https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6515
--- Comment #5 from Mark Zealey <[email protected]> 2010-11-24 04:47:08 UTC --- > First, I'm a little confused. The spamd time limit options are command line > parameters. Is there documentation somewhere that leads you to believe you > can > specify them in a configuration file other than something like a sysconfig > file > that an init script interprets? Nope; however I see nothing in the docs about the time_limit local.cf setting being ignored when running under spamd. Hence whatever happens out of this thread I'd like to see some clarification in the docs as to how spamd and time_limit relate to each other. > Second, I would argue that amu response, positive, negative, zero, etc. if the > child timeout is reached is invalid OTHER than a response that says the child > timed out. I can see your point of view here, but I believe it should at least be configurable by the end user (which the current setup doesn't allow). The TIME_LIMIT_EXCEEDED synthetic rule is just fine for this - spamassassin can provide a default behaviour and the users can override this how they would like (I assume this is how it works anyway, not looked into the code itself). However the current behaviour is that this rule is never generated under spamd, it simply kills the child and abruptly closes the connection to the server. As I've said before, this (on our systems) leads the clients to believe there has been an error on the server and it will then retry on another server etc. I think the most flexible, logical and easiest change is to allow both time_limit and the --timeout-child option under spamd to work together with the time_limit being a nicer and more configurable way of sending an error status back to the client but the --timeout-child option being a last resort in case of a bug within spamassassin or the regexp engine. Mark -- Configure bugmail: https://issues.apache.org/SpamAssassin/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug.
