Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Andreas, It works for me! Could you share your failed build log, please? I built PyQi from your latest SVN and then built the latest biom-format against it. What's more, I had the build log from building against PyQi 0.2.0 and I diffed it with the new build log. They were essentially identical. The only things I changed in the PyQi build - 1) I used the pristine tarball rather that repacking it to a -dfsg. 2) I disabled dh-linktree as I don't have it on Ubuntu 12.04. Could it be due to my version of Python? I'm on 2.7.3. I'm due to start building stuff for Ubuntu 14.04 this week, so at that point I'll be on a system much closer to what you have with Debian Unstable. Right now I'm at the point of maximum divergence. Cheers, TIM On Sat, 2014-03-29 at 20:20 +, Andreas Tille wrote: Hi Tim, On Thu, Mar 27, 2014 at 03:46:01PM +0100, Andreas Tille wrote: My guess is that updating PyQi might help (https://pypi.python.org/pypi/pyqi/). Did you already do that? No, I did not. I just realised that this also has moved. To bad if projects are moving and our watch files are pointing to the wrong location. Since it might perfectly be the cause of the problem I could try to do this today evening / night if nobody will beat me. So pyqi is updated to the latest version in unstable and while biom was *running* with this latest version installed it does not *build* any more with this new version as (Build-)Depends. Any idea? Kind regards Andreas. -- Tim Booth tbo...@ceh.ac.uk NERC Environmental Bioinformatics Centre Centre for Ecology and Hydrology Maclean Bldg, Benson Lane Crowmarsh Gifford Wallingford, England OX10 8BB http://nebc.nerc.ac.uk +44 1491 69 2705 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Tim, On Mon, Mar 31, 2014 at 03:14:58PM +0100, Tim Booth wrote: It works for me! Could you share your failed build log, please? The cruxial part is $ pyqi make-bash-completion --command-config-module biom.interfaces.optparse.config --driver-name biom -o debian/bash-completion Traceback (most recent call last): File /usr/bin/pyqi, line 177, in module optparse_main(cmd_obj, argv[1:]) File /usr/lib/python2.7/dist-packages/pyqi/core/interfaces/optparse/__init__.py, line 276, in optparse_main result = optparse_cmd(local_argv[1:]) File /usr/lib/python2.7/dist-packages/pyqi/core/interface.py, line 40, in __call__ cmd_result = self.CmdInstance(**cmd_input) File /usr/lib/python2.7/dist-packages/pyqi/core/command.py, line 137, in __call__ result = self.run(**kwargs) File /usr/lib/python2.7/dist-packages/pyqi/commands/make_bash_completion.py, line 79, in run cfg_mod = _get_cfg_module(cfg_mod_path) File /usr/lib/python2.7/dist-packages/pyqi/commands/make_bash_completion.py, line 23, in _get_cfg_module mod = importlib.import_module(desc) File /usr/lib/python2.7/importlib/__init__.py, line 37, in import_module __import__(name) ImportError: No module named biom.interfaces.optparse.config which works with pyqi 0.2.0 but not with pyqi 0.3.1. I have $ python --version Python 2.7.6 I built PyQi from your latest SVN and then built the latest biom-format against it. What's more, I had the build log from building against PyQi 0.2.0 and I diffed it with the new build log. They were essentially identical. The only things I changed in the PyQi build - 1) I used the pristine tarball rather that repacking it to a -dfsg. This should not be the problem - I just excluded binary chunks from sphinx doc builds. 2) I disabled dh-linktree as I don't have it on Ubuntu 12.04. If this might be a problem for your purpose I'm fine with a patch which does not rely in dh-linktree. I think you have worked around the JS links issue in dh_installdocs of python-biom-format. I think the most elegant way to cope with this for the moment would be to comment --with linktree in d/rules and use the d/linktrees as data to set the links manually in override_dh_installdocs (similarly as you did in python-biom-format). If this plan might sound to vague I'm fine with implementing this to make sure we are working on the same code in Debian and BioLinux and can easily switch to the more elegant linktrees later. Could it be due to my version of Python? I'm on 2.7.3. This might be possible (see above). I'm due to start building stuff for Ubuntu 14.04 this week, so at that point I'll be on a system much closer to what you have with Debian Unstable. Right now I'm at the point of maximum divergence. Can you perhaps test the pyqi call to create a bash-completion file on 14.04. Than we can be quite sure that the Python version has probably some influence. Otherwise I'd go on asking on debian-python@l.d.o. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Tim, On Mon, Mar 31, 2014 at 06:48:03PM +0100, Tim Booth wrote: ... Normally, 'export PYTHONPATH=.' is redundant but here it seems to be effective. I can confirm this. So I'll spend some cycles tomorrow to fix the remaining lintian issues. Thanks for your hint. No, just stick with what you have and use dh-linktree. I'm so close to switching to the new 14.04 Ubuntu release that I might not even push this update for my 12.04 users unless someone specifically asks for it. I do appreciate the general importance of keeping our code in sync but in this case there's no point. Many thanks for the clarification and thanks for checking the package Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Tim, On Thu, Mar 27, 2014 at 03:46:01PM +0100, Andreas Tille wrote: My guess is that updating PyQi might help (https://pypi.python.org/pypi/pyqi/). Did you already do that? No, I did not. I just realised that this also has moved. To bad if projects are moving and our watch files are pointing to the wrong location. Since it might perfectly be the cause of the problem I could try to do this today evening / night if nobody will beat me. So pyqi is updated to the latest version in unstable and while biom was *running* with this latest version installed it does not *build* any more with this new version as (Build-)Depends. Any idea? Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi, On Wed, Mar 26, 2014 at 10:54:11AM -0400, Yaroslav Halchenko wrote: ok -- done -- see patch attached I just realised that Biom-format has moved to a different location and thus we missed the new version. When inspecting this I realised that it is not using Cython any more - so the problem will be fixed with the new version automatically (Yaroslav, sorry for keeping you busy for no use :-(). I have updated the packaging at svn://svn.debian.org/svn/debian-med/trunk/packages/python-biom-format/trunk/ and have the following questions (specifically also to Tim): 1. Did you forwarded this patch to upstream? http://anonscm.debian.org/viewvc/debian-med/trunk/packages/python-biom-format/trunk/debian/patches/smart_arse_fix.patch?view=markup I think upstream would be interested in incorporating it. 2. Tim, your latest changelog says: Rename add_metadata to something less generic. It will die in the next release anyway. Since the script did not died in 1.3.1 I wonder what next release you might mean. 3. Will the packaging which I prepared in SVN work together with the latest qiime package? This question is probably important before I'll upload it to unstable. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Tim, On Thu, Mar 27, 2014 at 02:37:44PM +, Tim Booth wrote: Hi Andreas, I have updated the packaging at svn://svn.debian.org/svn/debian-med/trunk/packages/python-biom-format/trunk/ and have the following questions (specifically also to Tim): 1. Did you forwarded this patch to upstream? http://anonscm.debian.org/viewvc/debian-med/trunk/packages/python-biom-format/trunk/debian/patches/smart_arse_fix.patch?view=markup I think upstream would be interested in incorporating it. No, I didn't send it. Wasn't sure how they would take it. Well, seems to be worth discussing. I'll do since I also want to approach them about unit tests (which vanished in the new version). 2. Tim, your latest changelog says: Rename add_metadata to something less generic. It will die in the next release anyway. Since the script did not died in 1.3.1 I wonder what next release you might mean. Don't take my word for it. Run the script! --- $ add_biom_metadata This script no longer exists. ... --- If the script denies its own existence who are we to argue? :-) Seems we should not install this at all any more ... 3. Will the packaging which I prepared in SVN work together with the latest qiime package? This question is probably important before I'll upload it to unstable. A quick build on my system failed. Or rather it builds but when I run it: --- % biom usage: biom command [args] The currently available commands are: The following commands could not be loaded: add-metadataError: cannot import name make_command_in_collection_lookup_f convert Error: cannot import name make_command_in_collection_lookup_f show-install-info Error: cannot import name make_command_out_collection_lookup_f subset-tableError: cannot import name make_command_in_collection_lookup_f summarize-table Error: cannot import name make_command_in_collection_lookup_f validate-table Error: cannot import name make_command_in_collection_lookup_f See 'biom help command' for more information on a specific command. --- My guess is that updating PyQi might help (https://pypi.python.org/pypi/pyqi/). Did you already do that? No, I did not. I just realised that this also has moved. To bad if projects are moving and our watch files are pointing to the wrong location. Since it might perfectly be the cause of the problem I could try to do this today evening / night if nobody will beat me. This also scores one more point for the unit test which simply should have made the build fail ... Also, can this now be arch=all or is there still something being compiled that I missed? I think so. I'm not done with the package anyway since it creates a lintian error about google-spying in the docs. Letting you have a look in parallel has proven to be valuable. ;-) Kind regards and thanks for testing Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742691: New version of python-biom-format (Was: Bug#742691: please make robust against ad-hoc (pre/post released) versions of cython)
Hi Andreas, I have updated the packaging at svn://svn.debian.org/svn/debian-med/trunk/packages/python-biom-format/trunk/ and have the following questions (specifically also to Tim): 1. Did you forwarded this patch to upstream? http://anonscm.debian.org/viewvc/debian-med/trunk/packages/python-biom-format/trunk/debian/patches/smart_arse_fix.patch?view=markup I think upstream would be interested in incorporating it. No, I didn't send it. Wasn't sure how they would take it. 2. Tim, your latest changelog says: Rename add_metadata to something less generic. It will die in the next release anyway. Since the script did not died in 1.3.1 I wonder what next release you might mean. Don't take my word for it. Run the script! --- $ add_biom_metadata This script no longer exists. ... --- If the script denies its own existence who are we to argue? 3. Will the packaging which I prepared in SVN work together with the latest qiime package? This question is probably important before I'll upload it to unstable. A quick build on my system failed. Or rather it builds but when I run it: --- % biom usage: biom command [args] The currently available commands are: The following commands could not be loaded: add-metadataError: cannot import name make_command_in_collection_lookup_f convert Error: cannot import name make_command_in_collection_lookup_f show-install-info Error: cannot import name make_command_out_collection_lookup_f subset-tableError: cannot import name make_command_in_collection_lookup_f summarize-table Error: cannot import name make_command_in_collection_lookup_f validate-table Error: cannot import name make_command_in_collection_lookup_f See 'biom help command' for more information on a specific command. --- My guess is that updating PyQi might help (https://pypi.python.org/pypi/pyqi/). Did you already do that? Also, can this now be arch=all or is there still something being compiled that I missed? Cheers, TIM -- Tim Booth tbo...@ceh.ac.uk NERC Environmental Bioinformatics Centre Centre for Ecology and Hydrology Maclean Bldg, Benson Lane Crowmarsh Gifford Wallingford, England OX10 8BB http://nebc.nerc.ac.uk +44 1491 69 2705 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org