|
SA Bugzilla – Full Text Bug Listing |
Summary: | Fix perceptron.c to allow for run-time read of logs | ||
---|---|---|---|
Product: | Spamassassin | Reporter: | Duncan Findlay <duncf> |
Component: | Score Generation | Assignee: | SpamAssassin Developer Mailing List <dev> |
Status: | RESOLVED WORKSFORME | ||
Severity: | normal | CC: | apache |
Priority: | P5 | ||
Version: | SVN Trunk (Latest Devel Version) | ||
Target Milestone: | Future | ||
Hardware: | All | ||
OS: | All | ||
Whiteboard: | |||
Bug Depends on: | |||
Bug Blocks: | 2853 |
Description
Duncan Findlay
2004-05-13 20:09:43 UTC
Needed to go with the new masses/ stuff I propose that this (and possibly 2853) get moved to 3.1 unless the code is ready to go now. I really don't think these are a 3.0 blocker, it'd just be nice to have. it'd be really really nice to have. It'll have major effect on the usability of 3.0.0 after 3 months or so. I don't think it's a really tough change, but I agree we're running a little late. Perhaps we could comprimise and move it to 3.0.1? Subject: Re: Fix perceptron.c to allow for run-time read of logs On Sat, May 22, 2004 at 02:21:07PM -0700, bugzilla-daemon@bugzilla.spamassassin.org wrote: > it'd be really really nice to have. It'll have major effect on the usability of > 3.0.0 after 3 months or so. ??? How so? The perceptron (and everything in masses) is, by in large, for developers only. There's nothing in there which is required for 3.0 to run, nor is it even installed by default. So I don't see how there'll be an effect on usability. > Perhaps we could comprimise and move it to 3.0.1? I don't think the masses directory matters for a release, so don't have a problem making changes to it through a "stable" release cycle. My issue for right now is that we're taking forever to get 3.0 out, and I'm really tempted to say "screw all the open tickets". Subject: Re: Fix perceptron.c to allow for run-time read of logs On Sat, May 22, 2004 at 03:17:36PM -0700, bugzilla-daemon@bugzilla.spamassassin.org wrote: > ??? How so? The perceptron (and everything in masses) is, by in large, > for developers only. There's nothing in there which is required for > 3.0 to run, nor is it even installed by default. So I don't see how > there'll be an effect on usability. We have always known that the effectiveness of SpamAssassin deteriorates over time between releases. If users are able to generate their own scores, SpamAssassin will continue to give them very good hit rates. In some ways, we're kinda like Debian. We make large releases fairly infrequently, and users are forced to resort to third parties in order to meet their needs (for Debian, this is third party backports of Debian packages). If we distribute scripts that make it easy to make your own scores, people can do this when their hit rates start deteriorating and it extends the useful life of SpamAssassin. My "3 month" figure is a bit of an exaggeration. Subject: Re: Fix perceptron.c to allow for run-time read of logs On Sat, May 22, 2004 at 03:47:19PM -0700, bugzilla-daemon@bugzilla.spamassassin.org wrote: > Debian packages). If we distribute scripts that make it easy to make > your own scores, people can do this when their hit rates start > deteriorating and it extends the useful life of SpamAssassin. Ok, so I think we agree that this doesn't need to hold up the 3.0 release, but it'd be nice to get done. Anyone have an issue if we punted the ticket to the 3.1 queue? If the code is ready earlier, we can consider it for a 3.0.x release... (imo, the masses directory is available for enhancement during a "stable" release cycle, since any form of score generation is technically "development"...) Subject: Re: Fix perceptron.c to allow for run-time read of logs
> Ok, so I think we agree that this doesn't need to hold up the 3.0 release,
> but it'd be nice to get done.
Fair enough. I'll agree to that.
ok, moving to 3.1 queue for now. move bug to Future milestone (previously set to Future -- I hope) Closing ancient stale bug. |