RE: [freesci-develop] FreeSCI release plan

[Hugues Valois]

Hello!

I have started testing stable on win32.  I built it from CVS using Visual 
C++ 2003.  I am testing with all defaults options.

First a few questions:

1. Most of the entries in bugzilla are old.  Is it worth trying to reproduce 
some of the bugs in there?

2. Is the game detection by hash code still used?  The console outputs a 
message with the hash code when it is not detected.  Looks like the new 
detection technique looks for the version in the sierra interpreter 
executable.  Should I report the missing hash codes?


And now for the results I have so far:


Sierra Christmas Card v1.04

Could not determine interpreter version; hash code 04B0B081
1st scene with gift plays music and works fine
2nd scene with fireplace appears for 1-2 seconds and then goes to 3rd scene 
(that's a bug)
3rd scene with angel does not have music and never ends (additional angels 
never appear, which is also a bug)


Fun Seekers Guide v1.02

Interpreter version: 0.000.506 (by executable scan)
Could not identify game by hash code: E4A3234D
Did not notice any problems


Hoyle 1 v1.000.104

Interpreter version: 0.000.519 (by executable scan)
Could not identify game by hash code: 85AFE241
Did not notice any problems (played a few games). It's true that memory 
usage goes up a little bit during a game (by about 2 megs), but it comes 
down when selecting a new game.  Never tested with previous versions of 
FreeSCI, so I'm not sure if it has improved or not.  It is certainly very 
playable on my machine.


Hoyle 2 v1.000.011

Interpreter version: 0.000.572 (by executable scan)
Could not identify game by hash code: E0E070C3
Did not notice any problems (played a few games)


Iceman v1.023

Interpreter version: 0.000.668 (by executable scan)
Could not identify game by hash code: D0B8794E
FreeSCI crashes when playing volleyball, when ball goes into water, after 
dismissing the "Oops there goes the ball" message.  I was able to reproduce 
4 out of 6 times.
Here's the console output:
Send: No object at 0000!
pc=b571 acc=0000 o=bb4e fp=5f20 sp=5f2a
prev=0 sbase=5f06 globls=0e04 &restmod=0
Step #3869027
b571: [W] send 04
 <invalid>::x[Kernel Oops in file \freesci\src\engine\vm
.c, line 1290: Heap address space violation on read
Kernel Oops in file \freesci\src\engine\vm.c, line 1290:
Heap address space violation on read
Lookup selector functions: Method selector offset 0000 of object at 0000 is 
inva
lid
]()


King's Quest IV v1.006.004: Int #0.000.502

Played only a little bit.  Noticed several instances of a priority problem 
where objects appear in front of Rosella when they shouldn't (bushes, rocks, 
etc.).  This seems to be the same problem as the existing bug 402.
Can't reproduce bug 2778 (waterfall screen crash).


The Colonel's Bequest v1.000.046

Interpreter version: 0.000.631 (by executable scan)
UNEXPECTED INCONSISTENCY: Hash code 94EA377B indicates game version 
0.000.685,
 but analysis of the executable yields 0.000.631 (for game '(null)').
My 5.25 floppy disk labels indicates int #0.000.631
Played through act 2, did not notice any problems.
Can't reproduce bug 346.


Astro Chicken

Interpreter version: 0.000.453 (by executable scan)
Could not identify game by hash code: 28543FDF
Did not notice any problems


Various

Bug 2785 could be closed, as the game selection screen no longer appears 
when no games are found in the config file.


readme.win32 lists 3 sound servers for win32 (win32 polled, win32 event, sdl 
threaded).  Looks like the win32 event sound server is no longer an option.  
It's commented out in the source, and when I tried to compile it in and use 
it, it was horribly slow and I couldn't hear any sound.  If it stays 
disabled for the release, the documentation should be updated.


I couldn't get the DirectX driver to work properly.  I get the following 
errors:
\freesci\src\gfx\drivers\dx_driver.cpp(384): 
(dx_state->pDevice->CreateTexture(dx_state->xsize, dx_state->ysize, 1, 0, 
D3DFMT_P8, D3DPOOL_MANAGED, &dx_state->pTexPrioritys[i])) from 
dx_init_specific (hr=D3DERR_INVALIDCALL (0x8876086c))
\freesci\src\gfx\drivers\dx_driver.cpp(384): 
(dx_state->pDevice->CreateTexture(dx_state->xsize, dx_state->ysize, 1, 0, 
D3DFMT_P8, D3DPOOL_MANAGED, &dx_state->pTexPrioritys[i])) from 
dx_init_specific (hr=D3DERR_INVALIDCALL (0x8876086c))
\freesci\src\gfx\drivers\dx_driver.cpp(881): (dx_state->pDevice->CopyRects( 
sbuf, &roldPnt, 1, dbuf, &poldPnt)) from show_pointer (hr=D3DERR_INVALIDCALL 
(0x8876086c))
First-chance exception at 0x10056120 (fsci.dll) in freesci.exe: 0xC0000005: 
Access violation reading location 0x00000000.

Sometimes even after these errors the game will start.  The cursor is 
usually a mess, and things don't work for very long.  I was wondering if 
anyone got the DirectX driver to work for them?  I have an ATI Radeon 
9200SE, and I haven't had any problems with other programs that use 
DirectX/Direct3D.

bye,
hugues