dolfin team mailing list archive

["Garth N. Wells" <g.n.wells@xxxxxxxxxx>]

Johan Jansson wrote:
> On Thu, Oct 26, 2006 at 03:12:02PM +0200, Garth N. Wells wrote:
>> Johan Jansson wrote:
>>> On Thu, Oct 26, 2006 at 02:39:48PM +0200, Garth N. Wells wrote:
>>>> Repositories are so easy to make, so just make some local ones to share
>>>> developments against the stable version.
>>> Ok, but then we're in agreement, since this is exactly what dolfin-kth
>>> is. 
>>
>>> I think it's a good idea though to cc notifications to dolfin-dev,
>>> because there might be changes in local repositories  which are not
>>> only module-specific, 
>> Shouldn't these changes be made in DOLFIN?
>>
> ...
> 
>> I looked at the log for dolfin-kth, and couldn't see why most of the
>> work (which is by Johan J.) isn't done directly in DOLFIN?
>>
> 
> It seems we're back to square one :). You said this:
> 
> "They work against the latest release. That is the stable version :)."
> 
> Our stable version is dolfin-kth. 

We do our best to work against an official DOLFIN release. Perhaps a
part of the issue is the prominence that has been given to the
dolfin-kth repository gives the impression of a fork.

You don't point your
> students/collaborators to the DOLFIN development repository, and
> neither do I. 

Yes, I do at times :). If there is an important addition/change in the
development version and it is stable enough to use, I suggest that they
update to the development version.

If I'm working with people who work against dolfin-kth,
> then I need to push changes they need there. Eventually they will be
> merged to the development repository. Since most of the development is
> at the module level, a merge should not involve much work.
>



> It seems the argumentation is inconsistent. I'm getting criticized for
> not working against the main repository, when in fact nobody is doing
> that with their locally-run projects either (with other non-core
> developers involved). 

I wouldn't say anybody is criticizing anybody. We're just discussing the
best way to negotiate our way around a difficult development issue.

Surely you are then aware of the impracticality
> in doing so? Your plasticity project sounds very interesting, why
> isn't it developed in the main repository? 

Simple. The code is not yet up to the standard expected in DOLFIN. When
it is, it will be added.

Why isn't Dag developing in
> the main repository? The reason is probably the same as mine.
> 
>> I agree. What about more frequent releases then? Would that help? If the
>> latest release needs to be patched to get something working, let's make
>> a release.
>>
> 
> Either that, or a stable branch, which is essentially the latest
> release, with important fixes backported. This is how Linux kernel
> development has worked, and also Debian as far as I know.
>

This sounds good to me.

Garth


>   Johan
>