[MERGE] Remove the ReopsitoryKnit3 class

Martin Pool mbp at canonical.com
Mon Oct 8 05:49:40 BST 2007


Martin Pool has voted tweak.
Status is now: Conditionally approved
Comment:


+    There is one and only one format for every disk format. The actual
+    repository types do not indicate disk format at all - only 
repo._format can
+    be used to determine the disk format of a Repository instance.
+

'actual repository types' seems ambiguous.  It would be clearer to say

   There is one and only one Format subclass for each on-disk format.
   But there can be one Repository subclass that is used for several
   different formats.  repository._format can be used to determine etc.

Incidentally maybe _format should be public, but renaming it is not a
priority.


+    # make an manually, or incorrectly initialised KnitRepository 
object
+    # invalid

I don't understand the comment; these are set from mandatory constructor 
parameters.  I think it's just out of date.


@@ -388,7 +389,7 @@
          fid_bar1 = u'b\xe5r-01'.encode('utf8')
          fid_sub = u's\xb5bdir-01'.encode('utf8')
          fid_bar2 = u'b\xe5r-02'.encode('utf8')
-        expected = [(u'', fid_root, None, None),
+        expected = [(u'', fid_root, None, rev_id_2),
                      (u'b\xe5r', fid_bar1, fid_root, rev_id_1),
                      (u's\xb5bdir', fid_sub, fid_root, rev_id_1),
                      (u's\xb5bdir/b\xe5r', fid_bar2, fid_sub, rev_id_2),

This is in the tests for xml5.  I can see xml5 can include a
root revision id if there is one in the input, but I don't see
why this test is changing.

Otherwise ok.


For details, see: 
http://bundlebuggy.aaronbentley.com/request/%3C1191552370.13959.3.camel%40lifeless-64%3E



More information about the bazaar mailing list