Bzr development stopped

[Robert Collins]

On Fri, Nov 30, 2012 at 4:37 AM, Thorsten Glaser <t.glaser at tarent.de> wrote:
> On Wed, 28 Nov 2012, Robert Collins wrote:
>
>>  - parallel imports [files with the same name and content appear
>> different and are hard to merge]
>
> Speaking of that… is there a recommendation how to handle
> them? I’ve had that a few times in the past, where some
> previously local patches got applied upstream and then
> got re-imported with their svn fileid…
>
> TIA,
> //mirabilos


There isn't a good polished answer today. Here are a couple of
approaches that might work.

One is to extend rebase to be able to handle differing file ids,
detect them, and then use it to rebase your thing that is tracking svn
(and that would ripple out through any active branches).

Another is a merge strategy that would deal with conflicting paths by
choosing (consistently) one of the file ids, so that merging A->B or
B->A selects the same winner, and falls back to a 3-way merge along
the revision graph rather than per-file graph.

You can emulate the second one by hand, by deleting and readding the
file using --ids-from to 'bzr add', and applying any local delta by
hand,

-Rob
-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Cloud Services