Sorting threads (and articles)

Sorting threads (and articles)

[Lars Magne Ingebrigtsen]

As was suggested earlier, Gnus now creates composite sorting functions
based on the functions in `gnus-thread-sort-functions'.  Gnus uses
this composite sort function, and does just a single `sort' call
instead of one `sort' per sorting function.

I wasn't really sure whether this would be faster or not, since the
composite sorting function isn't byte-compiled.  And...  the new
method seems to use 20% of the length of time the old method used.
Heh.  

(I've only tested with sorting over article number and subject;
sorting over date is still just as slow as always, I would think...)

-- 
   Lars Magne Ingebrigtsen * larsi@ifi.uio.no
      (a red leaf that falls from the purple tree)

Re: Sorting threads (and articles)

[Sten Drescher]

Lars Magne Ingebrigtsen <larsi@ifi.uio.no> said:

LMI> As was suggested earlier, Gnus now creates composite sorting
LMI> functions based on the functions in `gnus-thread-sort-functions'.
LMI> Gnus uses this composite sort function, and does just a single
LMI> `sort' call instead of one `sort' per sorting function.

LMI> I wasn't really sure whether this would be faster or not, since
LMI> the composite sorting function isn't byte-compiled.  And...  the
LMI> new method seems to use 20% of the length of time the old method
LMI> used.  Heh.

	Given this result, are you going to look at creating composite
group sorting functions?  And does this still leave the bug with
gathering orphaned articles, or has that sorting been moved to a
different function now?

-- 
#include <disclaimer.h>				/* Sten Drescher */
1973 Steelers    About Three Bricks Shy of a Load    1994 Steelers
1974 Steelers         And the Load Filled Up         1995 Steelers?
Unsolicited email advertisements will be proofread for a US$100 fee.