Mon 16 Jul 2007 11:11:41 AM EDT from [EMAIL PROTECTED]
Subject: Re:SVN commit log: revision 5304
Sun 15 Jul 2007 05:15:08 PM EDT from [EMAIL PROTECTED]
Subject: Re:code for consideration
ok, i like it. do you think that this can work for serv_funambol.c too (as this one does s.th. like that? )
my latest commit should give patch a hook it can lean on, so this won't happen anymore.
I gues you could modify it to work in a similar manner but I'm not sure it would be appropriate. My patch is to replicate the Sieve processing. The new hook is called every time a new message is created/saved so that Sieve gets to it quickly before a user might read it in their mail box. Funambol is different in that the message isn't in a normally readable room and it makes sense to batch them up for delivery in bulk once a minute rather than adding the load of doing it every message save. Or are you asking for some reason I'm missing?
Actually, thinking about it I didn't see where the message gets posted into the funabol queue room in serv_funambol.c. Is that what you are referring to? that must be happening in another file (though I may be blind, been a very technical day today).
Riiigght, I've looked deeper at this, found where the message is posted to the funabol queue and, well, erm.
If you mean could a hook be implimented then sure it could be.
If you mean would the same hook work then I don't know enough about how messages pass through magbase.c yet to answer that as the message is added to the funabol queue in a different function.
Actually what I'm working on is a module to change parts of a message body based on a search and replace type thing. I want this to happen before Sieve gets its hands on the message for redirecting and above all I want to ba able to add my module by adding only one line to serv_extensions.c and altering nothing else. As a result I have focussed my efforts to this end.
Looking at the way the funabol stuff works I think it could be rewritten to do things via hooks, possibly even existing ones. the most likely cabdidate being the PerformMessageHooks(..., EVT_AFTERSAVE). Yes a little more code would be needed but it would make msgbase.c a little cleaner.
I think we really need a comment from the author of the funambol module.
