This is the mail archive of the xconq7@sources.redhat.com mailing list for the Xconq project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: More display problems with tcltk interface


>The bug doesn't seem to effect a unit's ability to see other units; just
>its ability to map unknown terrain.  If you turn off world-seen in
>gettysurg.g, you should be able to reproduce the problem.

OK, I tested Gettysburg with no-see-all no-world-seen, and could confirm
the bug. Moreover, those unseen cells that should have been revealed by
daylight also fail to be revealed even if the unit moves into an adjacent
cell. Other unseen cells are uncovered as soon as the same unit enters
within vision range. Strange. It suggests some kind of side view area layer
corruption.

This is a kernel bug since it happens also in the Mac interface. Which is
consistent with a problem in the side view code. It further seems that
AI-controlled units that end up in this situation (standing adjacent to an
unseen cell) become confused and unable to move at all.

>It happens on a cylindrical map when the "reach" of an advanced unit
>stretches beyond the left or right edge of the map, so that it covers
>cells on the other side of the map.  It's very similar (at least in
>appearance) to the recently-fixed bug that caused problems when an
>advanced unit's reach included non-existent (off-the-map) cells.  It has
>nothing to do with mobility.  Although I don't think that any of the
>games that use advanced units (except for the experimental "Colonizer")
>use cylindrical maps by default.

This seems to involve plotting of resource icons for cells that are used by
the advanced unit. This was also the case with the other bug, and even
though I fixed it, problems at the dateline (see below) appear to trigger
the same thing again.

Consistent with this, the displacement seems to affect only the current
unit of your own side, which is the unit for which resource icons are being
drawn. The icons are drawn one cell to the northeast of where they should
be, but only in those cells that are west of the dateline.

>I'm referring to the line on the coordinate plane where X=0.  I think
>that it's equivalent to the dateline at a 30-degree angle, with the
>lower end on the lower-left corner of the map.  On the earth-1deg map,
>this area contains Alaska and the western-most part of Canada (although
>this problem with connections appears to be inconsistent with hand-drawn
>maps such as earth-1deg).

X=0 is what I mean when I talk about the dateline. There are still some
problems with it in the tcltk interface. These problems have been fixed in
the Mac interface, but it took a lot of work. I could confirm this bug,
too, and as expected it is tcltk-specific. I'm willing to bet that the same
underlying dateline problem is causing both this bug and the one affecting
advanced units. It is the same displacement of one cell to the northeast in
both cases.

Hans



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]