Re: [freesci-develop] Please test: Updated FreeSCI packages available

[Christoph Reichenbach]

Hugues,

> A few notes on the 0.6.1b package (with VSNET 2003):
> 
> Needed to remove inline keywords from fmopl.c in order to compile (inline 
> is C++ only)

I guess I can import the CVS fmopl.c now.

> The following obsolete .c files are still in fsci_dll.vcproj and should be 
> removed:
> - heap.c
> - simplesaid.c
> - wrapper.c
> - resourcecheck.c

Weird, I could have sworn that I had fixed these with Matt's updated
vcproj file (maybe I only copied that to CVS? not sure...)

> VERSION preprocessor definition is incorrect (in both sci_dos.h and 
> sciv.vcproj)

Since no one is maintaining the DOS port at the moment, sci_dos.h
doesn't really matter all that much. I've tried to update these things
by hand, though (fortunately, they're using XML files to store that
stuff).

> ----------------------------
> RUNNING SCI1 GAMES USING SDL
> 
> major crashes in sdl_driver.c when running SCI1 games, right at startup:
> - sdl_grab_pixmap, ln 755: free(pxm->data);
> 
> crash always originate from kAnimate or kCursor
> 
> commenting out the line helps and the games are sometimes playable (SQ1 
> works best)
> 
> also, cursor is not drawn when mouse is moved, but is drawn when clicking
> 
> very often the crashes I see are cursor related.

Does this also affect SCI01 games? It's the first time I've heard of
this; do you have any additional information as to why it is crashing
(trying to read out of bounds or something)?
  Which version of SDL are you using?

> ----------------------------
> RUNNING SCI0 GAMES USING SDL (LSL3)
> 
> arrow keys don't work in menus
> in some screens (like first screen with the binoculars), larry can walk 
> about 10-20 pixels lower than he should

That's the default port boundaries bug. Right now, the port boundaries
are hardwired for Jones16 and SCI1 games, i.e. all SCI0 games will have
incorrect boundaries (should be off by precisely 10 pixels). We're
probably not going to fix that for 0.6.1b, as a correct fix would
require an infrastructural change (while minor, this is not something
I'd like to add right now).

  I'll send out another e-mail when the updated 0.6.1b package is
available.


-- Christoph