Re: Exchange Updating Info Problem

[Ted Zlatanov <tzz@lifelogs.com>]

On Wed, 09 Apr 2008 16:35:44 -0400 James Cloos <cloos@jhcloos.com> wrote: 

>>>>>> "Ted" == Ted Zlatanov <tzz@lifelogs.com> writes:
Ted> For me it generates a long list of integers (UIDs) on an Exchange
Ted> 2007 server (see at end).  Are you seeing something different?  If
Ted> it's just the 15K article UID numbers,

JC> Parsing those 15k digit strings into ints and then compressing them into
JC> ranges can take a lot of CPU when done in elisp.

JC> Especially when there are gaps.

Sure, but can it be avoided?  I don't think the Gnus flags map directly
to IMAP flags, so it's possible there are SEEN messages that need to be
retrieved.  I think someone with in-depth knowledge of the relevant
nnimap and imap code should comment.

In addition, I don't think processing a list 15,000 integers is too slow
with Emacs Lisp on a modern machine.

The other issue is that the SEARCH SEEN itself takes a long time, over a
minute according to Jake's calculation.  See his followup.  So the
number of UIDs may be a red herring.

Ted