Rev 2151: Correctly decode utf8 revision ids from knits when parsing, fixes a regression where a unicode revision id is stored correctly, but then indexed by the utf8 value on the next invocation of bzr, rather than the unicode value. in file:///home/robertc/source/baz/knit-index/

[Robert Collins]

------------------------------------------------------------
revno: 2151
revision-id: robertc at robertcollins.net-20061128010415-b004ee4bfebfad9c
parent: pqm at pqm.ubuntu.com-20061126153508-40a2ce4d019aefea
committer: Robert Collins <robertc at robertcollins.net>
branch nick: knit-index
timestamp: Tue 2006-11-28 12:04:15 +1100
message:
  Correctly decode utf8 revision ids from knits when parsing, fixes a regression where a unicode revision id is stored correctly, but then indexed by the utf8 value on the next invocation of bzr, rather than the unicode value.
modified:
  bzrlib/knit.py                 knit.py-20051212171256-f056ac8f0fbe1bd9
  bzrlib/tests/test_versionedfile.py test_versionedfile.py-20060222045249-db45c9ed14a1c2e5