diff src/server.c @ 5786:f8bee920b14e

[gaim-migrate @ 6211] Matthew Hall (mthz) writes: " In server.c, "plugin_event(event_buddy_idle, gc, b->name);" gets fired to the plugins BEFORE the call to "gaim_blist_update_buddy_idle(b, idle);" is made a few lines later to update the the buddy struct's idle member. So, if a plugin catches event_buddy_idle and attempts to figure out the buddy's idle time, it will not be correct (in some cases when one first signs on, it will even be completely blank, as it is still zeroed). As far as I can tell, there is nothing wrong with reversing the order of the calls. Attached to this is a patch for the latest CVS... as of 4-9-03 Noon." because he put the patch in the bugs tracker, it took me this long to find it. patches should be in the patch tracker. committer: Tailor Script <[e-mail]>

--- a/src/server.c	Sat Jun 07 02:43:30 2003 +0000
+++ b/src/server.c	Sat Jun 07 03:42:16 2003 +0000
@@ -1064,6 +1064,8 @@
 		gaim_blist_save();
 	}
 
+	gaim_blist_update_buddy_idle(b, idle);
+
 	if (!b->idle && idle) {
 		gaim_pounce_execute(gc->account, b->name, GAIM_POUNCE_IDLE);
 		gaim_event_broadcast(event_buddy_idle, gc, b->name);
@@ -1075,7 +1077,6 @@
 		system_log(log_unidle, gc, b, OPT_LOG_BUDDY_IDLE);
 	}
 
-	gaim_blist_update_buddy_idle(b, idle);
 	gaim_blist_update_buddy_evil(b, evil);
 
 	if ((b->uc & UC_UNAVAILABLE) && !(type & UC_UNAVAILABLE)) {