Re: [Outreachy] predictprotein and other RostLab packages
Hi Tanya,
On Sat, Jul 02, 2016 at 01:50:20AM +0300, merlettaia wrote:
>
> It's optional in debian/control file, and it should be - pdb2pqr works
> without it, and it simply shows warning message. All my tests pass without
> apbs. Some of pdb2pqr upstream tests won't pass, because they suppose that
> this optional package is installed and compare output file with expected
> output, which differs when apbs is present or is not (I don't use these
> tests).
I need to admit that I have no idea how sensible it might be to install
apbs when working with pdb2pqr. I'm personally a big fan or stronger
depencencies (like Recommends instead of Suggests). In this sense we
also should test what we are recommending - so it would be better to
reach a state where our test works as well with apbs.
> For now installation of apbs is considered by pdb2pqr to be correct if in
> python REPL command "from apbslib import *" doesn't fail with "ImportError:
> No module named apbslib". For now even when apbs is built with
> -DENABLE_PYTHON=ON flag (as suggested by pdb2pqr) and installed, this
> command still fails.
>
> I'll spend some time on apbs tomorrow and if I'll fail, I'll probably
> continue working on other packages.
That's fine. May be you add some remarks in README.test with the
outcome of your investigation.
> > PS: Please let me know what packages from rostlab are ready for upload.
>
> For now I made simple tests are ready for these packages:
> librg-blast-parser-perl, librg-utils-perl,
> librostlab, librostlab-blast,
> libzerg, libzerg-perl,
> profisis, disulfinder,
> rate4site, reprof.
>
> For some of them I haven't wrote correct descriptions\README.test yet.
> There is a simple test in disulfinder with Readme.test, but I wanted to
> check this package if I could improve tests somehow
OK. I decided to upload disulfinder as is. Its perfectly fine if you
decide to enhance the test later but for the moment I think our users
are served better if we fix the RC bug and upload with a simple test.
> (and probably do the
> same for profisis and rate4site). All current tests should pass when
> packages are installed correctly.
Passes for me. :-)
> I haven't wrote tests for these packages yet: pssh2, profnet, profphd,
> profphd-utils, predictprotein.
>From my point of view its easier if I upload single packages one by one
rather than a larger set of packages. So if you think while the tests
beeing more simple for the moment but working and functional I tend to
do uploads for now. But I'll leave the final say for the upload to you.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: