Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Chaoswizkid]

I searched for this, but nothing quite came up like it.
This is the second time this has occurred to me, in a previous version, though I forget exactly which previous version, though I suspect regular 40d. I will describe the bug in 40d9, however.

I entered the "Design New World With Parameters" section. I then started modifying the advance parameters of the preset "Large". I changed 5 things: I set the end date, normally at 1050, to 900. I changed the temperature minimum to -30, and the maximum to 90. Minimum volcanism was set to 50. I also changed the number of wetland tiles briefly, then changed them back to their default before world generation. I saved and started to generate the world.

I got nothing but a mass of green n's with instantaneous rejections. I aborted once it came to nearly 300 rejections. Thinking the most important change was the temperature, I set both the X and Y temperature variation to 800, not exactly knowing what that would do. I saved, and there was no effect. I tried repeatedly, to no avail. I re-entered the advanced parameters, loaded a previous version of DF (40d, instead of 40d9), and ensured that I had every parameter set to their default on my 40d9 version. I started world gen to make sure...  but all I had was green n's. I started the world gen on 40d, and it perfectly generated a new world. After triple checking in game, I decided that there must be a text file with the information. I dug into both files, found the world_gen.txt for each one, and found what appears to be the culprit.

I checked both versions, and they seemed to be exactly the same... however, at the end of the LARGE parameters for 40d9, I got something like:

[EN_SU]:200:200:200:200:200:200:200:200:200:200:200:200...

There were MANY [EN_SU] (could have been _SL or something else) entries with many more :200's than I displayed. There was also a different type that had :50's instead.

There were so many of these that the world_gen.txt file for 40d9 was around 1200 kilobytes, WELL over what it should have been. I deleted every suspicious entry seeing that my other version didn't have them. I then copy pasted the LARGE parameters from 40d to 40d9, saved, and retried to world gen on 40d9. It worked perfectly.

I have since tried to reproduce the bug, going step by step, but it appears the game will still operate the world gen fine. Though I am getting many rejections, I at least see some development, rather than a green sea of n's.

[Variance]

I don't know what the individual technical problems you were having were, but if you had stuck through with worldgen in the original attempt you made, it would have told you what was wrong and asked if you wanted to ignore that rejection type after about 550 rejects. I've genned worlds with much harsher changes and told it not to reject anything when it asks at about 550, and it works fine (most of the time).

[Untelligent]

Did you turn off all the corresponding reject parameters?

[Dorten]

It somehow used the drawn map function (which is en_su thingie)...

[Chaoswizkid]

I don't know what the individual technical problems you were having were, but if you had stuck through with worldgen in the original attempt you made, it would have told you what was wrong and asked if you wanted to ignore that rejection type after about 550 rejects. I've genned worlds with much harsher changes and told it not to reject anything when it asks at about 550, and it works fine (most of the time).

Even if it had told me what was the problem at that point, there'd still be something very wrong considering the minor changes were just handing out hundreds of worlds complete composed of green n's. Also, the massive log in the world_gen.txt of EN_SU wouldn't have been solved from that report, or at least I highly doubt it.

Did you turn off all the corresponding reject parameters?

What are the reject parameters?

It somehow used the drawn map function (which is en_su thingie)...

What is the drawn map function?


Also, as I said, the problem has already been solved by removing the EN_SU and the other EN_ etc., but because this was a technical error that seems to be of the program and not of my accord, I think that Toady should look into... unless it occurs only to my PC, then it isn't worth the time.

[Sleet]

There's a terrain editor function in the advanced parameters section of world-gen.  EDIT:  Reached by hitting (e) for "enter advanced parameters" and then (p) for "set preset field values." \EDIT Even if you just glance at it, the game remembers that you went to the editor and thinks those are the parameters you want-- set square by square (hence all the EN_SU's.)  The default setting in the editor is middle value for each parameter for each square-- 50 volcanism, whatever the middle height and drainage and rainfall are, etc.  The result is world with nothing but hills everywhere, which the game rejects out of hand.  I don't remember exactly how to clear it the terrain editor's values ingame-- you might have to make a new parameter set and reset all the advanced parameters.

So, not exactly a program error, but not intuitive either.  But what in DF is?

[Variance]

Even if it had told me what was the problem at that point, there'd still be something very wrong considering the minor changes were just handing out hundreds of worlds complete composed of green n's. Also, the massive log in the world_gen.txt of EN_SU wouldn't have been solved from that report, or at least I highly doubt it.

You'd be surprised how adaptable worldgen is, as long as it's allowed to run normal erosion and water processes it'll try to create a good world out of anything, even a perfectly flat plain. You just might have a large number of temperate grasslands. That said, the green "n"s do sound like a bit of a bug; were they all in a row, the hundreds of green n maps being one right after the other without interruption? You make it sound like there were a few that were not just green n maps.

If the green n maps were like that, then you may have triggered the drawn map function by accident and it was using a blank picture as its guide. It's unlikely, but possible. You might've just set the preset values like Sleet said, and the preset values are NOT the default values.

What are the reject parameters?

DF just looks at the number of biomes created and randomly generated topography, and decides if they match up with the advanced parameters for your world. If you changed the advanced parameters, then chances are you'll see at least a couple hundred rejects before you strike home.

What is the drawn map function?

You can import an image with the right file extension and features (I'm not exactly sure what file type and color codes are required, I've never done it) and use that image to determine the layout of your world.

[Avelon]

What happened is you accidentally entered the editor and spaced out of it without doing anything. Just by entering that editor (which will show you a bunch of hills), you told the world generator that you're using that instead of randomly generating. I did this a few times, first by accident and then to test, and the only way to fix it is to exit the Design With Parameters menu without saving, then re-enter.