[MERGE][#321066] "bzr log -r 1.1.1..1" dotted revnos in the wrong order raises ValueError
John Arbash Meinel
john at arbash-meinel.com
Mon Jan 26 15:04:46 GMT 2009
-----BEGIN PGP SIGNED MESSAGE-----
Marius Kruger wrote:
> with patch
> 2009/1/25 Marius Kruger <amanic at gmail.com <mailto:amanic at gmail.com>>
> Don't raise a ValueError when the user passes dotted revnos in the
> wrong order to log
> (eg. `bzr log -r 1.1.1..1`), by comparing revision ids in stead
> of revnos.
> This is needed because for some reason we always get a revno=1
> for dotted revnos,
> which was allowed past this validation.
> I added a assertRevnos() to the blackbox test but didn't refactor
> the other tests
> to use it because of all the other log changes in BundleBuddy and I
> didn't want
> to cause unnecessary conflicts. (I can do that when all the other
> log changes have landed)
I think we should trap this at a different point. Specifically, doing
"is_ancestor()" is a very expensive check, without much of a benefit.
Instead, I think we should just trap the location that is causing a
ValueError and give a better exception at that time.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the bazaar