Page 3 of 7

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 10:34 am
by John Adams
[quote="zippyzee":3c2kh3hd]How do we feel about trains[/quote:3c2kh3hd]
Asking the players [url=viewtopic.php?f=49&t=1574:3c2kh3hd]here[/url:3c2kh3hd] as to not derail the Combat thread.

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 11:44 am
by elquar
Chello

No a raider here. As john said from my memory both when VG was released and all the sunset i played. MOBS would not agro anyone else once the lost leash to the initial person who had agro them. They would run back toward their spawn point and nothing anyone else did would stop that till mob reached its spawn point.

IF multiple people had engaged the mob and initial agro person left leash range then mob would attack next person on hate list . example: grip of 3 players A, B , C . All three are fighting the bugbears in three rivers named mob in the mines. It is to tough for them so A and B player retreat leaving leash range . A player was highest on hate list, the mob will then turn attack player C till he or she leaves leash range.

This is how i remember it hope it helps.

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 12:36 pm
by Jammer
[quote="Xinux":3vra5x2h][quote="zippyzee":3vra5x2h]Also, on the way back to its origin location, it should be tagged as non-attackable, correct?[/quote:3vra5x2h]
Correct leashing mobs should be tagged as non-attackable.[/quote:3vra5x2h]
Just a heads up that the wasplings are not pathing their way back to their origin location once they reach the end of their leash distance, they are disappearing and respawning at their origin location.

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 12:43 pm
by zippyzee
Yes they are mega pathing back!

I am aware and going to adjust it. It has not been fully implemented yet but the framework is there. Thanks.

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 4:17 pm
by shargash
I took a couple of my guys out yesterday and beat on things. It was hard boxing because of the "you can't see your target" and "you are too far away" issues, but I could do it with some work. There were no crashes(!), and things were dying (often one or both of my guys). Good stuff.

Re: Combat - Phase 1 Testing

Posted: Sat Jul 25, 2015 5:19 pm
by shargash
Something strange seems to be going on with mob visibility. Almost none of the wasplings will agro my DK (though they often run right past the DK and go for the cleric). Sometimes the cleric gets attacked by mobs the DK cannot see and cannot target, even with /assist. Sometimes either of my characters may be attacked by most neither of them can see and which cannot be targeted (mobs attacking you are supposed to auto-target when they hit you). I am guessing that all the above problems are caused by my characters (the DK in particular) not seeing the mobs where they actually are.

Re: Combat - Phase 1 Testing

Posted: Sun Jul 26, 2015 8:40 am
by John Adams
Just for clarity, when you say you cannot see them, you mean the client reporting "Cannot see target", correct? You can still visibly see the mob in the game world.

Re: Combat - Phase 1 Testing

Posted: Sun Jul 26, 2015 9:06 am
by shargash
No, I mean I can't see it in my main display window. I can't target it with tab, nor can I target it with /assist. My assumption has always been that the two boxed clients are reporting mobs in different locations.

After I posted that, I was also being hit by mobs I could not see on both my clients. I was standing near the bank entrance in Khal, when my DK started taking damage. I could not see or target anything. I was able to camp (I should have bugged that, probably, as you're not supposed to be able to camp when taking damage). An unseen mob then attacked the cleric. I camped the cleric, and then logged the DK back after a couple minutes. Immediately I started taking damage from an unseen mob.

I also note that when one of my accounts watches the other account move (usually on auto-follow), the movement is extremely jerky. This is the sort of thing you would see another player do when either his account or yours was suffering badly from lag. However, the lag meters in Vanguard aren't showing any lag.

So, putting all those things together, I think there may be something amiss in terms of synchronizing mob/player locations between client and server.

Edit: note that both my clients will be coming across on the same IP address.

Re: Combat - Phase 1 Testing

Posted: Sun Jul 26, 2015 9:16 am
by John Adams
Okay thanks shargash. We do know positioning is not perfect, and that something we did a while back "broke" the smooth movement Lokked provided us with. It's likely the position update packets are too few, or not fast enough considering everything is being checked on the same thread, but that is all part of the performance enhancements we'll get to later. We decided to make it work first, then tune it later.

Keep the observations coming.

PS: definitely keep /bugging non-combat related issues you find so we do not forget about them in this sea of posts.

Re: Combat - Phase 1 Testing

Posted: Sun Jul 26, 2015 10:29 am
by zippyzee
There is something deeper to this that I'm trying to work out.

If you are not using auto-attack (i.e. your character is not actively participating in combat), the wasplings will follow you all around and attack as you might expect. They are visible as they should be.

When you attack back, they will sometimes fly off into the distance, visually. The code underneath does not report any changes to the npc movement, as in it is not actively trying to move the npc. If you do a ".spawn details" on the npc it will show it being right next to you, as it should. So the npc is there, but you can't see it.

I'm working on that part right now, but that may be something hidden pretty deep that could take a while.

EDIT: By the way, it seems like I was able to eliminate this while working on something similar. I think the problem was having the npc land exactly on a character when putting on the brakes (so to speak) to its target location. Setting the location to a character-occupied position may have confused the client and it was attempting to find a landing spot for the npc. In any case it is better than what I was experiencing this morning, so it may still occur; it is not during my brief testing now.