--=-sr2yz1O+BEl0IMdrmTBm Content-Type: text/plain Content-Transfer-Encoding: quoted-printable > Feature Requests >=20 > Have some option so that yum just looks at it's cache instead of going= to=20 > the server and checking the headers again. > For nightly yum's this isn't a problem, but when you doing a compile (or=20 > something) and you need to do lot's of 'yum list' and/or various 'yum=20 > install', it get's a little slow when each time it goes and checks the=20 > headers. I would gladly put in a extra -C (for cache) so that it just us= es=20 > it's cache, and save the extra time. ok so I implemented this and tested it for speedups. the speed is improved only marginally. from 8s when downloading the header.info file to 5s when using the cached copy. I loaded up the python profile module (thanks jeremy) and did some fun tests. It seems like clientStuff.rpmdbNevralLoad() is taking up a whole second and the cumulative time spent in returnObsoletes is huge. so I could speed things up by only calling returnObsoletes when an upgrade is desired. the time difference is amazing - it goes from 8-9secs for the startup to 3-4secs. It makes a huge difference for the list operations. anyone have an objection or could think of a situation where that would break things? -sv --=-sr2yz1O+BEl0IMdrmTBm Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQA9S21Z1Aj3x2mIbMcRAuqBAJ4kQ+CLbSxIglB7r2pV1CGbuuE1BwCeOuLV 11+3LuIyt9MFoduguZFE6/k= =WtRF -----END PGP SIGNATURE----- --=-sr2yz1O+BEl0IMdrmTBm--
