Harem

UI is requesting threat from Library every 0.5 seconds. Library is updating it's list from everyone else's Library every 1.5 seconds.

 

Antiarc

The TPS update speed is solely client-side and just determines how often TPS is re-calculated based on the threat information that Threat-1.0 itself has received from other clients.. Raising it (or disabling it) will free up more CPU; lowering it will make the display smoother at the cost of more CPU. You never make a request for threat info to other clients; clients publish info as they have it on a timer ranging from (right now) 0.5 seconds to 1.5 seconds based on group size. Note that you're not going to fire off threat updates on that period unless you actually have changes to report, so the actual reporting period tends to be lower.

 

Vulpes

I think this post got lost in the tread but Necrotid brings up a great point. I'm wondering why is it not possible to separate marked mobs through some other means outside combat log. Asheyla seemed to have figured out the way to do it, maybe so can threat-1.0?

 

chase

So any TPS update quicker than the broadcast update only makes TPS smoother for the owner of the client. Granularity for everyone elses TPS is the broadcast frequency. Makes sense now.

e.g. My TPS will be smooth on my UI. The TPS I see for the tank will never have granularity of under the broadcast update.

 

Antiarc

Any spell that can hit something you don't necessarily have to have targeted can't be tracked by any means other than the combat log. Multishot is a good example.

 

◊ Kalman

Actually, even easier, any spell that can be targeted via mouseover is *probably* untrackable by any means other than the combat log.

http://mmorchive.net

The WoW forums, explained:

 

◊ Praetorian

Random question: I know tanks that use mouseover Sunder Armor macros. Is that untrackable?

 

Antiarc

A mouseover sunder armor -might- be problematic. For some reason, I want to say it won't be, but I'll double-check the events tonight. Given Sunder's lack of a combat message when it lands on a 5-stacked target, this will need to be verified.

Edit: Mouseover should be fine - UNIT_SPELLCAST_SENT specifies the name of the target that you sent the spellcast to.

 

Demi9OD

Any chance of getting a toggle to turn off KTM interoperability? I have a strong feeling its far more trouble than good, yet some people will continue running KTM for weeks even after we mandate Threat use. Also, I couldn't see myself on Omen all night tonight. I saw others that had KTM install but had no idea where I was.

 

Antiarc

I committed a KTM filter in Omen tonight.

 

Imba

Just an idea how to different mobs with same name could be authorized by Threat... if they are marked by raid marks maybe its possible to see what marks are what and then recognize threat done to them? :S but i guess that would make even more problems if marks get changed :S

 

whiteshade

i did some trivial modification to the code to let it resize the width of the window & the bars cuase some my guildy stressed me about the dimensions of the window. It's very bad accurate but this let them use it and test a bit in KZ.

btw we tested OMEN with KTM on during our last Prince in KZ yesterday and well it does some mistakes:

a) shadowfiend always displayed
b) aggro was not equal in KTM window and OMEN window (it seems KTM was more precise).

Hope u can fix it soon.

Sorry for my english again.

 

• Chicken

This isn't possible to do accurately. It works for DoT Timer addons as you can always know which exact mob someone targeted and after that the duration of the DoT is pretty much fixed, but it'd get inaccurate for a threat meter which doesn't just have to track the spellcast, but also the damage done by the spell, and the damage done by the spell can vary through partial resists, damage increasing debuffs etc.

Plus there's simply no good way of handling AoE through it.

buff /bʌf/ Pronunciation[buhf]
–verb (used with object)
- to reduce or deaden the force of

 

radiante

There are some know alrorithms that u could use:

- First you could use several windows information an take the average which would have an efect of having a larger window
- You take serveral windows with weight. For example you have two windows one from the last 0.5secs and one from 1 sec ago, you want to give the last one priority then it would be someting like1*WindowOne+2*WindowTwo)/(1+2). This gives a higher priority to the most recent information
- ActualTps =(1-A)*LastTps+A*ThisWindowTps were A is between 1 and 0. For example take A=0.5 and in the first period you a WindowTps=30, your ActualTps=30 because u dont have more info. In the second perioriod you have a Tps=100, then: ActualTps= 0.5*30+0.5*100=65. In the third period you have WindowTps=100 again, then ActualTps= 0.5*65+0.5*100= 82.5
Etc. The closer A to 1 the more weight new information has.

 

Zanthor

Filters out KTM? Highlights the KTM data? Or?

 

• Chicken

KTM Filter. Read it a few times, and you'll get it.

(In case you still don't: Your first option)

buff /bʌf/ Pronunciation[buhf]
–verb (used with object)
- to reduce or deaden the force of

 

Antiarc

I believe that shadow priest threat has problems; my Karazhan run last night showed everyone except our shadow priest right about where he should be, so there's work to be done there.

 

Ellurion

Sorry if I repeat someone, but had some errors today on Hydross the Unstable.

Btw nice mod.

 

Antiarc

Yeah, I've had a few people tell me about the Hydross errors. I think it's a problem in the core, rather than the Hydross module itself, but damned if I can reproduce it.

 

• Chicken

I reported this error on WoWAce before with Hydross, but the exact same 'yell_arg2' error happened to me on Magtheridon as well today. The moment it happened is when he released.
For reference this is the error I'm speaking of: Threat-1.0\Threat-1.0\ThreatNPCModuleCore.lua:187: attempt to call local 'yell_arg2' (a table value)

Could be related to the fact I'm using an enGB client. Or perhaps not.

buff /bʌf/ Pronunciation[buhf]
–verb (used with object)
- to reduce or deaden the force of

 

Antiarc

Hm. I wouldn't think that enGB would cause problems. When did you last update - I made a fix in that area yesterday afternoon.

 

• Chicken

Ah that could be it, I didn't update since yesterday afternoon in my local time, which is before your yesterday afternoon.
Edit: Currently I'm on r38409

I'll try to remember updating before the raid tomorrow, we'll be going SSC again and I'll be able to test if the Hydross error still pops.

buff /bʌf/ Pronunciation[buhf]
–verb (used with object)
- to reduce or deaden the force of

 

Mulva

It's not just an enGB client issue - I had the exact same error on Magtheridon, and it occurred 16 times total according to bugsack. It seemed to correlate to when he was doing his blast nova.

 

Roana

That's probably due to a bug in KTM I discovered today, namely that KTM gives Shield Slam too much threat (my guess is that the innate threat gets added both for "You perform Shield Slam on <mob>" and "Your Shield Slam hits <mob> for <damage>", but I may be off -- I'm speculating as to what the reason is).

Briefly, a 529 damage Shield Slam made KTM 19.18 jump by over 1.7k threat, when it should be less than 1.3k for me.

Omen, in comparison, seems to calculate Shield Slam threat correctly.