Hi, concerning bug #24806, here's a patch which changes the initialowner and initialqacontact fields from tinytext to mediumint. Martin

This would be very cool, especially for sites like Mozilla.org where people seem to like to switch their emails once a month. Putting it on 2.12 radar.

Yay!

*** Bug 55245 has been marked as a duplicate of this bug. ***

So waht I did was take the existing patches, applied them to the trunk, fixed the rejects and then re-diffed. The only part I couldn't merge were the changes to importxml.pl, which appears to have dramatically changed. I'm going to review these now. Would appreciate help since this is essentially a schema change and converstion.

ooof. don't use that patch. i screwed up the merge. new patch coming up.

taking ownership, the patch looks good for the most part. testing it now.

adding dave to cc.

installed the patch on my test install... First thing I tried was editcomponents.cgi. It got this: Global symbol "$initialowner" requires explicit package name at /usr/local/home/justdave/public_html/bugzilla/editcomponents.cgi line 396 (#1) Adding the line: my $initialowner = trim($::FORM{initialowner} || ''); immediately before that line seems to fix this. After fixing that, I tried to change the initialOwner on a component to a user that didn't exist. It correctly reported to me that I needed to choose a user that had a bugzilla account, and offered me the chance to go back, but then instead of the footer, I got this: Table 'logincookies' was not locked with LOCK TABLES at globals.pl line 171. Not being an expert on table locks, I'm at a loss on fixing this one.

Changing an initialOwner or initialQAContact to valid addresses seems to work fine.

Strange, this bug is marked as a blocker of a fixed bug?

I have the table locking error as well. We could remove the trailer calls but I'd like to understand why we're getting the table locking error in the first place.

If you apply this you're likely also going to make a small tweak to the sanity checking that was checked in for bug #56822 to get it to check against the user id rather than the user's email address.

Matthew, your patches didn't apply clean. And the table locking problem can be fixed by moving SendSQL("UNLOCK TABLES") before PutTrailer(). So I think we're pretty close to having this be in good enough shape to checkin.

Oh well, too much work for such a small patch, just apply it manually. It should be obvious what I mean and should take about 5 seconds.

patch comitted to trunk. thank you!

Sorry for the spam, but I needed to be able to query for all of these correctly.

Verified this is definitely changed and I can't think of anything else to check that hasn't already been.

Moving closed bugs to Bugzilla product