[Freeciv-Dev] Re: Buglet.

["Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>]

I'll put out something by the end of the weekend that has most of these
sort of things fixed in passing. And may be sufficiently sanitized to be 
considered for CVS. However it would be really nice to run a variant 
branch for more serious overhauls while they are shaken down, rather than
letting everyone take the hit :-)

I've been running Flat and Donut worlds with the latest corecleanup stuff
and it has been useful to track down a lot of the residual hand coding and
normalize_map_pos/adjust_map_* confusion. 

Cheers,
RossW

At 12:33 PM 01/08/10 +0100, Gaute B Strokkenes wrote:
>On Fri, 10 Aug 2001, Thue wrote:
>
>>On Friday 10 August 2001 01:34, Gaute B Strokkenes wrote:
>>> There seems to be a bug in tile_move_cost_ai() map.c.  I think the
>>> right half of the assertion on line 977 should be y < map.ysize,
>>> rather than y <= map.ysize.  Or?  If this is intentional, I think it
>>> is sufficiently non-obvious that someone who knows this part of the
>>> code should add a comment.  Otherwise, we should replace the line
>>> with `assert(is_real_tile(x, y);'.
>>
>>It should an "assert(is_real_tile(x, y);". But tracing the x,y args
>>means going through most of the source. Could we wait until just after
>>1.12.0 for fixing this? (It seems to work as it is...)
>
>Sure.  The code is just a bit too permissive.  I thought this might be
>intentional because you might have to worry about crossing the edge of
>the map or something.
>
>-- 
>Big Gaute (not to be confused with LG)
>Mit der Dummheit kämpfen Götter selbst vergebens.
>
>
>