[GRASS-dev] devbr6

Hamish hamish_b at yahoo.com
Sun Mar 10 04:04:33 PDT 2013

[Subject: Re: [GRASS-dev] [GRASS GIS] #1889: wxPy loc'n wiz: don't auto-launch into PERMANENT]

> Comment(by martinl):
> just small note: it's probably related to the fact, that
> you are probably only developer who wants to do "development"
> in GRASS 6 branches.

I don't mean new development, I mean maintenance.

> There were several discussions in the past, and I don't want
> to open this topic again.

you didn't; I did, since there was a small problem with out of order development happening in the release branch which I wanted to make a noise about before it got too far away.

> The result will the same. Please bear in mind, that you are
>  probably only one from devs who really wants to keep
> `develbranch_6` alive.

I'm not into making threats or big ego shows so I'll just say the way it is. I see my long term role with GRASS as gardening the stable branch, making sure bug fixes get implemented, tested, & sync'd, etc. To do that properly a devbr6 is needed as a staging and testing area. If that's not respected, it seems to me that trying to keep relbr64 in a sane state will be impossible, and so a waste of my time and the best thing that I could do for my own motivation is simply walk away and work on something else more productive + fun. You may agree with it or not, but that's my POV.

I have added helper scripts in addons/tools/ to help make merging between the branches very simple, it should not be such a burden.

> Development should be done only in trunk!

again, perhaps I wrote a bad choice of words in the bug report, I didn't mean development of new features and modules (although unnecessary refactoring or breaking the API/ABI is the critical thing to avoid, the occasional new flag, python wrapper function, or nice gui improvement would be ok IMO), I mean maintenance of the stable branch, with "development" of g6-specific bug fixes and a place to put things while the relbr is frozen. Experiments for fixing WinGrass problems is one good example, parallelization of i.* scripts done in devbr6 now but not meant to go into the relbr until after the 6.4.3 release is another.

(more interested in fixing bugs than having another long discussion which goes nowhere :-)

More information about the grass-dev mailing list