[MERGE] [BUG #172657][1.0] make 'bzr status' after merge faster

John Arbash Meinel john at arbash-meinel.com
Wed Dec 5 20:23:40 GMT 2007


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Aaron Bentley wrote:
...

> Consider this graph (oldest nodes at the top)
> 
> A
> |\
> B |
> | |
> C |
> |\|
> D F
> 
> If we look at the traversal from D and F:
> 1: D, F (no common ancestors)
> 2: C, A (C common, A possibly unique)
> 3: B (search in A terminates.  All searches terminate because all
> remaining active searchers are searching through common ancestors.)
> 
> This will lead to A appearing to be a unique ancestor of F.
> 
> Aaron

As a small aside, because of how the common ancestor walker works, the
find_difference() code actually passes this test. Because the searchers
continue on to the 'null:' revision.
Anyway, if you extend it by 1 more node to:

A
|\
B |
| |
C |
| |
D |
|\|
E F

Then it fails as you described. Though I guess in fairness, your A could
have been 'null:', as you certainly didn't have it as a node in your graph.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHVwjMJdeBCYSNAAMRAj9WAJ96CyE67UZONcSSMrqqL7+20yLtPwCggdGu
Hq93hL3p6O9HrdTYvyXQIyc=
=1KB7
-----END PGP SIGNATURE-----



More information about the bazaar mailing list