[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Proposal: move win32-loader in SVN repository



Christian Perrier <bubulle@debian.org> writes:

> Quoting Otavio Salvador (otavio@debian.org):
>
>> I'm still of opinion to choose another dir for them. I've proposed[1]
>
> Depends if we have many of them or not. All propositions you made were
> indeed OK for me.

We already have:

win32-loader
busybox
mklibs
libdebian-installer (if we opt to move it too)
<others>

>> few alternatives for directory names but noone commented on them; the
>> move has been done without a widely agreement from the team and people
>> that were involved discussing on the thread didn't agree too (me
>> included).
>
> Well, that doesn't really makes a problem for me. Action was needed,
> Frans did the action...and even enforced the discussion by doing
> so..:-)...What would have happened if he hadn't? Very probably a dying
> discussion, all of us going back to our various tasks and nothing
> made.

Sorry I disagree. We always preferred to hold changes until we get
agreement. Many times it has been complained here by mistakes made by
me, or others, and this is really serious when it has SCM metadata
involved since it does complicate a future migration so it need to be
well decided.

> I'm all for action when a topic emerges (this is even something I
> debated with Frans sometimes because I'm often even more for immediate
> and quick action than he is...).

Yes, sometimes it is really important and I agree that we shouldn't
hold every action however a package move on SCM won't (and it didn't)
change our lifes and neither solve any urgent trouble we were having
so no reason to hurry with it.

>> It can be moved again, for sure, but as Joey has already spot this is
>> another thing to worry when migrating from svn to <put any other scm here>.
>> 
>> This enforces the need of those svn moves being betther thought to
>> avoid double work.
>
>
> I don't really understand why multiples moves are a problem for a
> future SCM change but, well, I'm mostly ignorant when it comes at
> these complicated things.... 
>
> About lost time: a "svn mv" is a matter of seconds, so we shouldn't
> worry that hard, I think.
>
> I still favor the move.

I'm favor of the move but I think we should agree to _where_ it should
be done.

About the migration in SCM changes, it does matter.

When we're migrating from SVN to another SCM we'll need to do partial
migrations for all previous places the source where and do a move add
another point where we'll need another migration.

Example:

r100:
  /packages/foo
r101
  /trunk/foo

If we were migrating foo, the source, we'd need to make a migration
from r1 to r100 using /packages/foo, get the result and then do
another migration using r101 /trunk/foo to allow it to continue to
grab the code history.

Now think about doing it for all moves since CVS, source renames, and
wrong moves .. this can be really boring.

-- 
        O T A V I O    S A L V A D O R
---------------------------------------------
 E-mail: otavio@debian.org      UIN: 5906116
 GNU/Linux User: 239058     GPG ID: 49A5F855
 Home Page: http://otavio.ossystems.com.br
---------------------------------------------
"Microsoft sells you Windows ... Linux gives
 you the whole house."


Reply to: