TAK Pathfinding Fix for plain TAK (not The Iron Plague).

Hi. I use the TAK release by gog.com. I don't like feel of the new race from The Iron Plague, so I have chosen not to install it. However, I saw this patch and thought it sounded promising. I simply want to know if I can apply the same tweak for 'Kingdoms.exe' instead of 'pgtak.exe'. The old pathfinding value listed in the readme says that it should read as '00000535', but reading this offset in 'Kingdoms.exe' list it as '006700FF'.

So in conclusion, is the ''TAK Pathfinding Fix' compatible with the plain game, or does it just change the wrong offset?

Okay, so I just installed TAK + Iron Plague, and there's not even an executable called 'pgtak.exe'. The old offset value still reads as '006700FF' and not '00000535' like the readme says it should. Which version of TAK are you guys even using, I don't understand.

I forgot to say that it's again the gog version that I'm taking about. I don't understand why there isn't a pathfinding fix that's actually compatible with the release that most people probably have (gog). I'm quite sure this patcher reads the wrong offset in the 'kingdoms.exe', if so, it's not going to work for the gog release.

Hey, I think you would like to check this out: Download TA:K PathFinding

Hey. Yes, I applied the fix to 'kingdoms.exe'. As said, what is confusing is how the 'readme' talks about the default value being '00000535' in hex and '1333' in decimal, but when the program itself reads the default value it reads as '006700FF' in hex which is '6750463' in decimal. It does this both for 'kingdoms.exe' and 'pgtak.exe'. This number is actually a lot higher than the new value of the fix, and this is confusing in itself. I don't know if I'm missing something obviously, but however the case, the program probably works correctly, at least I just assume it does. I mean, I'm sure you guys would have realised if the fix didn't work.

Maybe I'm hard to follow/understand. But I think it's all good.

I guess he (the writer of readme) is talking about the pgtak.exe, pgtak.exe is a PocketGeek TA:K version, which had many versions hacking TA:K to not use cd, use mods, etc. So, I guess the version he was talking about had 00000535 value, and he would change to 0001045A to get better performance. Original Kingdoms.exe has 006700FF value, made by Cavedog. We can't say that TA:K PathFinding fix would work for original .exe (maybe yes, need testing).

So, that's it, we can't guess many things, as it weren't us that created that TA:K Pathfinding fix.

ps.: don't double post, or triple post in the case lol.

Okay, I see. Maybe the fix isn't even needed when the original CPU cycles value for pathfinding is higher, anyhow.

About posting, sorry but I see no edit button, and so I have to post new replies. How can I edit a post?

Ye.

Can't you edit your own posts? Try now.

It should look like that:

Yes, now I can. Thanks.

Btw, I did a little test by comparing the default CPU cycle value vs. the new cycle value. Watch and judge for yourself. Video is a full game speed and accelerated by 150% for fast viewing.



Note! The video will come online in about 60 minutes because my slow upload.

Good Job! I don't know what to say about the video, lol, but I guess in the original value the units seem to not even know that there is a road in their side to pass, the ones that passed, just got stucked, walked, got stucked, walked, then found the place, and in the new value I guess units saw 'better' that there was a road there, but they got stuck in each order.

Original value (is there a road here? barbarians stucked like what the \"****\' I am doing):




New value (oh there is a road on our side let's go there (barbarians) but later then they got stucked each other :p):



Lol, at least it was what I saw.

Hehe, thanks. Well, it was only a single sample test, so I think it's hard to judge based on that fact alone, however, eventually all units will move down to the spot that I clicked on with both of the values. I might test it out more if I get bored.

anyway, the worse map is angvirs maze, for that path problems
but on this game they can stuck even on flat maps :S
btw with what program, you record game?

I use OBS. It's the best and open source.

I have the same issue. I got the GOG version, which worked fine except for the terrible pathfinding. I updated it to 4.1BB only to find out that it suddenly needs a CD for whatever reason. So I've tried the switcher and pgtak.exe works fine, except there's just a black screen instead of movies.
Anyway, both kingdoms.exe and pgtak.exe have 006700FF value at the 0000DED6 offset so I'm pretty sure the offset's wrong. Is there anyone who can confirm it actually works on any version? Because I've compared pathfinding.cfg for TA:K and TA and they use the same offset, even the readme has only the name of the game changed.

Hey WorMartiN. The black screen is because your renderer is set to Direct3D. Change it to Software MMX going to C:\Cavedog\Kingdoms (or wherever you installed your game) and opening ChooseRenderer.exe. It's going to fix the black screen.

I don't think so neither that the offset is the same for TA:K and TA. If you change it to 1, the units shouldn't get an updated path, but they did in my test. So, it doesn't work for TA:K. I've tried to search the TA:K pathfinding offset but couldn't.

It would be way easier if we had developer mode (like TA with +now film chris reload assert or something like that) and used +search command. =/

I get it, somebody just took the TA pathfinding fix and thought it's gonna work for TA:K. Really smart...
Thanks for the renderer advice though.

Yeah... That's what I thought when I just realized it didn't work for TA:K , lol. But we really need it, I'm getting bored of making 5000 units and not being able to move them through the map efficiently. =(
I've tried to look it up in the .icd (the real .exe) though didn't find anything. :S I should learn a bit of assembly and those things in order to get something for us... That's not gonna be that fast maybe.

Yeah Vhaerun, though I don't think changing maneuverleashlength is going to help that much:



It just says how much the unit will go far from their course to attack a ranged enemy that attacked it. For example, if you have maneuverleashlength = 0 on swordsmans, and a ranged unit (kirenna) attack it while they are patrolling, they are not going to move outside of their route to attack her, they will just keep walking. (tested in game)

Source: http://units.tauniverse.com/tutorials/tadesign/tadesign/fbidesc.htm

Faster computer isn't gonna help. The engine is almost 20 years old. TA has a working fix and my computer handles the pathfinding for thousands of units just fine and so can any other current computer.

WorMartin,

Do you know if the TA fix also works in tak?

Have you read any of the previous posts? It doesn't,

Now the original pathfinding in this topic did not work...I assumed it would due to the two games similarities which was not the case; it does not mean I gave up on it, however in a new topic I posted, there appears to be promise as I do not give up so easy. Doing lots of research with different programs, I spent way too much time comparing the sections of code with the TA.exe, trying to find similarities. As I posted in my new topic, the AIsearch section to a point in both the icd and the TA.exe are almost similar. There was only one address(016626, hex C7 46 48) in the whole icd file that appears like in the TA.exe; it was not in the AISearch section and no calls to that address from the AISearch section. It just did not make sense due to the offset. So I found a new address within the AISearch section, hexed that, tested it on a few games and the fix looks promising. Therefore, I need some testers to test it and give some feedback.
Now if PocketGeek or Cheetah were still active here or at least lurking, there could be more promise.

Consequently, I am no programmer...just a modder who has been playing this game for awhile and loves to tinker with it.