Dear Andreas, ok, I added simple test-suite for mothur. There is TestMothur folder with code of tests, but the data for this tests is missing (I checked upstream repo, it's missing there too). So for my test-suite I just repeated analysis from here https://mothur.org/wiki/Analyzing_H1N1_swine-origin_Influenza_A_isolates (It uses the smallest dataset among other analysis in that wiki).
Thanks for keeping pushing me out of my comfort zone) I'll look into bwa issue. Regards, Nadiya On Tue, Jun 20, 2017 at 1:11 PM, Andreas Tille <[email protected]> wrote: > Dear Nadiya, > > On Tue, Jun 20, 2017 at 12:52:04PM -0700, Nadiya Sitdykova wrote: > > after build mothur creates 4 deb files: > > > > - mothur_1.39.5-2_amd64.deb > > - mothur-dbgsym_1.39.5-2_amd64.deb > > - mothur-mpi_1.39.5-2_amd64.deb > > - mothur-mpi-dbgsym_1.39.5-2_amd64.deb > > > > but only one of mothur-dbgsym and mothur-mpi-dbgsym can be installed. > > if I try to install mothur-dbgsym after installation of > mothur-mpi-dbgsym I > > got an error: > > ... > > if I try to install mothur-mpi-dbgsym first and then install > mothur-dbgsym > > I got a symmetric error: > > ... > > Is that an expected behavior? Or am I doing something wrong? > > Admittedly its not really expected but by the nature how mothur and > mothur-mpi are build it is not astonishing since the build process is > repeated with different options and thus the debug info ends up in the > same place. > > If you ask me in practice there is no real harm done since a user will > most probably not debug the two executables at the same time and should > simply install the debug package that really fits. Strictly speaking a > Conflicts should be injected into the control information but I have no > idea how to do this in auto generated packages. > > I appreciate that you investigated this issue and reported it here but > for the moment I do not think this should stop us from uploading the > package with the autopkgtest. > > > I have another issue that might be worth spending some time into it. I > have reported a issue 108 to bwa upstream[1] which provides a test case > where bwa crashes. It seems upstream has given up on this since my > pings were not answered. If you have some experience with gdb and would > like to investigate this it would be a valuable target. May be using > valgrind might help as well since I suspect a very nasty memory issue > which happens somewhere else than where the crash finally occures. > > While I think upstream has better chances to spot the issue they seem to > do nothing and if you consider this problem interesting it might be > worth spending some time on it. Feel free to tell me if you rather > prefer to stick to your timeline - that's perfectly fine as well. > > Kind regards and thanks for your work so far > > Andreas. > > [1] https://github.com/lh3/bwa/issues/108 > > -- > http://fam-tille.de > >
