----- Original Message -----
From: "Lars Skovlund" <[EMAIL PROTECTED]>
To: "Rink Springer" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Saturday, January 06, 2001 3:41 PM
Subject: Re: SCI1.1 kernel calls


> On Fri, 5 Jan 2001, Rink Springer wrote:
Hi Lars,
> >
> > I'm working my way through the SCI1.1 decoding, mainly by looking at a
> > disassembly. Christoph says you know all SCI1.1 kernel calls... care to
> > share the list with me? I've found a lot of stuff now, but it *appears*
that
>
> I've attached a list from one of the latest SCI1.1 interpreters. Yours may
> be different (specifically, from KResCheck onwards).
I'll see what I can find.
>
> > about all functions are different. I've found some functions I know they
are
> > the same (Animate and DrawPic for example), but I think it'd be much
better
> > if we'd tell eachother what we know. I'll forward the email I've sent
> > Christoph to you too, it contains some very interesting information I've
> > found out.
>
> I'll try to address some of that below:
>
> Some of the opcodes are different due to changes in the script format. In
> SCI1.1 only the .HEP files are stored on the heap (and are thus accessible
> to scripts) - .SCR files are stored in hunk memory. Now, .HEP files
> contain mostly text and class headers (both of which were frequent targets
> of the lofsa instruction, along with the now obsolete said specs). As
> written in the FreeSCI docs, lofsa used to use IP-relative addresses.
> Since the actual code has now moved outside the adressable heap, lofsa
> needs new semantics. It's that simple.
I have already found that lofsa actually *is* different. As you said, it
only sets the accumulator to whatever argument was supplied.
>
> You're right about savegames starting with the sequence 0x8E 0x00.
> However, I think that you are jumping to conclusions when you say that
> they sync resources. Though being compressed, they are still just memory
> dumps - importing savegames is out of the question. This, BTW, was part of
> Sierra's reason for moving away from SCI.
Who says a resource cannot be a memory dump? I don't think importing
savegames is very useful BTW, but it *would* be nice if we knew what SCI
actually does when saving a game, because this _might_ help us uncover more
stuff.

PS. How did you find all those kernel func names? I have disassembled and
commented about all of SCI1.1 now... could you be in the #sci channel at
about 16:00 GMT?

>
> Hope this helps.
>
> Lars
>
Thanks!
--Rink


Reply via email to