Hi Andreas, I've managed to finish patch for metastudent (I've checked it several times and tried to make it to be compatible with old version of blast as well, although I didn't checked "studentC" part with legacy blast, only with latest version).
Some details on how I checked: I've run metastudent with --keep-temp part, after that looked at contents of temporary directory. If all goes fine, this directory contains 3 directories - "groupA", "groupB" and "groupC". They contain temporary files produced during stages A, B and C. If all goes fine, they contain actual data (graph nodes with their names, etc.), if not, these files are not empty, but contain just a template with no data in it. Results from patched metastudent and latest blast look similar to unpatched metastudent+legacy blast with the same version of metastudent-data package. Numeric results look similar, but I assumed that if they differ, the reason must be a difference in blast and legacy-blast implementation and didn't try to fix if some of numbers differ. I assume that this patch makes metastudent work with blast+ correctly. Tanya. 2016-12-20 22:14 GMT+03:00 Andreas Tille <ti...@debian.org>: > Hi Graham, > > thanks for the bug reports. > > On Tue, Dec 20, 2016 at 06:58:23PM +0200, Graham Inggs wrote: > > Metastudent 2.0.1-3 has been consistently failing its autopkgtests > > since 2016-12-05 [1]. > > > > This corresponds to the upload of ncbi-tools6 6.1.20160908-1. > > Previous tests were successful with ncbi-tools6 6.1.20120620-12. > > I'm not sure if this is relevant. > > That's most probably the very same issue as #848394. Tanya intends to > investigate into this and I guess she will come up with some solution. > Tanya, feel free to ask here for further insigt if you might face any > blockers. > > Kind regards > > Andreas. > > -- > http://fam-tille.de > -- Best wishes, Tanya.