[Bug 405264] Re: MYSQL/BZR P3: "bzr log" reveals revisions with duplicate parents

[John Arbash Meinel]

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

GuilhemBichot wrote:
> Could you please provide feedback on my request (2) in "GuilhemBichot  wrote on 2009-08-04: " :
> <quote>given that "bzr merge" already refuses to work (unless --force) if there are uncommitted file changes, could it also refuse to work (unless --force) if there are uncommitted revision history changes? To me it would be logical, and would catch the empty merge case.</quote>
> I'd like to know if this will be implemented, and in what time frame... Thanks!
> 

That was a bit hard to read.... :)

Anyway, I created a test case to reproduce what Guilhem is talking about:

bzr init t1
echo foo > t1/foo
bzr add t1/foo
bzr commit t1 -m 'init'
bzr branch t1 t2
bzr branch t1 t3
bzr commit --unchanged -m "noop2" t2
bzr commit --unchanged -m "noop3" t3
cd t1
bzr merge ../t2
bzr st
bzr merge ../t3
bzr st

You can see at the end of 'bzr merge ../t2' that there is a pending
merge, but no file changes. And the merge of t3 succeeds without
warning, and at the end there are 2 pending merges at the end.

I was surprised that the 'is tree clean' check doesn't include a check
for pending parents. So I agree that we could add that support.

We probably should propose it on the list, to make sure people are okay
with it, since it will be a UI level change.

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkqf5f8ACgkQJdeBCYSNAAPzvQCfUZJFlJ4CbDY7DrU+fHSme46/
TW0AoLdIGf0cxJL5xx2IVv/f88AR2Dbz
=BVdq
-----END PGP SIGNATURE-----