Re: Move is slow

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

Mats Lidell <Mats.Lidell@contactor.se> writes:

> Moving articles has become painfully slow for me. It has been so for a 
> while. I can't say from what version it started. Maybe it has
> something to do with my IMAP server rather than with nnimap. (On the
> ther hand splitting feels as fast as before so why should moving be so 
> much slower?)

Gnus does some stuff when moving articles, but that hasn't changed. I
added bug-workaround code for Cyrus IMAPD in 0.116 which slow things
down when moving several articles at once (but not by a lot -- all
stuff Gnus does should take much longer than that, for comparison).

If you have a mailbox with very large article UID's then the
gnus-range-* functions will be very slow. This is a major problem and
I don't think it will go away until someone implement the range
functions in C.

> So how is it. Is moving articles slower now than before? If not. What
> should I look for in order to find out what is happening.

Two ideas:

. M-x untrace-all RET, M-x elp-instrument-package RET gnus RET, M-x
elp-instrument-package RET nnimap RET, M-x elp-instrument-package RET
imap RET. Move an article, then M-x elp-results. Then you'll get a
list of functions where time is spent, mail it to me if there isn't
anything obvious about it.

. Kill *imap-log* and *nnimap-debug* buffers, move an article and send
me the output -- then I can find out if it's doing something it
shouldn't do which could take time. Sadly there are no time stamps so
I couldn't tell exactly what part take time.