Hi, David Lang wrote: >> Another change in 1.1.2 that may bite people is that it introduces the >> capability for rules to be multiline by allowing newlines inside a tag >> definition. If someone has a bad config (uneven number of % on a line), the >> newline will no longer terminate that bad rule, all rules follwoing it will >> not be parsed (until it hits the next line with an uneven number of % on >> the line) >> >> [...] >> >> Personally, I would probably concentrate on testing liblognorm-2.0 >> and the upgrade process to it rather than spending much effort on >> 1.1.2
Well, liblognorm-1.1.1 to 1.1.2 is a *minor* bump according to the version number. This shouldn't require any test to be rewritten or break any software which was working with liblognorm-1.1.1. If it does, it isn't a minor bump and the version number/SONAME is wrong ;) We bumped liblognorm because we thought it is just a bugfix release... Regarding your first paragraph: You said "has a bad config"... if that mean that the user wrote something which was violating specs but was working due to a bug and now because the bug is fixed in v1.1.2 it will break like it should from the beginning, that's ok. If it is a changed behavior, it isn't OK. I created a PR (https://github.com/rsyslog/rsyslog/pull/492) which enables mmnormalize in Travis to keep track. -Thomas _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE THAT.
