Just a bit cleaner
I need to test on PostgreSQL now so yay!
The most likely reason is that we already did it. We could store the $obj in the cache instead of just the $new_id but that would use a LOT more memory for little gain.
The "normal" accessors don't actually take arguments, they just mean Me and Now
Creator and Created get set when we import the ticket (yay!)
I was hoping it would keep LastUpdated but it doesn't seem to work. Something to look at.
No idea why it was there in the first place but I don't want it.
Safer, the values we probably need.
I need the user mappings to make things work properly.
I had a problem, this fixed it.
I want to re-use bits so to make it easy, just do it.
They are probably wrong
It was really useful for me, so may help someone else.
I was going to have to modify several of them, so rather than that, refactor!
it seems that is what RT likes.
Just make it a bit tidier
Allows me to put some things in a common routine later. nitpick my ($user) = @_ instead of my $user = shift;
because we need all the old tickets and transactions imported before we can create the links, we need to queue them until the end.
This one actually runs later so it ended up more like the original.
Same as the rest, just different.
Much fixups to make moar betterer
Just handle the "special" stuff here.
Gets things ready and then calls the proper import routine. Also handles retrying if we had an id collision. Plus accessor methods for each type.
Looks up the new QueueID from the name, defaulting to General if there isn't one that matches.
So we can lookup what to do with them rather than large if/else structures
Read tickets from a directory of xml instead of a single file. Handles what we need it to better than what we had.
Because now it should. This reverts commit 2cf673e. Conflicts: rt-ticket-importer
Makes the logic easier to follow
Should help keep EmailRecord transactions happy