Comment 31 for bug 536076

Personally, from a graphician's vantage point, I prefer the current
implementation. The presence/absence of the playercolor key in the conf
file has value in alerting/reminding the graphician as well as the
application to the presumed presence/absence of the mask files.

On Fri, Jun 18, 2010 at 2:54 PM, Nicolai Hähnle <email address hidden> wrote:

> I'm in favor of keeping it this way.
>
> The reason is that it avoids confusion with "ghost files". If you look
> at Chuck's commits, there are several ones where he changes the FPS at
> which things are rendered simply because some people end up with
> different versions of Widelands data files in the search paths. This
> causes the frame loading code to load files that are not supposed to be
> there (but end up in the search path by accident). This is very
> counterintuitive to people.
>
> Yes, and while not everyone maintains multiple versions of Widelands, the
search paths remain a wildcard that should be dealt with. (No pun
intended.) :)

> ...(In fact, given the above, it would make sense
> to *add* an additional key/value pair indicating the number of frames in
> an animation. Or to radically change the way the file loading works. Oh
> well.)
>
> IMHO I think parameterizing the number of frames in an animation is a
bandaid that still would not mitigate the search path vulnerability.