Re: Mail delays (was: Sending the beta4 release announcement, don't approve ...)
От | Tom Lane |
---|---|
Тема | Re: Mail delays (was: Sending the beta4 release announcement, don't approve ...) |
Дата | |
Msg-id | 20195.1280937654@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: Mail delays (was: Sending the beta4 release announcement, don't approve ...) (Alvaro Herrera <alvherre@commandprompt.com>) |
Ответы |
Re: Mail delays (was: Sending the beta4 release announcement, don't approve ...)
|
Список | pgsql-www |
Alvaro Herrera <alvherre@commandprompt.com> writes: > Weren't we supposed to pass through unmoderated all emails coming from > addresses subscribed to any of the lists? No, that'd make it very easy for spammers, not to mention destroy the main usefulness of the anti "unsubscribe" filters. My recollection is that the way that that bit actually works is that if you're subscribed to any of the lists, you're allowed to post to any other list as if you were a subscriber; we did that mainly to avoid delaying cross-posted mails. Presumably all these mails got caught by one or another of the moderation filter rules, but I didn't look to try to guess which, except for Simon's commit message: Log Message: ----------- Reduce lock levels of CREATE TRIGGER and some ALTER TABLE, CREATE RULE actions. Avoid hard-coding lockmode used for many altering DDL commands, allowing easier future changes of lock levels. Implementation of initial analysis on DDL sub-commands, so that many lock levels are now at ShareUpdateExclusiveLock or ^^^ I suppose this got trapped by the anti-subscribe-command filter. Could we possibly narrow the command filters down so they're only applied to very short messages, like one or two body lines? I've had stuff delayed like this from those filters before, and it's seriously annoying, particularly for commit messages. regards, tom lane
В списке pgsql-www по дате отправления: