Hello Andreas, > I would have uploaded but I get: > > > run_lefse comparative pass > format_input comparative pass > lefse2circlader comparative pass > plot_cladogram comparative fail I cannot reproduce this and I just did a clean build and installed a fresh *.deb. I did however do some refinements and any fails will result in script execution which will display the output so please pull. > I've observed your commits to phyx but did not acted upon it since you > did not confirmed that you are ready here. More about this: this is ready. Both the run_tests.py and gui interface work. A note about the run_tests.py. Before the conversion, 9 tests failed (I think 31 passed) and after the conversion, there is the same failure rate. Not sure if this is some upstream fault because I have not been able to get a fresh upstream copy and directly test the run_tests.py, but I assume it will be the same as what I get now. > I'm aware of this - but I wonder whether this really small contribution > compared to all those free software code served by us couldn't be > expected. I know ... probably not. But I might be permitted to dream a > bit. I agree with you. I still would have expected *some* form of awareness or feedback. Kind regards, Shayan Doust On 11/09/2019 16:52, Andreas Tille wrote: > Hi Shayan, > > On Wed, Sep 11, 2019 at 01:37:01PM +0100, Shayan Doust wrote: >> Thanks for the run-through of this package. > > Thanks to you. > >> When I started porting this package, I ran all the binaries to have a >> baseline of what to expect as "normal functionality". At that time, it >> was "No module named 'lefse'", which I assumed this to be "normal" as >> maybe lefse needed integration and not direct execution. I assumed that >> it being uploaded to unstable that it passed quality assurance, however >> some things always slip through which is understandable. >> >> Actually, lucky that I picked this package because I would rather have >> this rectified now than any later. > > :-) > >>> Once the script was running I realised that some Build-Depends were >>> missing. I now added debian/createmanpages to create manpages for each >>> script (+some manual editing). That way at least each script was >>> running once. >> >> With this package already being in unstable, I refrained from using a >> clean environment just to significantly speed up the build process so >> not realising missing build-depends slipped through. That reminds me, a >> commit persists on my side for the phyx package which has createmanpages >> but no generated man pages included. > > I've observed your commits to phyx but did not acted upon it since you > did not confirmed that you are ready here. > >>> Would you volunteer to write some autopkgtest - even if it would be >>> very simple to just call the script? >> >> I'd gladly volunteer. > > Good! > > I would have uploaded but I get: > > > run_lefse comparative pass > format_input comparative pass > lefse2circlader comparative pass > plot_cladogram comparative fail > > > No time to check this in more detail right now. Please git pull for > some minor changes of mine (DEP3 and spelling). > > >> End users are not always the reliable source of information. Most people >> want something that "works", so you'll find people installing the >> software, seeing that it doesn't work and looking elsewhere. Unless you >> have a higher margin of a sustained userbase, the ordinary population >> tends to not bother with bug reports and following up on a software. >> Could be those people ended up compiling from source instead. > > I'm aware of this - but I wonder whether this really small contribution > compared to all those free software code served by us couldn't be > expected. I know ... probably not. But I might be permitted to dream a > bit. > > Kind regards > > Andreas. >
Attachment:
signature.asc
Description: OpenPGP digital signature