Paul J Stevens wrote: > Mikhail, where's the db_header_speedup patch?
I'll post a roundup of all my patches on my Web site today. I just hope to make one more patch before that (fixing the SEARCH UID bug). Then I'll put all my patches online together, so they would not be scattered in the mail archives. > I've been testing the icfetch_speedup patch. Works great. > > My only objection is that it extends the coding style used before in > _ic_fetch. A style I'm desperate to get > rid of. Fully agreed. But, this patch is for the 2.0.x line, where this style is a given. > But the basic idea is valid and I'll make an issue of applying it to > the cvs-head codebase. Please also read my post on "ideas of further speedup" - some more basic ideas there. (I'm afraid that a correct C implementation is beyond what I can do). > I agree. If it works out ok during the client tests, it should be > applied to the 2_0 branch. > Same goes for the dbmysql and my own db_getmailbox patches although > that last one requires an addition index > on dbmail_messages to be effective. Whether to allow an added index in the 2.0.x brahcn is not for me to decide. Yours, Mikhail Ramendik
