On 08/11/2016 05:05 PM, Craig Rodrigues wrote:
On Thu, Aug 11, 2016 at 6:48 AM, Petr Viktorin <pviktori(a)redhat.com
<mailto:pviktori@redhat.com>> wrote:
Instead, we should shift our focus from porting specfiles to
upstream projects. At this point, if some software is easy to port
it was probably ported already; what we're left with are either
tough nuts to crack or projects with few people relative to the
codebase size. Some projects that come to mind that could use
attention are GTK, Mercurial, Samba, wxPython, PySide, Koji & Fedora
infra, Ansible.
I don't know yet what our priorities should be here, but that's the
general direction.
You should decide which projects have strategic importance to be ported
to Python 3,
and make a list on this project's web page of the ones that are of most
interest to focus on.
Right; I'll certainly evaluate strategic importance. But I don't plan to
post those results on this project's pages.
We (Red Hat python-maint) are just one team contributing to Fedora, and
our strategic goals differ from anyone else's.
We'll definitely share our strategy related for the shared goal: success
of Python itself on Fedora. But when it comes to individual projects
(say, choosing Ansible or Salt), we can't speak for the Python SIG, and
I'd like to keep that distinction clear.
Some of the projects you listed are slower to port to Python 3 for
various reasons.
For example:
- Ansible needs to maintain compatibility with Python 2.4 (
https://docs.ansible.com/ansible/developing_modules_python3.html ).
This makes porting really hard.
- wxPython has project Phoenix which runs under Python 3 (
https://wiki.wxpython.org/ProjectPhoenix ) but not
all wxPython widgets have been ported to Python 3
- Pyside requires changes due to differences in the Python C API in
Python2 and Python3 (
http://wiki.qt.io/PySide_Python_3_Issues )
That kind of information is valuable, and we'll find a way to share what
we find so anyone can evaluate what to focus on.
I've been spending a lot of time in the past month contributing
Python
patches to the Twisted library,
and I have learned that for serious Python codebases, porting to Python
3 can be non-trivial.
It's not just size; network-related frameworks were probably hit the
worst by the change. As the easy stuff is done, in general the porting
effort will have a tendency to get worse.
Hopefully we can help balance that with tools and documentation.
--
Petr Viktorin