[Debian-med-packaging] MNE v0.8 packaging

Andreas Tille andreas at fam-tille.de
Mon Aug 4 07:10:36 UTC 2014 

Hi Alexandre,

On Sun, Aug 03, 2014 at 06:44:59PM -0400, Alexandre Gramfort wrote:
> hi,
> 
> I pushed a fix on the copyright file.

I'd (strongly) recommend to use

    cme fix dpkg-copyright

You find what packages you need to install for cme in Debian Med policy
document.  By doing so I found two remaining incompatibilities in
d/copyright.  When fixing them I changed the file to match the original
intention of DEP5 to use a main license section if the very same license
is used more than once.

> git-buildpackage reported no error for me.

That's good to know. :-)

However, the package builds but since the compressed JS files are
remaining in the upstream tarball it is considered as affected by a
"binary without source" problem.  You can reproduce this by installing
the latest lintian from unstable (lintian 2.5.25) and run

   lintian -i python-mne*.changes

which will print the errors I was quoting in my last mail.  Any upload
with this lintian error will be rejected.

There are two ways to fix this:

  1. Provide the uncompressed source for all mne/html/*.min.js files
     and make sure you are using these in the final package.
  2. Use

      Files-Excluded: mne/html/*.min.js

     in debian/copyright and create a DFSG free tarball for instance
     by using

      get-orig-source:
		uscan --verbose --force-download --repack --compress xz

     as new target in debian/rules.  The tarball should be named
     python-mne_0.8.2+dfsg.orig.tar.gz and you can easily approach this
     by adding

      opts="uversionmangle=s/$/+dfsg/"

     in debian/watch.  Than `git import-orig --pristine-tar` this tarball.
     Finally you need to Depend from the libjs-* packages needed and
     create symlinks to the relevant places.

Since as far as I can see all JS libraries are packaged for Debian I'm
storngly in favour of the later way since this the lest questionable way
regarding the usage of compressed JS files as far as I'm interpreting
past discussions on debian-devel.  If anything remains unclear feel free
to ask here on this list (BTW, there is no need to CC me any I also
think Yaroslav is reading the list - if you confirm that you are reading
it I also stop CCing you personally which is the usual list policy).

> hope we're good now

As said above we need to care for all lintian errors first.

Kind regards

      Andreas.
 
