[MERGE] Make bzr log FILE faster by looking up text_keys one at a time.

[John Arbash Meinel]

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

Robert Collins wrote:
> Robert Collins has voted comment.
> Status is now: Waiting
> Comment:
> This is counter-intuitive - generally GraphIndex performs better if 1
> query for N keys is made rather than N queries for each of those 1 keys.
> What data set are you working with?
> 

I would actually say that I've seen that as well. When I was playing with
optimizing heads() I did a lot of work to batch the requests together, only to
find it was *slower* than doing them separately.

I'm certainly happier going forward to BTree indexes, because I feel I both
understand the code better, and that they are tuned to avoid some of the GI
issues.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIvImVJdeBCYSNAAMRAgdvAKDI5f41TO4gLxTJi6OpQOXXVHKBBACeJulK
MTIOjKOqImep0kx6nhflxr8=
=rxSQ
-----END PGP SIGNATURE-----