Re: Holes in article sequence

[Hrvoje Niksic <hniksic@srce.hr>]

François Pinard <pinard@iro.umontreal.ca> writes:

> I've often seen complaints about inefficiencies induced by big holes
> in the sequence of read article numbers, but I do not know the deep
> reason behind such inefficiencies.  Would someone accept to
> enlighten me?

The one inefficiency I know of is when you use total-expiry and nnml.
Because of the way total expiry works, Gnus has to stat all the files
in the article range, and delete the ones that should be deleted
(usually those older than N days.)

Now, if you have a big hole in your article range -- e.g. your article
range is 128 - 12300 because you ticked article no. 128 -- total
expiry will stat tens of thousands of non-existant files every time 
you leave the group.  That tends to get a bit slow.

The correct solution would be to expand Gnus' concept of "active
range" to be a list of ranges instead of a single range.  Then
total-expiry could work sanely.  David Moore had interesting ideas
about this, but he disappeared before doing anything.