Print

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

[Deathworks]

Hi!

I know that this bug has been hinted at before, but I think it is really bothersome, especially if people are sharing world gen parameters.

The bug I am referring to is that repeatedly generating worlds in a single session somehow causes the game to create different worlds from the same generation data. It hit me as I created a world, found that it had just one surviving megabeast but would have had 3 megabeasts if I stopped creation 100 years earlier. I extracted the gen data and created a new world in a new session stopping at the assumed date - and only two megabeasts, all caves have different names and history seems to be different. The map itself is similar but not identical:

The old map at the map archive
The new map at the map archive

The original, unrecreatable save
One of the new saves (I tried a few times, trying to avoid all things that could have caused the problem.)

Sorry about bringing this up again, but since people are sharing generation data I feel that this is kind of problematic.

Deathworks

[Tahin]

It looks like the same exact world, just with differently-placed evil/good areas.

Other than that, I can't help.

[Draco18s]

There are some small geologic things too, left lower middle there's a river that's got it's origin in a slightly different spot (different erosion occured).

[Toady One]

I didn't spot the different river, though I see a road going over a river in that location.  Where exactly is it?  The exact point where differences occur is really the only clue I'm going to have if we don't have a method of reproduction.  It seems like the seeds definitely go sour at or prior to evil/good region generation, so a clear confirmation of a river difference would be useful.

[Draco18s]

Must have been the road.  Looked again and it was just black space (due to lost pixels from the zoom level) that made it look like some mountains changed (it was a tunnel)).

[Toady One]

Okay, I'll create a bug report, with a note about the possible good/evil start point for the problem.  I can't reproduce it though.

[Deathworks]

Hi!

First of all thank you for your hard work, Toady (by the way, I also apreciate the name change you have done for the river location entry).

I haven't had much time lately for playing DF, but I tried to recreate the bug, without any success.

The problem is, when the problem occurred, I was doing serious world generation, so I would change a lot of parameters, start world gen, abandon after 1000 to 3000 rejects, adjust some parameters, try again, maybe abandon even a world generation during history when all elven civs are extinct, fiddle again with the parameters (I regularly delete the reject log when it reaches more than about 30MB). Thus, I can't really reconstruct all the steps that went before the bug occured so isolating the trigger is currently not possible.

Since no other reports of this has come in, it seems to be relatively rare, so that if people occasionally restart the game during world gen sessions, they should be fine, I assume.

Sorry that I can't give any more details on it, but I was not really out to find such a bug at that time.

Deathworks