Re: Windows Specific Fixes

[Neil Hotmail]

> This is an updated version of the .dsp project file (shall I resubmit this
> file every time it changes?)
| Getting a new version whenever a change is required is definitely usefull
but
| it may be better to find a way to create those patches automatically.
It takes about two seconds to fix, so its not too much of a problem, I'll
try a perl script - but I've only got a small bit of Perl experience.

> It is also a fix to #include <algorithm>, I'm not sure if this needs to be
> #ifdef WIN32'd, or moved into a different header file, but it certainly
> doesn't work on windows as is.
| It's ok to include it without #ifdef - it's very likely that it's included
in
| Linux anyway by some header.
It may be called something else on Linux - thats the C++ standard for you.
If this is the case then it might be best to include it in an equivalent
place to where it is on Linux, i.e. in the project file/windows headers or
wherever.

> It also patches the display to not update moving backgrounds under
> windows, this stops the slow updating - but still keeps the processor at
> 100% (I imagine it turning out 1000's of frames per second of the same
> thing). Keyboard input still suffers (not as much as before), maybe a
> separate (higher priority) thread is needed for input?
| Pingus doesn't use multiple threads but I don't think that would be the
right
| approach in any case. If there are severe performance problems in Windows
| they ought to be fixed.
| BTW since Pingus runs in an event loop I'd guess the processor to be at
100%
| all the time. Games usually aren't very cooperative about sharing the
| computer with other apps.

But they should be - in a windows message pump processor useage is about 2%
tops, its only when you do a greedy pump that it exceeds this. The taking
100% processor is the problem, not that it is doing anything with it. The
fix I supplied does fix the performance (as mentioned with ClanLib and alpha
blending etc.), but it opened up the problem of the game taking most of the
processor of the input system, which results in the game going slower (or
less responsive)

This fix should probably not be applied without further testing - the
current Pingus's copy on my machine won't play levels (problem posted to
list) so I can't test the levels aspect of it - which is obviously quite
important.

Ginge



Bye David



_______________________________________________
Pingus-Devel mailing list
address@hidden
http://mail.freesoftware.fsf.org/mailman/listinfo/pingus-devel