[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [mdt-ocl.dev] 'archive' branches


what about 349157? The bug is still open, but Ed seemed to have a plan how to fix this that exceeded what I had in mind doing. Shall I close / delete the branch?

I've "archived":

 - 348502
 - 348504

344368 and 349117 I consider important work in progress. I just currently lack the time to continue on them and hope to be able to pick them up early October again.

-- Axel

On 08/29/2011 10:36 AM, Ed Willink wrote:

I've eliminated all the (my) archive branches and pruned my obsolete bug

It looks as if there are a couple that Axel could prune and a few more
that Adolfo could prune.

In EGIT, just delete all remote tracking archive/bug branches and then
fetch from upstream to get a cleaned up display.

I'm finding that neither rebase nor merge works very satisfactorily when
there is a conflict, and I find the pop-up dialog and OURS/THEIRS
bracketing unfathomable when both are MINE, particularly when changes
seem to be whitespace related. I find a manual cherry pick iteration a
much better way to rewrite an out-of-date branch onto local master. Once
complete the out-of-date branch can be reset hard to local master, which
can then be reset hard back to upstream master.



On 26/08/2011 16:54, Adolfo Sánchez-Barbudo Herrera wrote:

+1. Let's see if the SR1 RC2 is smoothly done the next week, so that I
might spend some time sorting out the bugs and branches I have in
hands.... I think that some of them are solved and they could probably
be removed.


El 26/08/2011 9:58, Ed Willink escribió:

I think that we felt that archive branches might be useful for
retaining the history of a particular branch.

This does not appear to be the case. Once the successful branch is
rebased onto master, all the relevant history is 'merged', so there
is no particular use for the archive branch which is just an
end-of-activity marker. The true history is identifiable from the
[xxxxxx] commit comment prefixes and the compressed time recap at the
rebase transaction. If a real marker for the end-of-activity is
required then a tag is of course possible.

It seems that the only 'bug' branches worth keeping are those that
have not been exploited, so I propose that

bug/xxxxxx is a work in progress that will probably be exploited
archive/xxxxxx is a work abandoned, that might be revivable later

There should be very few archive/xxxxxx, and only a few bug/xxxxxx.



_______________________________________________ mdt-ocl.dev mailing list mdt-ocl.dev@xxxxxxxxxxx https://dev.eclipse.org/mailman/listinfo/mdt-ocl.dev