[merge] Use utf8 file-ids

[John Arbash Meinel]

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

This attached patch changes the xml serializer and the appropriate
interfaces to use utf8 file ids, rather than passing around Unicode file
ids.

The biggest gains from this won't be until dirstate lands, but once
there, we will be able to avoid lots of encode/decode calls while
processing the inventory.

This includes my earlier patches for making sure
fileids_affected_by_revision_ids can handle unicode data. I was a little
surprised to see that we pass file ids around quite a bit less than we
pass around revision ids. Thinking about it, I probably shouldn't have been.

I'd like to see this land so that we can get dirstate optimized a bit
more, but I'm happy to address any concerns people might have.

At this point, I'm not sure whether we should be using utf8 for paths or
not. But both file_ids and revision_ids are "arbitrary" in the sense
that it isn't really user visible data, so it is pretty clear that we
can define them however we want.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFF154kJdeBCYSNAAMRAn6FAJwP7nPG4jQUWBlvOYK9W+C6ZHKZ0gCgzjuh
DqNpiHwTsLZdykeJGjw4YJc=
=usCV
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: utf8_file_ids.patch
Type: text/x-patch
Size: 95721 bytes
Desc: not available
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20070217/e6e19d01/attachment-0001.bin