Last modified: 2008-09-13 12:35:27 UTC
When a blocked user uses [[Special:Emailuser]] to email another user, the autoblock is triggered, even if they have not been blocked from sending email. This happens both on enwiki and on a vanilla install from SVN.
It looks like a user's autoblock was being reset to 24 hours as a side effect every time they performed something where the software checked for a block, including when isBlockedFromEmailuser() was called. (Rather confusing and unpredictable, and possibly causing other users to suspect the blockee is sockpuppeting.) I changed this in r40771 so autoblocks will instead be spread every time a user begins their session.
Don't we want to spread blocks as often as possible?
I can't think of any benefit we gain by spreading blocks more than once per session. The user's IP is rarely going to change within the same session, in most cases. The one case it _will_ change is if they deliberately reset their IP within a short timeframe, with the same browser window still open. However, if they're still logged in at this point, they'll still be blocked directly; while their new IP will not be caught in the autoblock, the same would have happened under both the old and new approaches if they had simply logged out or cleared their cookies when switching IPs, which presumably anyone savvy enough to reset their IP would be doing in the first place. :)
(In reply to comment #3) > I can't think of any benefit we gain by spreading blocks more than once per > session. The user's IP is rarely going to change within the same session, in > most cases. The one case it _will_ change is if they deliberately reset their > IP within a short timeframe, with the same browser window still open. However, > if they're still logged in at this point, they'll still be blocked directly; > while their new IP will not be caught in the autoblock, the same would have > happened under both the old and new approaches if they had simply logged out or > cleared their cookies when switching IPs, which presumably anyone savvy enough > to reset their IP would be doing in the first place. :) > Well, why did this bug get filed, then? :)