[Scummvm-devel] SCI code breakage

[Walter van Niftrik]

Hi Max,

Max Horn wrote:
> OK. So if I understand right, then KQ4 was released in variants using  
> 0.000.274, 0.000.409 and 0.000.502?

It seems that at least these PC versions exist: 0.000.253, 0.000.274, 
0.000.409, 0.000.502. Most of them have probably been released on both 
3.5" and 5.25" floppies, and then there's at least one version for Amiga 
and one for Atari ST.

> If files in these versions are  
> different, we should be able to distinguish them, somehow (if all  
> files are identical, there wouldn't be a need to distinguish  between  
> them, after all...)
>   

Certainly, but one of the current problems is that we're only using 
resource.map and resource.000 etc, and this is insufficient to tell some 
games apart.  We need to analyze what files are used by the later games, 
so we don't end up with people sending us md5sums that we can't use. ;)

Also, we currently have entries such as the following:

                {"resource.map", 0, "55e91aeef1705bce2a9b79172682f36d", 
5730},
                {"resource.001", 0, "8b55c4875298f45ea5696a5ee8f6a7fe", 
490247}
                {"resource.002", 0, "8b55c4875298f45ea5696a5ee8f6a7fe", 
715777}
                {"resource.003", 0, "8b55c4875298f45ea5696a5ee8f6a7fe", 
703370}

It seems that the resource files all have the same resources at the 
beginning, leading to the same md5sum for the first 5000 bytes. As we 
also use the filesize, I guess that in practice this won't matter much, 
but I thought I'd mention it anyway.

Regards,

Walter