Dnia 29 września 2016 09:06:59 CEST, Brian Dolbec <[email protected]> napisał(a): > >OK, I've got the new twisted ebuild almost ready. I have the tests >working, only diabling the network test group and one other that was >disabled in the split pkg ebuild. > >But, the twisted-r1 postrm function is not working correctly. I have >been overriding it in the ebuild so far. It seems that there is an >EPYTHON issue... > >I'll let the traceback speak for itself. > >* python2_7: running distutils-r1_run_phase >twisted-r1_update_plugin_cache >* python3_4: running distutils-r1_run_phase >twisted-r1_update_plugin_cache >Traceback (most recent call last): > File "<string>", line 1, in <module> >File "/usr/lib64/python3.4/distutils/sysconfig.py", line 13, in ><module> > import re > File "/usr/lib64/python3.4/re.py", line 336, in <module> > import copyreg >File "/usr/lib64/python2.7/site-packages/copyreg/__init__.py", line 7, >in <module> >raise ImportError('This package should not be accessible on Python 3. ' >ImportError: This package should not be accessible on Python 3. >Either you are trying to run from the python-future src folder or your >installation > of python-future is corrupted.
At a first glance, it looks like you have some package installed that gets auto-imported magically and is broken itself, therefore breaking everything. > > > >What I can't figure out is of it isn't EPYTHON that is screwed up, how >the hell is >this happening... a simple py3.4 import re causes the 3.4 re.py to try >and import a >py2.7 version of copyreq. > >I can't reproduce this in a terminal window with py3.4 session. >"import re" is fine and >works correctly. So, it can't be a bad python install. > >I'll work on pushing the changes to the twisted ebuilds in the morning >to my buildbot-dev overlay. >But I have attached the WIP ebuild for 16.3.0. You can grab the files >dir from [1] if you want >to try emerging it. > >[1] >https://github.com/dol-sen/buildbot-dev/tree/master/dev-python/twisted/files -- Best regards, Michał Górny (by phone)
