Re: Error in Process Filter

[Simon Josefsson <jas@pdc.kth.se>]

(Cc:d back to the list, hope you don't mind.)

Jake Colman <colman@ppllc.com> writes:

> I see that the start-up time for nnimap is dramatically faster than 0.29
> (where I've been stuck for a while).  It was getting to the point where I
> dreaded fetching new mail and kept MS Outlook up just to see if anything
> important had come in.  I would fetch only when necessary.  Now, I can fetch
> new mail as often as I wish.  Are you doing asynchronously now or is it
> something else that changed?

No, I've turned off tracing of the IMAP library.

If someone wants to work on doing the active info updates
asynchronously, I'd be happy. Same goes for asynchronous article
fetching, of course.

Sorry, my focus isn't speed.

> Will article fetches (that is, fething and article to read) be sped
> up at some point too or is that not an nnimap problem?

This is partly a nnimap problem, I think the IMAP library is very
ineffecient -- it does =lots= of re-search-forwards on each article to
escape character that `read' can't cope with. I think we shouldn't use
`read' at all, but I might be wrong.

The IMAP library won't be seing any development unless we can work out
the copyright issues. If we can't solve the copyright issues, I'll
split nnimap into two trees; the current code that would only see bug
fixes, and one complete rewrite.

/s