On 18-03-13 16:30:16, Antonio Terceiro wrote:
> On Sun, Mar 04, 2018 at 10:04:46PM +0100, Georg Faerber wrote:
> > On 18-03-03 16:30:36, Cédric Boutillier wrote:
> > > ruby-factory-girl-rails fails to build because of dependency checking
> > > looking for a factory_girl gemspec file instead of a factory_bot.
> > >
> > > I don't know if it is better to fix factory-girl-rails gemspecs, or to
> > > ship also the factor_girl.gemspec file too with ruby-factory-bot.
> >
> > Hm, I'm not really sure how to deal with this, as I already (tried?) to
> > describe in one of my initial mails: Not only to name changed, but also
> > code within, and no "compat helper / level" was introduced. This means,
> > that code which depends on factory_girl needs to change as well.
> > Upstream described the necessary steps regarding this [1].
> >
> > Given these circumstances, I'm not really sure if using 'Replaces' in
> > d/control and providing a transitional package is the right way to go.
> >
> > What about introducing ruby-factory-bot as a new package without
> > 'Breaks', 'Replaces' and a transitional package, and removing
> > ruby-factory-girl once all upstreams of the reverse dependencies have
> > changed to ruby-factory-bot?
>
> If the new package is not backwards compatible, it should not have
> breaks/replaces or anything like that. you need to make the old package
> deprecated, report bugs against reverse dependencies, and wait for them
> to switch to the new library.
Alright -- so I'll create an ITP and package ruby-factory-bot
accordingly? After this, both ruby-factory-{bot,girl} will exist in the
archive then, but I guess this is okay? Reverse dependencies are then
able to change their dependency anytime they wish.
Thanks,
Georg
Attachment:
signature.asc
Description: Digital signature