I didn't know aiotest even existed. I've also never used tox. I've always just run asyncio's own unittests using runtests.py. Maybe, once aiotest is fixed, we should update the Travis CI setup to run tox with aiotest? Currently it just runs runtests.py.
The policy for new asyncio releases has been to match these up with CPython releases. We used to sync these with 3.4 releases (the last is 3.4.3) but now 3.4 is no longer receiving updates maybe we should do one to match 3.5.1? It should contain exactly the same code as was released with 3.5.1 (not the current repo head). On Mon, Feb 1, 2016 at 4:05 AM, Victor Stinner <victor.stin...@gmail.com> wrote: > Hi, > > I synchronized GitHub asyncio with CPython stdlib asyncio. A few files > were outdated :-/ > > When I ran "tox" to validate my changes on CPython stdlib asyncio, I > noticied that aiotest is now failing. Nobody ran tox recently? > > I updated aiotest to support the new stop() semantics: > http://bugs.python.org/issue25593 > > aiotest is a test suite to validate an implementation of asyncio. When > I ran aiotest tests on Trollius, stdlib asyncio of Python 3.4 and > asyncio from PyPI: all tests failed with the new stop() semantics. > > Trollius is outdated, but it's a known issue, and I'm not intested to > update Trollius (see the other thread about Trollius). > > My system /usr/bin/python3 is Python 3.4.3 and asyncio of this version > still has the old stop() behaviour. > > https://pypi.python.org/pypi/asyncio/ looks quite old: last release > almost one year ago. > > Questions / TODO: > > * kill Trollius :-) > * new asyncio release to PyPI? > * new aiotest release supporting the new stop behaviour (I already > pushed a change, it's configurable: old/new beheaviour, > config.stopping=False) > > Victor -- --Guido van Rossum (python.org/~guido)
