• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Andrei Gijgorev

Civilian
3 Badges
May 23, 2007
686
65
  • Arsenal of Democracy
  • Darkest Hour
  • For The Glory
I have a mod that is based on Watkabaoi. In this mod, some province connections are missing in one direction. The effects are basically identical to what happens if you remove those connections from adjacent.tbl, including borders within countries that are drawn in some places. So I thought it might be a faulty adjacent.tbl (I had made some changes to the adjacent.tbl from Watkabaoi) but I could not find any errors there. Then I took the original file from Watkabaoi and put it in my mod. Voilà, the same thing still happens. So I boot up Watkabaoi, and everything works fine there. I hadn't modified any other .tbl files and no other files should have any impact on this. I replaced the adj_defs.txt (just in case) but of course this had no effect. I also made sure that _MAP_MAX_PROVINCES_ is 2020 (same value as in Watkabaoi).

Here's a picture. I send my army directly to Kalmykia and then directly to Terek. Also notice the red borders between Terek and Kalmykia as well as Terek and Abkhazia. For some reason, all provinces that have this problem seem to be in this region. I can send an army into Daghestan, but I can't move it out again as Daghestan has no outgoing connections.

I have no idea what is going on here. On one hand, it looks like an engine bug since there are no errors in adjacent.tbl. On the other hand, Watkabaoi works just fine. Just to be sure, I tested my mod with the old 1.2 version, and the exact same thing happens. So it's definitely not an issue with the latest beta; I had only noticed it now as I was testing the impassable adjacency type.
 
Hi,

I personnaly use the latest patch with my modifies map, and i don't see any problems. But i never touched the adjacency.tbl, except by the Magellan use. I tested the impassable adjacency type and it's working very well :
adjacency = { from = 427 to = 425 type = impassable through = 0 } #Bearn-Aragon
, except that there isn't graphical sprite to suggest this impassability.

As you've modificated the WATKABOI map, do you modificate the ID-layer ? i mean, modificate the ID of the provinces ? I experimented some similar weird things when i changed my map. At the final, it appeared that i left ONE or TWO small pixels of the old ID province, and it was sufficient to create those strange ways.

If this can serve you...
 
What Polak says. Province adjacencies are based on the ID-layer so if you modify adjacent.tbl on its own you're bound to mess things up. What sort of changes are you doing to adjacent.tbl?
 
Guys, I don't think you read my post entirely. If I copy the adjacent.tbl from Watkabaoi into my mod, the exact same thing still happens. Which is why I said it doesn't make any sense; otherwise the problem would be easy to find, no?

I know that impassable adjacencies work; that's not the point. It's only that the reason why I found this issue in the first place was because I was paying attention to adjacencies since I was testing the impassable adjacency. As I said, the same thing happens in 1.2.

Also, I know how to edit .tbl files; before the latest beta patch I had tons of connections removed between provinces with very short land borders, especially those with quadripoints, and I know that connections being removed in one direction is something that can happen if you forget to remove them for both provinces or mess up the IDs/offsets. But this is definitely not the issue here. The issue here is that I have a map with .tbl files that work in Watkabaoi, but not in my mod. Also, the provinces that now have missing connections are some that I haven't touched, ever.

So, to sum it up: Currently, my mod uses the original Watkabaoi map with all original .tbl files (and the original adj_defs.txt which shouldn't even matter) and I still have this problem. Which doesn't make any sense. At all.
 
Alright, I've narrowed down the issue to... *drum roll* provinces.txt.

Unfortunately, I couldn't narrow it down any further. So, I've modified my provinces.txt so it can be used with Watkabaoi (replaced new terrains etc.) and uploaded the file here: http://www.mediafire.com/view/5s46ygbz7atiis2/provinces.txt

I would be grateful if someone could test this provinces.txt in Watkabaoi (any version, start 1337 scenario, pick Golden Horde, move army to Kalmykia and from there to Stavropol) and look whether they can replicate the problem. I have absolutely no idea what could be wrong with this file, or in fact how any provinces.txt could lead to this kind of bug. There's definitely nothing wrong with the definitions of the provinces that have faulty connections (so far the list is 468, 471, 1960, 2006, 2015, and 2017).

Any help would be much appreciated.
 
Thanks for the confirmation! So it's 100% not due to a faulty installation or whatever, and it's 100% certain that provinces.txt indeed causes the bug. I guess I'll have to replace it bit by bit with the original file and figure out what exactly causes this error to occur. This is going to take some work...

Edit: Aaaaand I found the issue: I had made the Caspian Sea (1327) into a river province and forgot to update adjacent.tbl accordingly. So when reading adjacent.tbl, the program finds connections to a province that is neither land nor sea, doesn't know what to do with them and discards any following information, skipping directly to the next province offset. Upon removing all connections to 1327 everything works fine now!

So, problem solved. :)
 
Last edited: