On 8 September 2010 16:16, Vincent Ladeuil <email address hidden> wrote:
>>>>>> Martin Pool <email address hidden> writes:
>
> > I think moving them into a top-level directory is fine. I'd like
> > the patch for this to abstract the handling of them so that we can
> > later add the others.
>
> What do you mean by others ?
Other strategies for replacing ignored or unversioned files. istm
users may want to configure that they should be just deleted, cause an
error, be saved to a trash directory, etc.
> >> 0 is not appropriate as the problem may occur during a pull or a
> >> switch (the infamous .pyc problem).
>
> > I think it's appropriate [to give you conflicts on the deletion]:
> > you can already get other kinds of conflicts on pull or switch.
>
> Oh, I don't dispute that, I was talking about breaking the natural flow
> by creating what could be considered as spruious conflicts, if there are
> othe conflicts, they will still be detected.
+1
>
> > The main thing we have to do there is make sure deleting the
> > directory is treated as implicit resolution, because it is not at
> > present.
>
> Right, I consider that to be a different bug (see bug #344013, bug
> #138803, bug #287979, etc). If you think I've high jacked this one I can
> create a new one.
No, that's fine. It would be good to fix that too.
On 8 September 2010 16:16, Vincent Ladeuil <email address hidden> wrote:
>>>>>> Martin Pool <email address hidden> writes:
>
> > I think moving them into a top-level directory is fine. I'd like
> > the patch for this to abstract the handling of them so that we can
> > later add the others.
>
> What do you mean by others ?
Other strategies for replacing ignored or unversioned files. istm
users may want to configure that they should be just deleted, cause an
error, be saved to a trash directory, etc.
> >> 0 is not appropriate as the problem may occur during a pull or a
> >> switch (the infamous .pyc problem).
>
> > I think it's appropriate [to give you conflicts on the deletion]:
> > you can already get other kinds of conflicts on pull or switch.
>
> Oh, I don't dispute that, I was talking about breaking the natural flow
> by creating what could be considered as spruious conflicts, if there are
> othe conflicts, they will still be detected.
+1
>
> > The main thing we have to do there is make sure deleting the
> > directory is treated as implicit resolution, because it is not at
> > present.
>
> Right, I consider that to be a different bug (see bug #344013, bug
> #138803, bug #287979, etc). If you think I've high jacked this one I can
> create a new one.
No, that's fine. It would be good to fix that too.
--
Martin