CEONSS dedicated server test - Saturday Oct 8th, UTC+2 22:00

[captainsnarf]

Thanks trans_lux! I appreciated seeing you there

I have more unreg hits on the omni server with 60ms ping than I did the test server with 160ms ping. I guess YMMV. Player weapons had ping compensation but the vehicles didn't, so that took some adjustment.

[leon]

McLovin, ask Europe. Maybe someone recorded the matches.

[Enyo]

Sorry I couldn't make the test this weekend, my son had a baseball tourney out of town that I forgot about when I originally said I'd be here for it. So, what's the conclusion? Seems like the high cost of a dedicated server wouldn't be justified. Or would it? Is more testing needed?

[pooty]

So the take away was on that server, with 16 players, CPU peaked around 23%. I think that's pretty similar to what we see, so we should be able to compare UTComp on our current server with the stats below....and we can use the same maps to do it. We should test our current server with UTComp when we feel its ready and compare.

The bigger takeaway seems to be that UTComp has progressed to the point where its almost ready to test again on our server. No crashes with UTComp (We did play Minus-KS and HSM-Xmas)

There were a couple of items to work on (Snarf jump in a feel free to correct)

Must Haves:
- Flak Cannon hitting nodes at point blank range -- somehow this gets messed up with enhanced netcode. (FYI enhanced netcode works for player weapons only, not vehicles), potential fix is no enhanced netcode for Flak Cannon (works as it is now)
- Fixing enemy team skins. I think there's one option that works (Brightskins?) and the two other ones have issues with team switches -- including balancer.

Nice to haves:
- (my selfish add) getting vehicle names to the dots -- I know snarf you said that's a bunch of work -- I'd be happy to help with that.

I think everything we are using in ONSPlus is there in UTComp.

[Miauz55555]

I streamed it: https://www.twitch.tv/miauz55555 the quality of the video (should be there for a few more days) isn't the best .. just need to find some good settings .. but you can see the skin thingy in it .. and also how well the people with a higher ping were hitting (very good).

[captainsnarf]

Pooty, your "must haves" are the items I was thinking need fixed.

Your "nice to have" is already implemented on the test server

[pooty]

Your "nice to have" is already implemented on the test server

Boom. Freaking awesome.

[captainsnarf]

For the skins issue, I originally disabled 'enemy based skins' and set the default to 'UTComp style'. This gave everyone bright skins. Red team was bright red and blue team was bright blue. Since the checkbox was disabled you couldn't change the colors. This made the color you see match the name of the color for each team.

To allow enemy based skins to work, I need to add more options for the different teams. With enemy based skins enabled, everything (code, menus, etc) assumes your team is red and the enemy is blue. This becomes a problem if you are on the blue team and you set the enemy skin to be a blue color. Everyone will be blue and you won't know who to shoot at.

So instead of the dropdown having two options "Teammates" and "Enemy" there will now be "Teammates (Red)", "Teammates (Blue)" "Enemy (Red)" and "Enemy (Blue)" and you will be able to set each of the colors individually. It's a lot of work but I think this is the best way to support enemy based skins in team games.

I think a lot of people will that use this feature will set "Enemy (Red)" and "Enemy (Blue)" to the same color e.g. bright green, but having options for all of them allows it to work the same as with 'enemy based skins' disabled, i.e. red team is always a red color and blue is always a blue color as long as you set them up like that.

Does this sound right or should I do something else?

[pooty]

Is it worth a lot of work?

I'd probably use it as the red team is always harder to see for me.

[captainsnarf]

Hopefully not too much work, but yes more than I'd prefer