2014-04-03 0:45 GMT+02:00 Paul Franklin <pf.98052@gmail.com>:
On 4/2/14, Benny Malengier <benny.malengier@gmail.com> wrote:

> It was a big change to do in branches that normally only see bug fixes. I
> assume this is mostly because we have not talked about a 4.1 one yet. It
> should seem clear I'm not in the possibility to handle a 4.1 push, as I did
> for the previous spring-early summer releases in the last years.
> Is there somebody else who wants to do this this year? It involves a lot of
> testing and some extra eye for quality assurance.

I will repeat again my opinion that we shouldn't release
4.1 until 4.0.x has been in use by Windows users for
a long time, almost a year -- as would normally be true.

That hasn't been the case and bugs about 4.0.x are
still coming in, including lots of GUI-related and DB-
related bugs -- which are beyond the ability of many
people to investigate and fix.

I don't see any strong reason to push 4.1 out the
door when we are still in the early life of all the big
changes which appeared in 4.0.0.  For instance
there are very few users who are using 4.0.x under
Python3 and so I suspect there are still lots of bugs
lurking there.  I also am guessing that very few of
the developers are using a "new" Gtk (e.g. 3.10.x),
and if they were I would guess that more such bugs
would be being noticed, and fixed.

We don't have a lot of developers who are active at
the moment and in my opinion this is the wrong time
to be making any major change -- like releasing 4.1.0

All correct. On the other hand, bit-rot can be a problem if we wait too long. It is April and nobody really asked for a 4.1. So I can live with officially postponing to an October release data. Like that we all know the schedule and can work towards that.

I would not do it later. After all, we also don't want developers doing major changes in released branches because they don't trust that their improvements trickle through in a timely manner to the users if working in master. That is bad for stability, see the 2.2 series in the past.

So, others agree to this schedule? Even if so, one of the main developers should step forward to handle it. As my evidence work will probably not land in 4.1 (or are you continuing on it Tim?), I don't have much code of myself in master.


Gramps-devel mailing list