WTB old pathing back

[oksnoop2]

Just wanted to say that I've not noticed any change in pathing in playing ct. Don't know, might just not be observant.

[Forboding Angel]

Probably not, I'm not surprised. CT doesn't use retarded values for unit movement. I warned first Caydr, then Noize, and then TFC about the fail that is unit movement in BA YEARS ago, but no one listened. I even posted a fix for it. No one listened.

WHAT'S UP NOW BITCHES?

Edit: On a side note, I was lookign at the moveinfo.tdf for BA.

Inside the defs it says:

heatmapping = FALSE;

Unless I'm mistaken, since this is tdf, it needs to be:

heatmapping = 0;

[Johannes]

It's not game specific, and heatmapping is off by default. You're just not observant enough.

[hoijui]

@Otherside: the point is, it is actually YOU who is leaving the path finder in its current state.

[Google_Frog]

So do you suggest he grab 0.82.4 and call it 0.82.7?

[hoijui]

if you had read my previous posts in this thread, you could have saved yourself from looking stupid.

[Licho]

Ok, how hard would it be to just backport ingame joining to old engine version?

[hoijui]

Licho, why would you do that, instead of porting the old path-finder stuff to new spring?
we would not set old spring (eg. 0.81) with mid-game-join ported back live on the server. it is not the only change there was, next to path-finder stuff.

[FLOZi]

Interestingly enough, EvoRTS, GundamRTS, and S44 are largely unaffected by the pathfinder. Only issue I see in evo is that occasionally a unit will get stuck on some terrain.

Actually S44 has been quite badly affected:

http://springrts.com/phpbb/viewtopic.php?f=11&t=24312

And yes changing accel / decel / maxvelocity ratios can improve it, but we're trying to model realistic behaviour.

[Gota]

Wow, this thread is getting better by the minute...
I thought reverting is huge issue and no dev wants to do it?
Whats all this noize about..
Seems to me the only way forward is sticking with what we got and hoping it gets completed.

[Licho]

Licho, why would you do that, instead of porting the old path-finder stuff to new spring?
we would not set old spring (eg. 0.81) with mid-game-join ported back live on the server. it is not the only change there was, next to path-finder stuff.

Because from CA player perspective recent versions are much worse:
* dramatic reduction in performance (at least for ATI users like me)
* broken effects
* pathing makes it almost unplayable (units stuck in factories, stop because of slight irregularities or because of slight cratering)
* lua bugs like incorrect arguments now often cause full engine crash
* several widgets which worked for years now dont work (even basic stuff like minimap events)

Only visible pluses since older version are:
* ingame rejoin
* specular lighting fixed

[smoth]

Only visible pluses since older version are:
* ingame rejoin
* specular lighting fixed

* SSMF
* lua lobby stuff (should make having lan parties easier)

[Google_Frog]

from CA player perspective ... visible pluses since older version are

We don't use lua lobby and can live without SSMF.

Forb you seem to be saying that this engine change that forces high manoeuvrability on all your units is not a bad thing and that everyone should just get over it and buff their units. This is stupid. With your logic any engine restriction is fine as it is possible to make a game that works with that restriction.

[Forboding Angel]

Changing acceleration and braking rates does not make your units magically maneuverable (that's why god (a.k.a. The engine devs) invented turnrate ). It does however allow them to gather speed fast enough so that they can react to undesirable situations.

In the case of BA, a global set of accel and brakerate to a min of 0.1 should do the trick.

Seriously, I'm not kidding, do the math and see how long it takes a flashtank to reach full speed. Once you realize the time involved, then you start to develop a clearer picture of why they are having so much trouble.

Also, setting that in BA will not perceptively affect balance (if it did, then something inside BA is horribly horribly wrong). Would it help if I made a little mutator to illustrate the point?

[Google_Frog]

Acceleration is part of manoeuvrability. Forb are you being intentionally dense? Do you know what the issue is?

Here is how I see you see it:

I just discovered that many units have poor manoeuvrability, baww baww baww

At which point you suggest increasing their accel and deccel.

Yes this makes them more manoeuvrable (or reaction time as you call it) but it doesn't fix what Otherside was actually talking about. As in that the new pathfinding for low manoeuvrability units sucks.

[Forboding Angel]

Acceleration is part of manoeuvrability. Forb are you being intentionally dense? Do you know what the issue is?

Here is how I see you see it:

I just discovered that many units have poor manoeuvrability, baww baww baww

At which point you suggest increasing their accel and deccel.

Yes this makes them more manoeuvrable (or reaction time as you call it) but it doesn't fix what Otherside was actually talking about. As in that the new pathfinding for low manoeuvrability units sucks.

Yes I know what the issue is, and no that is in no way what otherside said.

His exact words were:

...the pathing is horrible and makes some games nearly unplayable without spamming line move.

Which is completely different from what you said that he said which was:

As in that the new pathfinding for low manoeuvrability units sucks.

And as an act of me being kind, here is much better pathing for BA in a cute little mutator (99.9999999% of players won't even be able to tell the difference).

BA719pathing.sd7

(241.82 KiB) Downloaded 12 times

If you can't say anything nice, don't say anything at all.

[hoijui]

@Licho:
obviously, you can not list pathing as an argument for your way of doing it, as it would be "fixed" in both methods just the same. i would bet on that there are more pros of the new version, but i am too lazy to go through the 1.5k commits too .
the thing is, in a few months there will be an other mayor version, and again, and again...
with your method, we would have two parallel spring engine projects, while with the other method, you only had to merge master/release into your separate "fixed" branch from time to time.
Who knows, maybe you will not be happy with Kloots new path-finder either, what then?
back-porting everything you like is like 1000 times the work of merging in master into a branch that has only some stuff back-ported. You may think, the two things are in essence the same, but they are not. The important thing in the branch would be the structural changes. if you do not have them (and you would not with your approach, cause it would be much more work then using the other approach), you get into a merge hell.

in short: your way of doing it, means maintaining a separate engine project (which also means, two versions of each mod/AI for example, if they want to be compatible with both projects). And yes, this is even a huge problem if you only want to maintain it till next mayor version.

it also means, due to more and more internal structural changes, it will be practically impossible to back-port stuff introduced from now on already.

[Licho]

Well I don't mean to maintain it, just one time merge to make beefed up old version.
Apart from those features I listed, old engine was fine and we could do almost all we wanted with it.

So we would not need to change it at all and just wait until new engine is more suitable than the old :)

Actually perhaps its worth for CA even without merge - even without improvements.

[Johannes]

Here's some Evo pathing for you
http://replays.adune.nl/?2821

As in that the new pathfinding for low manoeuvrability units sucks.

No that's not really accurate. Problem is that pathfinding for all units sucks. Some just suck a little bit less.


If someone really wants to defend the current state of the pathfinder, please list some improvement it has over the old one. So far I haven't noticed any single thing.

[Forboding Angel]

660 is old...