: Hmm @hoss did you check this manually, I mean the revision? ironically, i noticed because i initially tried running the 4x version of the smoke tester again the RC, and it complained about the duplicated changes sections. At which point i starting thinking that it was my fault for using the 4x smoke tester instead of the 4_3 smoke tester, because clearly the changes.txt fix to go along with the smoke tester fix probably only made it to the 4_x branch -- then i remembered *I* was the one that fixed CHANGES.txt, yesterday, and i knew for a fact i had put it on the 4_3 branch.
then i had that entire conversation over again in my head because it still didn't make any sense to me. then i looked at the revision number. : I think our smoketester should check the revision of the release branch and : the revision of the RC to make sure this doesn't happen? 1) i'm not sure how it would do that, it doesn't currently know anything about branches does it? 2) it would make it really hard for people to "keep testing" an RC${X} to look for more problems after it was already decided there should be an RC{$X+1} and someone had commited something for that. If we don't do so already though, it would definitely be good to have the release scripts force an "svn update" though in case the RM ddin't realize their working dir was stale (i realize that's not what happened in this case, but still -- good idea.) -Hoss --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org