Python 3.5 is stable and quick enough to be considered.
Cheasy answer: Patches are accepted! ;-)
Python 3 is still so much of a regression for tools like Mercurial that Python 3 not is an option yet when using Mercurial. Last time I checcked, the only dulwich version that actually works for us doesn't support Python 3 - it started a wack-a-mole when they added Python 3 support. Finally, Kallithea is based on Pylons which is Python 2 only. Work is in progress for switching to TurboGears.
Former userAccount Deleted
For mercurial, as I wrote in #192, I should be treated as an external dependency, as git.
For the turbogears2 migration, I can be tester, but I've no time to develop on a new framework.
Is there any update on this? The current setup.py still requires 2.7.
Python 2 is soon to be unsupported (https://pythonclock.org/), Therefore it’s critical that Python 3 be supported, at least optionally. I’d really like to use Kallithea, but I can’t knowingly deploy anything to it if there’s a chance it’ll be using an out of date, and arguably insecure version of Python.
Please don’t go the way of calibre, maintaining Python 3 support yourself isn’t viable!
Mercurial just made the first release with official Python3 support. The next Mercurial release in 3 month might be a good first version for us to target.
With Kallithea being an application and not a library, it shouldn’t matter much to users which Python version it is using. Rushing will just give us extra work. No doubt we will make it work out.
If you want to contribute, a good first step would be to do a quick’n’dirty proof of concept of getting something working. That will help understanding the work that is coming, and perhaps land refactorings that can make it easier.
The python2 clock has started ticking https://pythonclock.org/ … has a clock for upgrading kallithea to python3 been started?