Re: RFS: added autopkgtest for pplacer
Hi,
Am Mon, Apr 11, 2022 at 11:23:13AM -0000 schrieb Mohd Bilal:
> >1. There is no information given where you obtained the data in
> > debian/tests/data from. It is good style to add some script that
> > fetches the data from somewhere (may be via wget or whaterver
> > tool you prefer.
> >
>
> I've added a script[4] in d/tests/ that fetches the test data
Looks good.
> >2. The data should be also mentioned in d/copyright.
> >
> >3. Since the size of the data exceedes the size of the other files
> > it might make sense to create a separate tarball as it is
> > described in the Debian Med team policy[3]. Alternatively
> > you might consider xz compressing the data files (which needs
> > adding them in debian/source/include-binaries) and uncompressing
> > them once they are needed. The compression rate might be
> > sufficiently high to tolerate the size of the compressed data
> > files.
> > It might be also be sensible to create an extra pplacer-examples
> > binary package in case the compression rate is not high enough.
> > A sign for the issue is also given by lintian:
> > I: pplacer: arch-dep-package-has-big-usr-share 12192kB 39%
> >
>
> I've followed the instructions in the med-team policy and moved the tests data to debian-tests-data. The upstream repo[5] for test data doesn;t state a license . So what should be included in the d/copyright file?
May be its a good idea to ask upstream. In lots of cases those data are
in public domain - but this should be clarified.
Also good you created the additional test-data tarball. Please be so kind
to push it to pristine-tar branch. When I try `gbp buildpackage` I get:
/usr/bin/gbp buildpackage --git-pbuilder-options=--source-only-changes
gbp:info: Tarballs 'pplacer_1.1~alpha19.orig.tar.gz, pplacer_1.1~alpha19.orig-debian-tests-data.tar.gz' not found at '../tarballs/'
gbp:info: Creating /home/andreas/debian-maintain/salsa/med-team/build-area/pplacer_1.1~alpha19.orig.tar.gz
gbp:info: Creating /home/andreas/debian-maintain/salsa/med-team/build-area/pplacer_1.1~alpha19.orig-debian-tests-data.tar.gz
gbp:error: Can not find pristine tar commit for archive 'pplacer_1.1~alpha19.orig-debian-tests-data.tar.gz'
> >4. Unfortunately the test does not work in my pbuilder environment.
> > I've added a `set -x` to run-unit-test and so I'Ve got:
> > ...
> >Test 1
> >make: *** No rule to make target 'test'. Stop.
> >autopkgtest [07:44:45]: test run-unit-test: -----------------------]
> >
> >There is no Makefile provided and if I'm checking the build directory
> >the Makefile would need a binary _build/tests/tests.native which
> >contains the according calls to the pplacer binaries. Thus the test
> >only works in a full source tree ... and I admit I'm quite baffled
> >why the Salsa-CI test is passing. It obviously rebuilds the needed
> >binary but I have no idea why.
> >
>
> I've removed the part that uses the Makefile for testing and autopkgtest now uses commands from the tutorial[6] and verifies their checksums. So would this be ok for now?
Sounds good. I'll test once you've pushed
pplacer_1.1~alpha19.orig-debian-tests-data.tar.gz .
> Thanks a lot for this detailed review. Please see whether my changes are appropriate or is there anything more to be done.
You are welcome. We need to make sure to add some paragraph about the
license of the data since the change in the source package will probably
require passing the new queue which means license and copright will be
re-checked. As I said above an issue in the upstream Github repository
might be a promising way to find this out.
Kind regards
Andreas.
> > > [1] - https://salsa.debian.org/med-team/pplacer
> > > [2] - https://fhcrc.github.io/microbiome-demo/
> >[3] https://med-team.pages.debian.net/policy/#embedding-large-test-data
> >
> [4] - https://salsa.debian.org/med-team/pplacer/-/blob/master/debian/tests/get-test-data
> [5] - https://github.com/fhcrc/microbiome-demo
> [6] -https://fhcrc.github.io/microbiome-demo/
>
> >--
> >http://fam-tille.de
>
>
--
http://fam-tille.de
Reply to: