While writing it I came to a pretty big problem that's still bugging me, because right now the extension is quite unstable:
There seems to be no "normal" way to read/write a tb message using script. It is easy enough to get the headers (altough I have a bug with tb under linux there). But from then to get f.e. the message content as String is a pain. I actually have to get the offline stream and parse the WHOLE folder for my msg (in my case that was exactly what I wanted, so I didnt care for the "normal" way to get each message).
I am also really unhappy with the way I handle the rewriting of messages. Currently I create a temp folder, find out the message filename, delete the temp folder, create the known message file as filestream, write the messages, close the stream, create the temp folder - this is the ONLY way I found for creating any message on the fly. In my opinion, that sucks big time.
And because of the "weird" message writing way I now have to script the interface in order to the messages back where they orignally where (in the imap folder). If anyone wants to have a look how i did it, just unzip the xpi and jar, I think its pretty well documented.
Does anyone have any idea how to open a message (where I know f.e. the header from) in a read/write mode??? This would make the whole thing much more stable and quite a lot quicker since I do not have to do everything 3 times..
While I am at it - the mozilla address book should REALLY have a vcard ID field to store the vcard string. In my implementation I used the custom4 field to store the vcard id. This way I always know which card to update on my machines, and I dont have duplicate address cards...
