On Wed, Jan 23, 2019 at 09:58:45AM +0000, Nick Morrott wrote: > > Cool. Thanks a lot. Your outright contribution was very helpful and > > really welcome. The only change I made is to revert your commit > > 38022ac5 since there is no point in changing the upstream source that > > way. I agree upstream should not ship that file but if its in the > > upstream tarball I see no need to remove it here. > > Apologies about that.
No problem. :-) > My build failed immediately due to the presence > of that file when I started testing, and removing it got the build to > proceed so I could look at the test failures. > > That file is not present in their upstream github repository at the > release date of 0.9.37. pypi points to the repo as the definitive > source location so it's my fault for not investigating the pypi > tarball further. Strange. I have no idea how this file might have sneaked in. It might be that PyPI changed the content of the file without renaming it? I used what I once obtained and stored in pristine-tar. May be you did not used `gbp buildpackage` - thus the difference. > If it persists it can obviously always be removed > automatically via d/copyright if desired. It never harmed my build. I tend to not touch the original tarball via d/copyright to keep the md5sum equally to the download - which does not seem to be the case any more if you are right (I did not mind to check admittedly). I guess it will "heal" for the next version and I have more important things than this on my desk now. > > Definitely. Package is uploaded - one worry less. :-) > > Excellent! Thanks again Andreas. -- http://fam-tille.de