[Scummvm-devel] kMD5FileSizeLimit = 1024 * 1024

[Jamieson Christian]

Max Horn wrote:
> Am 04.08.2006 um 20:56 schrieb Stuart George:
>   
>> what if the there is an update > 100mb with two versions of the
>> same game? you wont find it. thats why I said it should be all, or  
>> nothing.
>>     
>
> Hu? I don't understand what you are trying to say here? What updates?
>   

This line did make sense to me, so I will attempt to clarify, since I 
did see it as being potentially valid.

Let's say a particular game has a couple variations out there, 
variations that we have to deal with in different ways, such that we 
need to be able to uniquely identify which variation we're looking at 
when we scan a directory. Furthermore, the variations to the file(s) we 
scan only occur somewhere beyond the 1-MB (or 128-KB, or whatever) limit 
of what we're scanning. In that scenario, we would come up with a 
fingerprint that is identical for both variations, even though they are 
different and have to be dealt with differently.

It is on this premise that Stuart condones scanning the entire file. I 
agree with Fingolfin, that doing so is not practical. However, I do 
think the question is valid: how likely is it that we will need to 
distinguish between two variations of a game whose only differences 
occur past the limit of our fingerprinting?

If the answer is, "It could be an issue," then my first thought would be 
to fingerprint a portion at the END of the file rather than at the 
BEGINNING. Changes that happen in the middle of a data file are very 
likely to cascade, resulting in a significantly different fingerprint at 
the end of the file, or even a completely different file size. (Do we 
take into account total file size when fingerprinting? Should we?)

Thoughts from the sidelines....

Jamieson630
ScummVM developer
Retired (no, really!)