Re: [STUMP] Patch: improvement for scripting frame splits

[Julian Stecklina]

Carlos Konstanski <address@hidden> writes:

> On Sat, 26 Sep 2009, Shawn Betts wrote:
>
>> Date: Sat, 26 Sep 2009 00:24:05 +0300
>> From: Shawn Betts <address@hidden>
>> To: StumpWM-devel <address@hidden>
>> Subject: Re: [STUMP] Patch: improvement for scripting frame splits
>>
>> On Fri, Sep 25, 2009 at 5:46 PM, Vitaly Mayatskikh
>> <address@hidden> wrote:
>>> At Fri, 25 Sep 2009 02:01:57 +0400, Michael Raskin wrote:
>>>
>>>> Maybe the actual steering could be in form of split authority. I.e.
>>>> while you can always act as a benevolent dictator, you could give a few
>>>> of us commit rights simultaneously with the idea that each patch not
>>>> picked up by you should be commented on by at least two of us before
>>>> being committed.
>>>
>>> Sounds reasonable. Also it'd be good to commit fresh patches in some
>>> kind of development branch, moving proven to be stable patches to main
>>> branch on monthly basis as suggested by Julian.
>>
>> Regular releases. stable and unstable branches. Geez, it sounds like a
>> *real* software project :). Let's make this happen!
>>
>> -Shawn
>
> Careful.  The Stumper has been the most stable and predictable project
> I have ever had the pleasure of using.  I have never checked out a
> broken build, ever.  Whatever you have been doing, it works great.

I don't think that having a separate development and "stable" tree is
really the way to go, given our limited development resources. What I
suggest would be the following:

1. If a release is imminent, create a release branch, while normal
development continues on HEAD.

2. People can test the release branch for a week.

3. If no regressions are found, make a tarball and release it.

As we seem to have a number of people who just use git HEAD, I fear
that there are few people left that will actually test the release
candidate. But perhaps some of them are going switch to using official
releases.

Another way would be to do it like SBCL: Just freeze the development
branch for some days and release it.

Any comments?

Regards,
-- 
Julian Stecklina

The day Microsoft makes something that doesn't suck is probably the day
they start making vacuum cleaners - Ernst Jan Plugge

pgpgJmmSiIa0c.pgp
Description: PGP signature