[EMAIL PROTECTED] said: > Currently, the code uses db_fetch_headers() for every message in the > search cycle. Despite the name, this function actually parses the entire > message, multipart boundaries and all. (Misleading function names are > possibly the worst thing about this code...) > > But, for header field searches, one could do with only parsing the main > header. This action is already available in the current code, as > db_get_main_header(); but it is only used for IMAP FETCH.
In combination with a query that asks the database to only return potential matches, this should make search much, much faster. I'm going to file it as a bug so that we make sure to get this into 2.0.x ASAP. Aaron --
