Follow-up Comment #2, patch #1983 (project freeciv):

After this patch, I noticed that file #10365 from bug #16708 is loading
differently on my machine: it's now coming out with the map taller than it is
wide, and giving some warnings.

Looking at the savefile, I think should *always* have been "tall" -- xsize is
48 and ysize is 64. But in old versions of trunk it loads as "wide". Bizarre.

I don't know why this is. Savegames from S2_2 load fine (but I don't have any
others from trunk in savegame2 format to try). Probably it just means that
some savefiles generated by medium-old versions of trunk are a bit dubious,
and we shouldn't worry about it. I'm mentioning the behaviour mainly in case
it puzzles anyone else.

While staring at the patch to work out what could possibly have caused this,
I noticed that it incidentally fixes a bug in setting_bool_str_to_bool(),
which was using a function pointer from the wrong part of a union. This could
have been causing widespread, possibly machine-dependent carnage with loading
boolean server options, I guess. I think the bug was introduced by r17554,
patch #1726, 2010-07-10. Perhaps this could explain the behaviour I'm seeing
above by some complicated chain of events I haven't worked out.


Reply to this item at:


  Message sent via/by Gna!

Freeciv-dev mailing list

Reply via email to