[MERGE] Comparison cache to speed up diff

[Aaron Bentley]

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

John Arbash Meinel wrote:

> I thought your patch meant that we didn't have to extract the basis
> text, since we already know the diff of the current file relative to its
> basis revision. Or am I misunderstanding what you have done?

No, it doesn't avoid extraction.  I believe it's reasonable to expect
really fast extraction of the texts we care about.  Maybe if we don't
achieve that, we can re-examine caching basis texts.  But fast extract
time is so important to so many aspects of performance that we really
need to try our best to accomplish it.

What I've done is to cache the get_matching_blocks output, which is not
enough data to display a unified diff with.  You need the basis text to
display the contents of removed lines, but I'm not caching the actual text.

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

iD8DBQFGoTgY0F+nu1YWqI0RAtLcAJ4yytHDEtWJtDGiBpjWiuRAyj7CVwCfWGlU
XAH7WM31AuNCCB+g9fpCxgU=
=Zxnm
-----END PGP SIGNATURE-----