Bloodboil kill, my profiler resets when the combat starts. LibQuickHealth-2.0 includes it's own copy of CBH. 2ms CPU in 8 minutes is ok I guess.
Those values are wrong because of a stupid addon called MTLove which resets the blizzard profiler. So I'll have to redo that test. LQH2 alone (w/o CBH) is at 0.1-0.2% CPU (80ms in 49s, 168ms in 132s, 375ms in 317s).
oUF_Caellian\oUF_Caellian.lua:97: bad argument #1 to 'len' (string expected, got nil)
oUF_Caellian\oUF_Caellian.lua:97: in function `UNIT_NAME_UPDATE'
oUF_Caellian\oUF_Caellian.lua:189: in function <Interface\AddOns\oUF_Caellian\oUF_Caellian.lua:114>
oUF_Banzai-1.0\oUF_Banzai.lua:42: in function <Interface\AddOns\oUF_Banzai\oUF_Banzai.lua:41>
oUF_HealComm-1.0\oUF_HealComm.lua:98: in function `PostUpdateHealth'
oUF-1.1\elements\health.lua:83: in function `UNIT_HEALTH'
oUF_QuickHealth\oUF_QuickHealth.lua:53: in function <Interface\AddOns\oUF_QuickHealth\oUF_QuickHealth.lua:45>
<string>:"safecall Dispatcher":4: in function <[string "safecall Dispatcher"]:4>
<in C code>: ?
<string>:"safecall Dispatcher":13: in function `?'
CallbackHandler-1.0\CallbackHandler-1.0.lua:91: in function `Fire'
LibQuickHealth-1.0\LibQuickHealth-1.0.lua:130: in function `?'
LibQuickHealth-1.0\LibQuickHealth-1.0.lua:38: in function <...ace\AddOns\LibQuickHealth-1.0\LibQuickHealth-1.0.lua:38>
Got 15 of those during a normal instance run. Using both oUF_QuickHealth and oUF_HealComm running at the same time...
Uhm... this may not be the right thread for it, but there was none for oUF_QuickHealth...
How would that even be necessary? SCT presents stuff fromthe combatlog. QuickHealth uses data from the combatlog. They use the same source. There may be some kind of throttling going on for SCT, but maybe you just need to find the switch to turn that off...
It would be a bit more accurate, but still not exact. To get exact OH values, you'd need to wait until the next UNIT_HEALTH event, then look at the combat log entries in retrospect, and analyze what exactly happened. I thought about adding that to LQH, but i won't since it's not necessary in wotlk.
It looks correct (didn't test), but also unnecessarily slow. When you receive an update for e.g. unitID=='raid3', can't you call UNIT_MAXHEALTH just for the frame showing raid3 instead of every visible unitframe?
Today I had the profiler running on our mother kill. LQH had it's own CBH, so it should include the CPU usage of CBH. One addon was listening to UnitHealthUpdated, one to HealthUpdated. CPU usage is ok in my book, I don't see a reason to bake our own CBH.
There appears to be a problem with the focus health. I'm not exactly sure what's wrong. I remember the focus not getting updates sometimes, or showing just plain wrong values (50/100 for someone in the raid). Couldn't reproduce it so far. Did someone else notice that?
Sorry about that. The problem occured when target or focus didn't change, but UNIT_PET, RAID_ROSTER_UPDATE or PARTY_MEMBERS_CHANGED fired. I assume your focus didn't change when you entered the arena, and RAID_ROSTER_UPDATE fired because you joined a raid. Should be fixed in the svn now.