Re: release 3.4.x

[Jaroslav Hajek]

On Wed, Sep 8, 2010 at 2:19 PM, Ben Abbott <address@hidden> wrote:
> On Sep 8, 2010, at 7:49 AM, Jaroslav Hajek wrote:
>
>> On Wed, Sep 8, 2010 at 1:41 PM, Ben Abbott <address@hidden> wrote:
>>> On Sep 8, 2010, at 5:18 AM, David Bateman wrote:
>>>
>>>> Jaroslav Hajek-2 wrote:
>>>>>
>>>>> hi all,
>>>>>
>>>>> what about forking the 3-4-x branch now, and start making release
>>>>> candidates?
>>>>> Thomas, would you be willing to host 3.4.x on hg.tw-math.de?
>>>>>
>>>>
>>>> Shouldn't the fltk legends be included before the fork?
>>>>
>>>> D.
>>>
>>> I'd also like to see the new legends included. For that I'll need to back 
>>> out the outerposition code I recently added to the gnuplot backend.
>>>
>>> Also, the colorbar isn't currently working. I'll fix that today.
>>>
>>> Ben
>>
>> OK. I'm afraid I won't be able to do the release, then. If there's a
>> volunteer for release management, I can share a couple of useful
>> recipes.
>>
>> --
>> RNDr. Jaroslav Hajek, PhD
>> computing expert & GNU Octave developer
>> Aeronautical Research and Test Institute (VZLU)
>> Prague, Czech Republic
>> url: www.highegg.matfyz.cz
>
>
> It was a bit of a hassle last time, but should be branch and then do parallel 
> commits when needed?
>

Our "branches" are in fact separate repos. Essentially, after the
branch is forked, changesets can be simply pulled as long as we need
all of them (most changesets currently arriving are bug fixes). After
a diversion occurs, hg transplant needs to be used. What I used to do
is producing a sequence of candidate releases using "make dist"
eventually converging to the actual release.

If anyone wants to propose a new method, you're welcome, but be ready
to actually carry it out.

-- 
RNDr. Jaroslav Hajek, PhD
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz