >XML files: > >$ grep "ERROR: Invalid file produced" >python-pbcore_1.4.2+dfsg-1_amd64.build >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpiOHCZI.alignmentset.xml.disabled >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmplAqR2T.alignmentset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpMyWNXh.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmps4GD0E.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmp71IS22dataset-unittest/test.alignmentset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpUn37tBdataset unittest/spaced.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpMouabqdataset-unittest/tempfile.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpN0pUz2alignmentset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpeQ4v3jalignmentset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmp0Kb6RWdataset-doctest/tempfile.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpj20209.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmp2XGDAY.barcodeset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpXtaXG3.consensusreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpJlLjAA.contigset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpDKGXFh.contigset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmptZ4Kro.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmp6ySKmX.contigset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpvYTcX1.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpiXc1ez.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpdiSRmI.subreadset.xml >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >/tmp/tmpveTAsa.subreadset.xml > > >Before I start diving into this: Afif, upstream seems to have released >1.5.0 which might be worth a try. While I asked for creating according >release tags[2] I remember you said that upstream does not care much >about this metadata. I wonder whether we should possibly use mode=git >inside the watch file, inject version 1.5.0 into packaging Git, cross >fingers that the issue might be fixed and if not report the issue >upstream. > I believe that error had to do with the package requiring an older version of some dependency. At least, the error I was seeing at the time was for that reason, which prevented me from uploading. regards Afif . > >[1] >https://salsa.debian.org/med-team/python-pbcore/blob/master/debian/patches/relax_test_constraint.patch >[2] https://github.com/PacificBiosciences/pbcore/issues/116 > >-- >http://fam-tille.de > >___ >Debian-med-packaging mailing list >debian-med-packag...@alioth-lists.debian.net >https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-med-packaging
Bug#900485: [Debian-med-packaging] Bug#900485: python-pbcore: FTBFS and Debci failure with NumPy 1.14
Hi Afif, On Mon, Jun 25, 2018 at 03:41:11AM -0400, Afif Elghraoui wrote: > >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: > >/tmp/tmpveTAsa.subreadset.xml > > > >Before I start diving into this: Afif, upstream seems to have released > >1.5.0 which might be worth a try. While I asked for creating according > >release tags[2] I remember you said that upstream does not care much > >about this metadata. I wonder whether we should possibly use mode=git > >inside the watch file, inject version 1.5.0 into packaging Git, cross > >fingers that the issue might be fixed and if not report the issue > >upstream. > > > > I believe that error had to do with the package requiring an older version of > some dependency. At least, the error I was seeing at the time was for that > reason, which prevented me from uploading. So we can either skip those tests (no idea how important is this for general functionality) or give 1.5.0 a try. What do you think is the better strategy? (And what do you think about watch mode=git?) Kind regards Andreas. -- http://fam-tille.de
Bug#900485: [Debian-med-packaging] Bug#900485: python-pbcore: FTBFS and Debci failure with NumPy 1.14
On June 25, 2018 5:07:48 AM EDT, Andreas Tille wrote: >Hi Afif, > >On Mon, Jun 25, 2018 at 03:41:11AM -0400, Afif Elghraoui wrote: >> >pbcore.io.dataset.DataSetIO: ERROR: Invalid file produced: >> >/tmp/tmpveTAsa.subreadset.xml >> > >> >Before I start diving into this: Afif, upstream seems to have >released >> >1.5.0 which might be worth a try. While I asked for creating >according >> >release tags[2] I remember you said that upstream does not care much >> >about this metadata. I wonder whether we should possibly use >mode=git >> >inside the watch file, inject version 1.5.0 into packaging Git, >cross >> >fingers that the issue might be fixed and if not report the issue >> >upstream. >> > >> >> I believe that error had to do with the package requiring an older >version of some dependency. At least, the error I was seeing at the >time was for that reason, which prevented me from uploading. > >So we can either skip those tests (no idea how important is this for >general functionality) or give 1.5.0 a try. What do you think is >the better strategy? I think we can give 1.5.0 a try, but it might have the same problem. > (And what do you think about watch mode=git?) > I don't know much about it, but I don't have any strong preference here. regards Afif
Bug#900485: [Debian-med-packaging] Bug#900485: python-pbcore: FTBFS and Debci failure with NumPy 1.14
On Mon, Jun 25, 2018 at 09:52:50AM -0400, Afif Elghraoui wrote: > I think we can give 1.5.0 a try, but it might have the same problem. Since it builds successfully I've just uploaded. > > (And what do you think about watch mode=git?) > > I don't know much about it, but I don't have any strong preference here. It might serve as a sensible solution for all upstreams using Git without proper release tagging. I guess the pbcore example will explain what to do. BTW, I remembered to late that you are not happy about the cme formatting. I just fired up cme for Vcs-fields, Standards-Version etc. I hope you don't mind - feel free to revert to your prefered formatting and I hope I will not forget if I touch some of your packages in future. Kind regards Andreas. PS: I'm currently checking pbseqlib since now upstream has tagged a release but there are some hdf5 issues which prevent me from uploading. Please let me know if you want me to revert the cme formatting to its previous shape. -- http://fam-tille.de
Bug#900485: [Debian-med-packaging] Bug#900485: python-pbcore: FTBFS and Debci failure with NumPy 1.14
On June 26, 2018 3:45:49 AM EDT, Andreas Tille wrote: >On Mon, Jun 25, 2018 at 09:52:50AM -0400, Afif Elghraoui wrote: >> I think we can give 1.5.0 a try, but it might have the same problem. > >Since it builds successfully I've just uploaded. > Ok >> > (And what do you think about watch mode=git?) >> >> I don't know much about it, but I don't have any strong preference >here. > >It might serve as a sensible solution for all upstreams using Git >without proper release tagging. I guess the pbcore example will >explain >what to do. > >BTW, I remembered to late that you are not happy about the cme >formatting. I just fired up cme for Vcs-fields, Standards-Version etc. >I hope you don't mind - feel free to revert to your prefered formatting >and I hope I will not forget if I touch some of your packages in >future. > Don't worry about this. > >PS: I'm currently checking pbseqlib since now upstream has tagged a >release but there are some hdf5 issues which prevent me from >uploading. Please let me know if you want me to revert the cme >formatting to its previous shape. > No need to revert anything. regards Afif