On 13 October 2015 at 22:34, Nathan Woodrow <[email protected]> wrote: > Hey all, > > I think I would just like to add to this that no matter what we do there is > always going to be someone/some company that doesn't like the process. No > release process is perfect and is part of the game of software development. > We can adjust if required but I suspect it will be raised again in the > future.
I agree... looking at this discussion we've got people on two opposing sides - those who want development to slow down and stabilise, and those who are disappointed in the slow speed of development and want faster reviews/merges. We'll never be able to satisfy both sides. So instead of this becoming an endless debate I think the important thing is to work out what practical changes we should make. I'd say the things we CAN change include: - clarifying the process and role of QEPs and when they should/must be done - put some kind of guidelines around PR reviews... when is code suitable for direct merging and when should it be sent via a PR? Who is responsible for this? - ditto for unit tests... our contributor guidelines state that all code added to core must be accompanied by units, so I assume the PSC or someone made this decision in the past. We just all choose to ignore it when convenient ;) I've got an idea for a half-way approach on this, which I'll submit sometime as a replacement to QEP #5 (which even I'll admit was too hardline) - decide on the timeline for LTR releases (1/2 years? overlapping LTR releases?) - and finally.... make a decision about the move to Qt5/Python 3.0/QGIS 3.0 and stop putting this difficult choice off! Nyall _______________________________________________ Qgis-developer mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-developer
