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

Bug#724718: Dependency packaging for 2.8.15



I have now more carefully investigated the dependencies and added blocking bugs:

REQUIREMENT               SRC PACKAGE      STATUS     BUGS
------------------------------------------------------------------
Safe>=0.4                 safe             ITP        #850088
colorclass>=2.2.0         colorclass       ITP        #850087
flask-cors>=2.1.2         flask-cors       ITP        #850091
flask-restplus==0.8.6     flask-restplus   ITP        #850089
rpyc                      rpyc             ITP        #850097
terminaltables>=3.1.0     terminaltables   ITP        #850093
FeedParser>=5.2.1         feedparser       OUTDATED
cherrypy>=3.7.0           cherrypy3        OUTDATED   #828942
subliminal >= 2.0rc1      subliminal       OUTDATED   #850098
vcrpy>=1.7.4              vcr.py           OUTDATED
flask-login>=0.4.0        flask-login      PY2-ONLY   #802614
pynzb                     pynzb            PY2-ONLY

Droppable dependencies: I think that coverage checks are not particularly relevant to the debian package, so I can avoid having to package codacy-coverage. Similarly, gitpython seems to only be used in a script to update the upstream changelog, so I can also omit that.

Pinned versions: Note that guessit and flask-restplus are pinned to old versions by upstream flexget. If flask-restplus can't be unpinned then I'll have to adjust the ITP to package the old version flexget wants. If guessit can be unpinned (which I suspect it can since the dependency conflict which cause the situation seems to have disappeared in new upstream guessit) then that will possibly also need to be updated before flexget can be packaged. See: https://requires.io/github/Flexget/Flexget/requirements/?branch=master

pynzb: Upstream seems abandoned as it hasn't been touched in 8 years. This is why it's a python2-only package. Only one flexget plugin needs this dependency so it could be reasonable to just patch that out for now to simplify things. Since pynzb is pretty small SLOC-wise it might be possible to eventually adopt it upstream or convince someone else to take over.

Since there is a lot of work to be done this could take me a while but I'll try not to let the effort lapse again! Hopefully I can get some help from the Python modules team.


Reply to: