Re: thinking twice about the new issue tracker

[Federico Bruni]

Il giorno mer 2 set 2015 alle 9:33, David Kastrup <address@hidden> ha 
scritto:
> I think that on the "evaluate other alternatives" angle it would 
> likely
> make more sense to go through with getting Allura up and running.

As we are so close to get Allura up and running, of course it makes 
sense.
But I was trying to make a different point: there's no migration tool 
from Allura to some other tracker. In case we realize after a month or 
so that Allura was not a good choice and we'd better use something 
else, we'll be forced to write our own migration script.
On the other hand, if we take some time to make the move and use a 
temporary solution (such as Github or Bitbucket) while testing other 
applications (such as Gogs), an import from those services is more 
likely to be available.

> >  Not Free services, I know, but at least their business model seems
> >  honest.
>
> For a project of our size, the business rates of those services are 
> not
> exactly cheap.

I won't reply to your joke. You know what I meant.

> >  Regarding the alternatives, I found Gogs, which seems really
> >  promising: http://gogs.io/ https://github.com/gogits/gogs/
> >
> >  The user interface is very similar to github. See for example:
> >  https://try.gogs.io/galaxycui28/init.el/issues/1
> >
> >  The problem is that migrating from Github/Bitbucket is not currently
> >  possible: https://github.com/gogits/gogs/issues/809
> >
> >  What do you think?
>
> At the current point of time, we aren't even on Github/Bitbucket.  At
> any rate, regarding "promising" solutions there is also the SFC-hosted
> Kallithea <URL:https://kallithea-scm.org/> where we likely would get
> more than marginal interest of the FSF if we got it working for us.

Unfortunately, it does not support issue tracking yet:
https://bitbucket.org/conservancy/kallithea/issues/129/issue-tracking
> The main problem I see is that questions like the usefulness of the
> documentation, scalability under duress, helpfulness of community,
> design of the codebase don't really gain a reliable answer before 
> taking
> a full plunge.

True, that's why the test at Sourceforge is valuable. But we are 
already adding new issues there, which means that the final decision is 
made. Too quick decision IMO.