Re: ITP: HTSeq

[Diane Trout <diane@ghic.org>]

Hello Andres,


> So the only way to properly build is using `gbp-clone --git-ignore-new`
> which is a bit nasty.  The reason for this is that the clean target does
> something else but restoring the original.  I could imagine that the
> Debian Python team might have some experience with things like this and
> knows a good way to work around.  IMHO git-buildpackage is terribly
> picky about things like this and autobuilders (or (p)debuild) are more
> tollerant.  But I'd be happy if this could be solved before we do a final
> upload if time permits.

I git rm'ed the files that get generated by the build process and added a 
debian/gbp.conf file to filter them out of master next time someone does a gbp 
import-orig.

I think the entries in debian/source/options will keep debuild's orig diff 
from complaining about those build-generated files as well.

> 
> Two additional remarks:  If you have *really* good contacts to upstream
> you can start nitpicking and might give them a hint about the lintian
> info (which you get via lintian -I) about

I saw the couple of spelling errors, but as they haven't gotten back to me yet 
for the fixes to the setup.py, I thought I should wait before sending more 
patches.

> The other thing is that I noticed you have implemented means to rename
> original upstream version 0.5.4p3 to 0.5.4.3 per uversionmangle in
> d/watch.  I just want to mention that leaving the 'p' inside the version
> string would be perfectly OK.  There is no rule that enforces numbers
> and '.' exclusively.  Feel free to leave it as it is currently if you
> have your reasons to do so - I just want to mention that it is not
> needed and in some case it could lead to confusions.

I undid the mangling, I looked at the debian policy and as long as the version 
number sort correctly letters are ok. 

> 
> The next step would be to file an ITP bug report.  Do you want to do
> this yourself or do you want me to do this (on behalf of the team - so
> the package wil remain yours).

I'd like to submit the ITP, though as I'm still learning how to make a correct 
ITP I'd like a review first.

Also while I was working on the ITP I changed my previous description a little 
bit.

--[ITP]--
To: submit@bugs.debian.org
Subject: ITP: htseq -- high-throughput genome sequencing analysis.

Package: wnpp
Severity: wishlist
Owner: Diane Trout <diane@ghic.org>
X-Debbugs-CC: debian-med@lists.debian.org

    Package name: htseq
         Version: 0.5.4p3
 Upstream Author: Simon Anders <sanders@fs.tum.de>
             URL: http://www-huber.embl.de/users/anders/HTSeq/doc/overview.html
         License: GPL-3+
Programming Lang: Python
     Description: HTSeq can be used for a number of common high-throughput 
genomics analysis tasks.
 .
   * Getting statistical summaries about the base-call quality scores to
     study the data quality.
   * Calculating a coverage vector and exporting it for visualization in
     a genome browser.
   * Reading in annotation data from a GFF file.
   * Assigning aligned reads from an RNA-Seq experiments to exons and
     genes.

--[ITP]--