Hi folks,
I know everyone here is very busy working on the 3.1.0 release, but I work on one of the distributions where major version upgrades are not an option. Duncan similarly is in need of a maintenance release for Debian. So I have been testing the 3.0 branch r122144 for many weeks now by pushing it to thousands of our FC4 testers and running it on several production servers. For all intents and purposes the current 3.0 branch is great, it seems to solve the worst RH Bugzilla reported problems from 3.0.1 and 3.0.2 and issues that I personally have observed.
While I believe the current 3.0 branch to be good enough for 3.0.3, for thoroughness I spent some time reading through all SA 3.0.x reports and 3.0.3 milestones. Below is what I propose for checkin to the 3.0 branch. Anything else you would suggest?
======= TRIVIAL STUFF THAT LOOKS SAFE ========== http://bugzilla.spamassassin.org/show_bug.cgi?id=4191 [review] uri_to_domain() is broken for urls with empty port
http://bugzilla.spamassassin.org/show_bug.cgi?id=4232 multipart message with 0 parts -> uninitialized in m// (with patch)
http://bugzilla.spamassassin.org/show_bug.cgi?id=4121 [review] Score for user defined rules become ignored
http://bugzilla.spamassassin.org/show_bug.cgi?id=3944 [review] get_envelope_from not handling received header
======== MAYBE =========== http://bugzilla.spamassassin.org/show_bug.cgi?id=4201 URIBL Tests are looking up email addresses No unanimous agreement whether this is the right thing to do. SA should decide if they want this behavior in both 3.0.3 and 3.1.0 (already committed in 3.1.0).
======== DOC CORRECTION ========== http://bugzilla.spamassassin.org/show_bug.cgi?id=4267 Undocumented obsolete -w switch in spamassassin(1) and usage Already fixed in 3.1.0 according to AJ_ZO Do doc fixes require R-T-C?
The only uncertainty here is... were there other serious issues fixed in trunk but not retargeted for the 3.0.3 milestone? It would take me far longer to dig through all the 3.1.x changes to find the most important stuff to backport, so if you happen to think of any very important thing fixed that probably affects 3.0.x, please let me know. I will personally do the testing to verify it.
I propose reviewing and committing the above changes, and give it a week or two for more testing before release. I will help in testing by pushing this to FC3 and FC4, running it in production on several servers, and running any regression tests recommended by the developers.
Thoughts?
Thanks, Warren Togami [EMAIL PROTECTED]
