|
SA Bugzilla – Full Text Bug Listing |
Summary: | -t flag adds to auto-whitelist | ||
---|---|---|---|
Product: | Spamassassin | Reporter: | Craig Hughes <craig> |
Component: | spamassassin | Assignee: | Craig Hughes <craig> |
Status: | RESOLVED DUPLICATE | ||
Severity: | enhancement | ||
Priority: | P2 | ||
Version: | 2.0 | ||
Target Milestone: | --- | ||
Hardware: | All | ||
OS: | All | ||
Whiteboard: |
jm's getting ready to trot off around the globe for a while. Reassigning all his unaccepted bugs to craig. Alternatively, have spamassassin -t clean up after itself -- ie do the test, update the whitelist, report, then revert the whitelist to pre-test state. Trouble with doing it this way though is you'll run into concurrency issues though probably.... This will actually be more or less fixed by switching to the new AWL algorithm since it will only modify the mean for that sender somewhat, not really push the sender over some threshold. Still might be worth implementing someday, but for now I'm just going to attach it to bug #23 *** This bug has been marked as a duplicate of 23 *** |
Justin Mason said: > Greg Ward said: > > > I think his suggestion was right on: don't update the auto-whitelist in > > testing mode. > > yeah, I agree. Must implement this... hmm, don't think I'm going to get a chance to ;) Craig/Matt -- the tricky thing here is that 2 of the tests in "t" rely on "-t" updating the auto-whitelist, so they can test it! So I reckon an undocumented command-line arg to do -t while updating the whitelist is the way around that, and just change the def behaviour of "-t". --j.