> Alex
> 
> On Sat, Aug 2, 2014 at 4:58 PM, Andreas Tille <andreas at fam-tille.de> wrote:
> > Hi again,
> >
> > while the tests are now working nicely there are some remaining
> > lintian errors:
> >
> > Successfully signed dsc and changes files
> > E: python-mne source: source-is-missing mne/html/bootstrap.min.js
> > N:
> > N:    The source of the following file is missing. Lintian checked a few
> > N:    possible paths to find the source, and do not find it.
> > N:
> > N:    Please repack your package to include the source or add it to
> > N:    "debian/missing-sources" directory.
> > N:
> > N:    If this is a false-positive, please report a bug against Lintian.
> > N:
> > N:    Severity: serious, Certainty: possible
> > N:
> > N:    Check: cruft, Type: source
> > N:
> > E: python-mne source: source-is-missing mne/html/d3.v3.min.js
> > E: python-mne source: source-is-missing mne/html/jquery-1.10.2.min.js
> > E: python-mne source: source-is-missing mne/html/jquery-ui.min.js
> > E: python-mne source: source-is-missing mne/html/mpld3.v0.2.min.js
> > E: python-mne source: build-depends-on-obsolete-package build-depends: python-support => use dh_python2 instead
> > N:
> > N:    The package build-depends on a package that has been superseded. If the
> > N:    superseded package is part of an ORed group, it should not be the first
> > N:    package in the group.
> > N:
> > N:    Severity: important, Certainty: possible
> > N:
> > N:    Check: fields, Type: binary, udeb, source
> > N:
> > W: python-mne source: syntax-error-in-dep5-copyright line 51: Cannot parse line "Copyright (c) 2005-2012, Michele Simionato"
> > N:
> > N:    The machine-readable copyright file didn't pass Debian control file
> > N:    syntax check.
> > N:
> > N:    This issue may hide other issues as Lintian skips some checks on the
> > N:    file in this case.
> > N:
> > N:    Refer to
> > N:    https://www.debian.org/doc/packaging-manuals/copyright-format/1.0/ for
> > N:    details.
> > N:
> > N:    Severity: normal, Certainty: possible
> > N:
> > N:    Check: source-copyright, Type: source
> > N:
> > gbp:error: '~/bin/git-pbuilder' failed: it exited with 1
> >
> > Since I'll go offline until tomorrow evenining right now I can not
> > comment on it.  Feel free to seek for other help on the mailing
> > list.
> >
> > Kind regards
> >
> >        Andreas.
> >
> >
> > On Sat, Aug 02, 2014 at 03:42:07PM +0200, Alexandre Gramfort wrote:
> >> hi Andreas,
> >>
> >> I think I pushed a fix.
> >>
> >> Sorry for the noob question but what do you mean by "you can avoid
> >> this kind of problems when using `dch` to edit
> >> the changelog" ? how does the packaging system run the tests?
> >>
> >> Can I check myself if the package built?
> >>
> >> Alex
> >>
> >>
> >> On Sat, Aug 2, 2014 at 3:09 PM, Andreas Tille <andreas at fam-tille.de> wrote:
> >> > Hi Alexandre,
> >> >
> >> > the state in Git does not build.  While I fixed the missing Debian
> >> > revision number in the changelog entry (please be so kind to always run
> >> > a test build - with your latest commit the package does not even start
> >> > building; you can avoid this kind of problems when using `dch` to edit
> >> > the changelog) the test suite is running into one error.  I attached a
> >> > build log.  Any hint how to run this successfully?
> >> >
> >> > Kind regards
> >> >
> >> >      Andreas.
> >> >
> >> > On Sat, Aug 02, 2014 at 12:31:52AM +0200, Alexandre Gramfort wrote:
> >> >> hi Andreas,
> >> >>
> >> >> I think i've made it :)
> >> >> I pushed v0.8 and immediately pushed v0.8.1 which is our quick fix release.
> >> >>
> >> >> Let me know if everything looks good on your side.
> >> >>
> >> >> thanks for your help and patience :)
> >> >>
> >> >> Alex
> >> >>
> >> >>
> >> >> On Fri, Aug 1, 2014 at 11:25 AM, Andreas Tille <andreas at an3as.eu> wrote:
> >> >> > Hi,
> >> >> >
> >> >> > just reading
> >> >> >
> >> >> >    http://lists.alioth.debian.org/pipermail/debian-med-packaging/2014-July/028210.html
> >> >> >
> >> >> > and wondering, whether this might be an alternative to the manual
> >> >> > merge ...
> >> >> >
> >> >> > Kind regards
> >> >> >
> >> >> >        Andreas.
> >> >> >
> >> >> > On Fri, Aug 01, 2014 at 11:17:58AM +0200, Andreas Tille wrote:
> >> >> >> On Fri, Aug 01, 2014 at 03:34:45AM +0200, Alexandre Gramfort wrote:
> >> >> >> > >> uscan --verbose --force-download
> >> >> >> >
> >> >> >> > this works
> >> >> >>
> >> >> >> OK.
> >> >> >>
> >> >> >> > >> git import-orig --pristine-tar ../python-mne_0.8.orig.tar.gz
> >> >> >> >
> >> >> >> > I had to specify debian as my debian-branch in .git/gbp.conf
> >> >> >> > as otherwise it was merging to my master branch.
> >> >> >>
> >> >> >> Usually it should merge to *upstream* branch.  I have no gbp.conf file
> >> >> >> (which rather would belong to the debian/ dir if I'm not totally
> >> >> >> mistaken) and it always worked fine that way.
> >> >> >>
> >> >> >> > Yet this now produces
> >> >> >> > merge conflicts. It there a way to use something like
> >> >> >> > --strategy-option theirs in the merge?
> >> >> >> >
> >> >> >> > is it fine if I do the merge manually ? that will keep all the history though.
> >> >> >>
> >> >> >> Do you get
> >> >> >>
> >> >> >> $ git import-orig --pristine-tar ../python-mne_0.8.orig.tar.gz
> >> >> >> ...
> >> >> >> pristine-tar: committed python-mne_0.8.orig.tar.gz.delta to branch pristine-tar
> >> >> >> gbp:info: Merging to 'master'
> >> >> >> gbp:error: Merge failed, please resolve.
> >> >> >>
> >> >> >>
> >> >> >> $ LC_ALL=C git status
> >> >> >> ...
> >> >> >> Unmerged paths:
> >> >> >>   (use "git add <file>..." to mark resolution)
> >> >> >>
> >> >> >>         both modified:   mne/tests/test_utils.py
> >> >> >>         both modified:   mne/utils.py
> >> >> >>
> >> >> >> ?
> >> >> >>
> >> >> >> Please always post such details rather than "this now produces merge
> >> >> >> conflicts".  Any yes, while I have no idea why this is happening this is
> >> >> >> probably the way to go.  Some more experienced Git experts than me might
> >> >> >> insist if I'm wrong.
> >> >> >>
> >> >> >> > thanks for your help
> >> >> >>
> >> >> >> Kind regards
> >> >> >>
> >> >> >>       Andreas.
> >> >> >>
> >> >> >> --
> >> >> >> http://fam-tille.de
> >> >> >>
> >> >> >> _______________________________________________
> >> >> >> Debian-med-packaging mailing list
> >> >> >> Debian-med-packaging at lists.alioth.debian.org
> >> >> >> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-med-packaging
> >> >> >>
> >> >> >
> >> >> > --
> >> >> > http://fam-tille.de
> >> >>
> >> >
> >> > --
> >> > http://fam-tille.de
> >>
> >
> > --
> > http://fam-tille.de
> 

-- 
http://fam-tille.de

More information about the Debian-med-packaging mailing list