Handling repeated text annotations

[John Arbash Meinel]

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

Robert Collins wrote:
| On Wed, 2008-02-13 at 17:07 -0600, John Arbash Meinel wrote:
|
|> I'll try using heads() though I'm not sure it is exposed easily down
|> at this
|> level :(. I'm a little concerned that all the heads calls are going to
|> kill
|> performance. But I guess we only have to do it for lines that seem to
|> claim 2
|> parents.
|
| When you merge a line from another text, it should not get a new
| version, so when you have two branches join, no changes on either side
| will not be a merge.
|
| BUT - if one side changes, then you have two versions - the unchanged
| and the changed.
|
| So yes, heads() is indeed the key to not showing 'merges' where its
| really just one side doing edits.
|
| -Rob

I also want to check the old knit logic. Because we might have used "match all
lines to left, for things that don't match, check right". Which would treat this
case as "left always wins" which is certainly easier to implement.

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHs4lrJdeBCYSNAAMRAuYQAKCtCqEjMtsHbDgMJHmfwTDpjLfefQCgimz8
FFexSDkokJmgSGed4xUkX6Q=
=5yNB
-----END PGP SIGNATURE-----