*MAJOR* issue that needs to be fixed before release

Cabbage · Post by **Cabbage** » 29 Sep 2006, 19:40

+ 1 to souless, if you could make all structures lose their targets once they are out of range then its sorted as far as i can tell. You dont need to worry about mobile units really, they'll either follow and get killed, or follow and succeed in their task, same as it is now, but you'll have to spot the unit first, rather than checking your homing LLT :p

NOiZE · Post by **NOiZE** » 29 Sep 2006, 19:45

When a unit goes Out of los the order should become a posistion attack order after 5 seconds..

Soulless1 · Post by **Soulless1** » 29 Sep 2006, 19:52

possibly we could have the option of an expired order either being dropped completely (unit stops) OR having an attack-move set (which might get messy and annoying for the player) to the location. Again, moddability is key

PauloMorfeo · Post by **PauloMorfeo** » 29 Sep 2006, 19:58

jackalope wrote:
Zirjin wrote:I don't know the exploit at all. But what about the orders are maintained for as long as the unit is visible on radar, not just LOS?
this sounds reasonable to me

Foe OfTheBee wrote:
Zirjin wrote:I don't know the exploit at all. But what about the orders are maintained for as long as the unit is visible on radar, not just LOS?
This plus maintaining the order indefinitely for buildings would be my preference.

All of that sounds good to me. Maybe with one obvious addition:
"This plus maintaining the order indefinitely for buildings would be my preference" until building is spoted destroied. That, if the building is not under radar, of course.

Also, maybe the unit attacking last known position if nothing else is targeteable.

That *bug* was an exploit that i see with good eyes gone.

Caydr · Post by **Caydr** » 29 Sep 2006, 19:59

Cabbage wrote:+ 1 to souless, if you could make all structures lose their targets once they are out of range then its sorted as far as i can tell. You dont need to worry about mobile units really, they'll either follow and get killed, or follow and succeed in their task, same as it is now, but you'll have to spot the unit first, rather than checking your homing LLT :p

Problem is that all you have to do to make a "unit" stationary is first give it an order to guard something, then give it the attack order. Since a guard order never ends (quiet), once again you've got the homing llt behaviour.

Cabbage · Post by **Cabbage** » 29 Sep 2006, 20:12

Machiosabre · Post by **Machiosabre** » 29 Sep 2006, 20:18

Cabbage wrote:

in your face Cabbage!

FoeOfTheBee · Post by **FoeOfTheBee** » 29 Sep 2006, 20:19

Caydr wrote:So what happens when you don't have radar, you're just fighting with kbots. The enemy kbots leave your LoS for a split second and yours stop moving and fighting.

Maybe all units should have radar capability of 1.5 los or something...

Post by **hollowsoul** » 29 Sep 2006, 20:31

Ideally units should carry on to last know location firing @ last known spot.

Unless they see another target ideally till they regain los on target (possible tag for unit to enable / disable targetting another target).

If no sight when @ last spotted location. Just go idle
Prob would need unique id for each unit. And prob be a pain to code aswell

10053r · Post by **10053r** » 29 Sep 2006, 20:32

Is my suggestion too difficult to implement? Here are several suggestions in increasing order of difficulty of implementation:

1) Keep firing at the last known coordinates until that spot enters LOS again (and the target isn't there) or the target is spotted somewhere else.

2) Do 1, and also attempt to follow the unit to those coordinates to get them into LOS again (assuming that the attacking unit isn't a building). This might get attacking units killed needlessly, but it fixes the problem and is slightly more intelligent.

3) Implement my suggestion from above. To recap, that suggestion was to create an expanding search radius and attempt a search until the unit was found or a time limit passed or the search radius got bigger than the map.

Any of these will be somewhat reasonable behavior, and will at least let us keep playing. 1 should only take a couple hours to implement at the most.

FireCrack · Post by **FireCrack** » 29 Sep 2006, 20:50

Rather than an expanding search aradius, it would be more intelegent to shoot at where the unit should be given it's last known velocity...

Soulless1 · Post by **Soulless1** » 29 Sep 2006, 21:00

Caydr wrote:
Cabbage wrote:+ 1 to souless, if you could make all structures lose their targets once they are out of range then its sorted as far as i can tell. You dont need to worry about mobile units really, they'll either follow and get killed, or follow and succeed in their task, same as it is now, but you'll have to spot the unit first, rather than checking your homing LLT :p
Problem is that all you have to do to make a "unit" stationary is first give it an order to guard something, then give it the attack order. Since a guard order never ends (quiet), once again you've got the homing llt behaviour.

that's why I suggested making a per-unit-defined order expiry timer

Caydr · Post by **Caydr** » 29 Sep 2006, 21:05

Will these solutions fix small groups of fighters being useless for ground attack because they keep on trying to land the instant they lose LoS (every 2 seconds)?

Geez that's irritating... Maybe I should increase fighter sight range a fair bit.

Tobi · Post by **Tobi** » 29 Sep 2006, 23:16

I'll fix this with a timeout of 15-30 gametime seconds, and possibly also have the order not be canceled on ghosted buildings if that's easy enough.

Post by **Tim Blokdijk** » 29 Sep 2006, 23:24

Tobi to the rescue!

I must say that I'm really impressed by the effort you and Jelmer are putting in this release.

rattle · Post by **rattle** » 29 Sep 2006, 23:30

It's the release with the most changes isn't it?

Post by **Tim Blokdijk** » 29 Sep 2006, 23:34

By far

Argh · Post by **Argh** » 30 Sep 2006, 01:54

... the list is incredible. I'm about to mess with the latest from Yeha now.

Caydr · Post by **Caydr** » 30 Sep 2006, 04:24

The changelog from this version is probably bigger than that from any other 5 versions combined, or more... Lots of amazing new stuff.

LordMatt · Post by **LordMatt** » 30 Sep 2006, 08:22

Thanks to all the people who worked on this, it's gonna be great.

Spring RTS Engine

MAJOR issue that needs to be fixed before release