After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 373523 - thread scoring is too slow
thread scoring is too slow
Status: RESOLVED FIXED
Product: Pan
Classification: Other
Component: general
pre-1.0 betas
Other All
: Normal normal
: 1.0
Assigned To: Charles Kerr
Pan QA Team
Depends on:
Blocks:
 
 
Reported: 2006-11-10 19:58 UTC by Carl Roth
Modified: 2006-11-16 21:23 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Compressed score file (186.98 KB, application/octet-stream)
2006-11-10 20:31 UTC, Carl Roth
  Details
0.119 patch (4.72 KB, patch)
2006-11-11 02:16 UTC, Charles Kerr
none Details | Review

Description Carl Roth 2006-11-10 19:58:19 UTC
Please describe the problem:
The "watch thread" and "ignore thread" functions are too slow.  If the score file starts to get large (mine has 6500+ scores, built up over a month) and the given newsgroup gets more than a few hundred articles, then each "ignore thread" operation takes 5-10s to complete.  This is on a 2GHz dual-core P4.

A workaround that I've been using is to periodically delete any extra headers whose score is below the "ignore" threshold.  This is a hack, and in any case, Pan should have an option to delete (expire) low-scoring headers automatically.


Steps to reproduce:
1. 
2. 
3. 


Actual results:


Expected results:


Does this happen every time?


Other information:
Comment 1 Charles Kerr 2006-11-10 20:20:00 UTC
Could you attach your scorefile to this ticket so that I can run Pan under a profiler with your scorefile?
Comment 2 Carl Roth 2006-11-10 20:31:37 UTC
Created attachment 76346 [details]
Compressed score file
Comment 3 Charles Kerr 2006-11-11 02:16:24 UTC
Created attachment 76365 [details] [review]
0.119 patch

Faster scorefile parsing.
Comment 4 Charles Kerr 2006-11-12 20:31:10 UTC
Carl, could you give this patch a try?
Comment 5 Carl Roth 2006-11-12 20:42:40 UTC
I did.  I've been running with the new patch for a few days now; I wanted to wait for my header cache to grow to a more realistic size before I declared it a success.

I'm noticing that the behavior is faster, with a few points:

 1. Adding the first score to a newly-entered group takes a few seconds
 2. Each successive score operation is quick (less than 1s)