[GRASS-dev] where to store GRASS settings: GISRC and wx settings

[Hamish]

Markus Neteler wrote:
> I am also not quite sure about the need of the 6.5 branch?

I think it is better to think of it by its name, develbranch_6,
than as the 6.5 branch. Then much the same as releasebranch_5_5
we should just leave it in peace to grow old and collect dust.
Calling it 6.5 implies a future 6.5.0 release of development
work, like we did for 6.3.0, but there are no plans for that.

> For me it is sufficient to have 6.4 as production system
> and 7 as experimental system.
> 
> Honestly, who's *using* it as a user?

I use it.. my normal workflow is to use it as my everyday scratch
pad and for experimentation, then switch to 6.4 when I need to
generate real numbers I will put in a report. (which also
encourages me to properly test and backport changes in the
weeks following development :)

I'm not comfortable using trunk for my job-work, and I'm not
comfortable developing directly in 6.4. As most of my current
contributions stem from fixing bugs/overcoming problems+needs
I encounter as I do my day job, I typically end up in the middle
of the two: 6.5.  I'm probably not a standard user, but perhaps
not too far from how a power-user might operate.

> Because: yes, it can (could?) be branch to be used for
> testing but I actually don't know anyone who really uses
> 6.5 for work.

fwiw my usual dev mode has to put any potentially dangerous
6.x specific change in 6.5 first for two weeks, then try not
to forget to back port it to 6.4 once it has been well tested.
(And as a rule only commit things to 6.4 that I am quite sure of)

> Please tell us if you use 6.5 for work to understand if
> this branch is really needed.

in its function as develbranch_6, and to preserve its svn
history, I would certainly argue strongly not to simply delete
it. but I'd be very happy to declare all 6.x to be in bug-fix-
only mode (except perhaps for already begun wxGUI projects)
and let it sit there idle.

aka it would seem to me like defeat to remove choice from users
because us developers couldn't muster discipline, and so needed
to implement technological fixes to what are essentially social
problems. (maintaining focus)

Glynn:
>> We should release GRASS 6.4.2 and end the version 6 series.

I am not against a 6.4.3 release which just fixed some bugs
people found. Similarly I am not against a 5.4.2 release if there
is some terrible bug (ie data corruption or copyright issue)
found which goes back there. That doesn't mean it has to be
anything like the place of development it has been up 'til now
though.

Madi wrote (& Massimo echoed):
> As a user, I can say that it is actually the most usable
> IMHO, G7 is not stable enough and the 6.4 is not so up-to-date.

since 6.4.1 was released a flood of backports have happened and
now AFAIK 6.4 and 6.5 are not anything like as out of sync as
they used to be. I wonder if I should set up a cron job on the
adhoc server to generate a weekly diff. I don't think that
everything currently in devbr6 must be backported to relbr64, but
we should be aware if we've meant to backport something but
forgotten to do it.


Any reason to wait any longer for 6.4.2rc1?


letting go is hard to do,
Hamish