Complete.Org: Mailing Lists: Archives: freeciv-dev: December 2001:
[Freeciv-Dev] Re: he server appears to have gotten into an infinite loo
Home

[Freeciv-Dev] Re: he server appears to have gotten into an infinite loo

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: <kayeats@xxxxxxxxxxxxxxxxxxxxxxxxx>
Cc: <freeciv-dev@xxxxxxxxxxx>
Subject: [Freeciv-Dev] Re: he server appears to have gotten into an infinite loop in the allocation of starting positions (PR#1135)
From: "Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>
Date: Thu, 13 Dec 2001 18:45:41 -0500

Mapgenerator1 has some builtin requirements that the first two continents 
are polar, that there is a polar sea passage at each pole and that real
continents come after these. These are I think a reflection of some CIV 
compatibility assumptions. The server makes explicit checks for these
conditions for this generator in addition to the mapgen code ramifications.

Of course, the polar passage code as well as many of the make_<feature>
elements should be optional or parameterizable, and any problems with 
starting positions making assumptions about where to start fixed as well 
as lobotomizing this check in the server code.

If you run wrapped over the N/S poles, having an icewall east to west
doesn't make a lot of sense :-).

Cheers,
RossW
=====

At 05:12 PM 01/12/13 -0500, Karen Yeats wrote:
>I don't really see why it should check if the poles are actually connected
>as if they aren't then the polar land will be so bad that the starting
>position allocation won't put starting positions there (and in the very
>odd case that this was not true and the land there was in fact on par with
>the other continents then you would want to be able to start there.)
>
>Do you find connecting the poles make that error more likely or was this
>just a fluke?  I get that error sometimes either way and hadn't noticed it
>being any worse with connected poles.  I get it particularly when I set
>the land to be bad (lots of mountains and swamps for example) or have very
>little land.  I like to play on bad land so this happens to me a lot.
>Really I think that the starting position allocation algorithm is too
>weak, but its a bit of a mess so I haven't done anything about it (yet?).
>
>On Thu, 13 Dec 2001 sigra@xxxxxxx wrote:
>
>> I just tried the new feature with unseparated poles. I got this:
>>
>> 0: The server appears to have gotten into an infinite loop in the
allocation
>> of starting positions, and will abort.
>> Please report this bug at http://www.freeciv.org/
>> Avbruten (SIGABRT)
>>
>> Could it be that it tries to put a player on a pole continent that is not
>> connected to more land?
>>
>> Here is the part of the patch I guess is wrong:
>>
>> -  /* don't want them starting on the poles: */
>> -  if (map_get_continent(x, y)<=2 && map.generator != 0)
>> +  /* don't want them starting on the poles unless the poles are
>> +     connected to more land: */
>> +  if (map_get_continent(x, y)<=2 && map.generator != 0 && map.poles)
>>      return 1;
>>
>> Does this code check that the pole IS actually connected to more land? It
>> seems like it just checks if the pole COULD be connected to more land
>> according to the map.poles rule.
>>
>
> ________________________
>
> It is generally held that the teaching of virtue demands the inculcation
> of falsehood.
>
>       Bertrand Russell, On Education
>
> ________________________
>
>
>
>



[Prev in Thread] Current Thread [Next in Thread]