[Pidgin] #4816: Oscar protocol performance improvement(s)

[Pidgin]

#4816: Oscar protocol performance improvement(s)
-------------------------------------+--------------------------------------
 Reporter:  oliver                   |        Owner:  MarkDoliner
     Type:  patch                    |       Status:  new        
Milestone:  Patches Needing Review   |    Component:  AIM        
  Version:  2.3.1                    |   Resolution:             
 Keywords:  oscar performance patch  |  
-------------------------------------+--------------------------------------

Comment(by MarkDoliner):

 Finally looking at oscar_ht.patch...

 I wonder if the memory overhead from the hash tables is worth the saved
 processing time?  Maybe one or two of the hash tables is used enough to
 justify its existence, but maybe we could get rid of 1 or 2 of the other
 ones?  And I wonder if a GTree might be more appropriate than linked
 list/hash table combination?

 A few things I've run into so far (I don't know if you're still around,
 but don't bother fixing these things, I'll post an updated patch soon):
  * There's a crash bug where aim_ssi_itemlist_del calls itself, which
 results in infinite recursion
  * aim_ssi_item_create_key() should probably re-use the oscar_normalize
 function instead of reimplementing it
  * aim_ssi_item_set_name() needs to remove groups from the
 list->idx_groups hash table when a group name is changed
  * The strcmp in aim_ssi_cleanlist() crashes for me because key_name is
 NULL for some people.

-- 
Ticket URL: <http://developer.pidgin.im/ticket/4816#comment:14>
Pidgin <http://pidgin.im>
Pidgin