On 02/16/10 05:00 PM, Shawn Walker wrote:
On 02/16/10 04:53 PM, Danek Duvall wrote:
Shawn Walker wrote:
- line 466: use pkg.fmri.fmri_match()?
- line 471: use pkg.fmri.glob_match()?
For both of these, using the function was much slower than in-lining
the logic. That's an optimisation that was made in the work for
get_pkg_list().
Ugh.
To be clear, I haven't measured the impact of it here in the logic the
client uses after the results have been returned by get_pkg_list(). My
experience is from it contributing (seconds) to get_pkg_list() when I
was writing get_pkg_list().
I'll double check here to see if the difference is as great as it was
for get_pkg_list() (internally).
I've checked; depending on the pkg list operation in question, it's
about 2-8% slower to use the fmri matching functions instead of
in-lining the logic. For pkg list (not with -a, etc.) that's a
half-second at most from what I can tell. But for the -a or -af cases,
it can expand to a few seconds or more.
--
Shawn Walker
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
