[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Exonerate Test Failures



Hi Canberk,

On Tue, May 31, 2016 at 04:43:45AM +0300, Canberk Koç wrote:
> 
> Finally i manage to figure out what is wrong in tests.

:-)

> At last resort i run tests manually and find out these results.

I admit I *always* try to design the tests that it can be run manually
on the local machine.  I also usually install the tests into
   /usr/share/doc/<packagename>
and add a README.test there to inform users how they can test the
package manually.  This serves in most cases as an informative example
how to use the software.

> First:
> 
> *fastaremove.test.sh <http://fastaremove.test.sh>line no: 25*
> 
> *$FASTAREMOVE $INPUTFILE CALM_HUMAN(ARGUMENT PART) > $OUTPUTFILE*
> 
> Upper test fails because upstream package have a little bug in test.sh.
> There must be >CALM_HUMAN for input argument but they forget ">" character :-)

Please forward this upstream (address is in debian/copyright) but
see below first.
 
> *fastaindex.fastafetch.test.sh <http://fastaindex.fastafetch.test.sh>
> 
> line no: 34
> $FASTAFETCH $INPUTFILE $INDEXFILE $IDENTIFIER
> 
> line no:52
> run_fastafetch A_MISSING_FROM_START 1
> run_fastafetch M_MISSING_FROM_MIDDLE 1
> run_fastafetch z_MISSING_FROM_END 1*
> 
> Upper test fail because in source file it take an argument  -1 for
> fail and in test.sh they use -1 parameter for showing fail situation.
> At these lines they test for what program act when there is a missing
> part in fasta file so test working correct actually :-)
> if you need more info i can show c.file line numbers too .
> 
> Now what we do for correct these mistakes :)

   quilt push -a
      # this applies all previous patches
   quilt new fix_test1.patch
      # this creates a new patch
   quilt edit <path_to>/test.sh
      Add the missing ">"
   quilt refresh
      # this tells quilt you are done with the patch
   quilt new fix_test2.patch
      # this creates another patch (feel free to leave
      # this out and block-fix everything in one patch
   quilt edit <broken_c_file_with_wrong_return_code>
      Fix the return code
   quilt refresh
      # done with patching
   quilt pop -a
      # unapply all patches
   editor debian/patches/fix_test1.patch
      Add the DEP3 fields:

       Author: Canberk Koç <canberkkoc@gmail.com>
       Last-Update: <timestamp>
       Description: <explain what you explained above
        to make upstream understand the problem>
   editor debian/patches/fix_test2.patch
      If you decided to use two patches add DEP3 header as well
   git add debian/patches/fix_test*.patch
   git commit -a -m 'Fix testsuite'
   git push

Once you are done with this I can check.  Then we can point upstream to
our patches to enable inclusion into next release.
   
Please do not hesitate to ask if the advise above is to short.

Kind regards

     Andreas.
 
> Best regards
> 
> 
> 
> 
> Canberk Koç
> [image: https://]about.me/canberkkoc
> <https://about.me/canberkkoc?promo=email_sig&utm_source=email_sig&utm_medium=email_sig&utm_campaign=external_links>
> 
> 2016-05-26 13:10 GMT+03:00 Canberk Koç <canberkkoc@gmail.com>:
> 
> > Hello Andreas,
> >
> > Actually hints give me some ideas but i am struggling with understanding
> > what is wrong. Package do these tests in building by itself no problem
> > occurs but adt-run give 2 test failure though one test works once also :-)
> > I look at build log see some warnings about deprecated methods i'll gather
> > all together and turn back to you.
> >
> > Best regards
> >
> > Canberk Koç
> > www.about.me/canberkkoc
> > 26 May 2016 12:50 tarihinde "Andreas Tille" <andreas@an3as.eu> yazdı:
> >
> > Hi Canberk,
> >>
> >> since you did not sent any answer I assume the hints were helpful
> >> enough.  Is this assumption true?
> >>
> >> Kind regards
> >>
> >>        Andreas.
> >>
> >> On Tue, May 24, 2016 at 08:09:11AM +0200, Andreas Tille wrote:
> >> > Hi Tanya,
> >> >
> >> > > ...
> >> >
> >> > thanks for your verbose and (without testing sounding) sensible input.
> >> >
> >> > On Tue, May 24, 2016 at 05:26:06AM +0300, merlettaia wrote:
> >> > > Hope this helps.
> >> > > Please, somebody correct me if my interpretation is wrong.
> >> >
> >> > Canberk, please let us know here if you need some further help or
> >> > testing.
> >> >
> >> > Kind regards
> >> >
> >> >          Andreas.
> >> >
> >> > --
> >> > http://fam-tille.de
> >> >
> >> >
> >>
> >> --
> >> http://fam-tille.de
> >>
> >>

-- 
http://fam-tille.de


Reply to: