Hi,

sorry for the late reply; my system load is a bit high ATM.

> i think the colors that FreeSCI uses in LSL2 are quite a bit off - at least 
> for the trees in e.g. room 71 (jungle, where one falls from the parachute).
> 
> In Sierra SCI, the trees are colored in mixture of
> #a8a800 + #000000 = #545400
> 
> In Freesci, it's
> #8a4500 + #633100 = #763b00

This might be a filling problem, but also note that FreeSCI (IIRC by
default) uses slightly different colors because it was believed that the
results look better.

Comparing the color values to the results of dither16 would probably be
worthwhile, but unlikely to help with the problem (if it really is a
filling problem).


[...]

> In the same scene/room, there are killer bees that are only rarely 
> triggered, while it's a lot more diffcult to evade them in Sierra SCI. 
> The control map is exactly identical in the original and FreeSCI.

This will be a bit difficult to debug, as it involves finding out how and
where these are triggered (which script checks for them etc).


> Room #81 has a problem with occlusion - one can walk behind the glacier in 
> the front right corner, but ego is drawn in front of it if one stands as far
> in front as possible. doesn't happen in SSCI. Again, the control maps are 
> exacly the same in the Sierra SCI and FreeSCI. the sierra.exe version is
> 0.000.343.

Might be caused by a bug I introduced recently and haven't had the time to
fix, regarding the priority/coordinate mapping.

> One correction for the debugging tutorial:
> the sequence to search in sierra.exe is only "01 00 00 00 01" instead
> of "01 00 00 00 01 00". With this version of sierra.exe there's a "21"
> at the end instead of "00". I think it should be mentioned that this
> is to be found towards the end of the executable.

That sounds useful, thanks for pointing this out!

> And if anyone is still with me: the scaling of the mouse cursor should
> be limited (in the X11 driver at least) to the size given by 
> xdpyinfo as "largest cursor". 

OK, thanks for pointing this out! XF86 4.0 IIRC has pretty much unlimited
pointer sizes (some tests with 6x5, 16x1 and 1x16 worked fine for me, at
least, so this might be an MGA feature), but I hadn't considered that X11
does not require this.

There are two approaches to fix this:

(a) Switch to software cursors in this case
(b) Scale down


Particularly in the light of remote access, (b) (which you suggested) 
seems to be the better approach. Note, however, that for cursors with more
than two colors we will have to make a similar, yet less straightforward
choice.


Note to self: XQueryBestCursor() seems to be the function that checks for
this.


llap,
 Christoph


Reply via email to