Web lists-archives.com

Re: Exclicitly or "implicitly" mark architectures a packages does not build




On Wed, Dec 20, 2017 at 07:24:28PM +0000, Holger Levsen wrote:
> On Wed, Dec 20, 2017 at 08:10:19PM +0100, Andreas Tille wrote:
> > May be I should write an according query and than close all those
> > bugs ...
> 
> please do ;)

OK, I'm working on this:

  https://anonscm.debian.org/cgit/collab-qa/check_build_depends.git/tree/query_missing_build_depends.sh

Its not finished but a start:

$ query_missing_build_depends.sh -m i386 | grep pysam
 ariba                                          | python3-pysam                    |         | 
 circlator                                      | python3-pysam                    |         | 
 fitgcp                                         | python-pysam                     |         | 
 htseq                                          | python-pysam                     |         | 
 iva                                            | python3-pysam                    |         | 
 paleomix                                       | python-pysam                     |         | 
 python-pbcore                                  | python-pysam                     |         | 
 python-pybedtools                              | python3-pysam                    |         | 
 python-pysam                                   | bcftools                         |         | 
 sga                                            | python-pysam                     |         | 

This has more hits for packages that do not build on i386 due to a
missing pysam on this architecture than respective bugs reported
in BTS.  The query is **not** finished and needs further polishing
but at least it proves my point that quering UDD for this kind of
problem is better than reporting bugs manually (or brute force
builds to find out problems like this).

I'll keep on polishing and will need to write some more documentation
than `query_missing_build_depends.sh -h` will provide.  I also have not
yet checked for false positives (the ones above are correct hits,
thought).

As I said I'd like to close those bugs that are not really bugs as
discussed here and which can be found by the query above easily.

Kind regards

        Andreas.

-- 
http://fam-tille.de