[Scummvm-devel] [Scummvm-cvs-logs] SF.net SVN: scummvm: [23747]scummvm/trunk/tools

[Max Horn]

Am 26.08.2006 um 01:53 schrieb Travis Howell:

> From: "Marcus Comstedt" <marcus at mc.pp.se>
>>> Exactly why are file sizes been added? we already check the md5  
>>> checksum,
>>> to
>>> confirm an exact match.
>>
>> Because if the file size differs, we can conclude that the file
>> doesn't match without computing the md5 (or even opening the file),
>> thereby saving a lot of time?
>
> Well it is already planned to reduce the file size limit used for md5
> checksums to 128KB for the SCUMM engine.

Yeah, but 128kb > 0 bytes, isn't it?


> If the file size details are required to reduce file access, we  
> might as
> well just update the md5 checksums with an even smaller file size  
> limit.

Ah... but even 1kb is more than 0 bytes, isn't it?

> Since the md5 details have to be updated in either case.

Who says that? Did you notice all the "-1" entries indicating that  
the value is unknown? Having the sizes is purely *optional*. In fact,  
right now, we don't use them at all. In the future, we *could* use  
them, if we want to (limited to those cases were they are known,  
i.e. != -1). As it is, I am simply adding some data to a table. What  
we do with it is optional.

Now, why do people on this list always get so upset over minor  
details that cause *no harm* at all (adding additional and purely  
*optional* information to a table, reducing the amount of data the  
MD5 has to be computed of, etc.), while for the real meat, everybody  
keep silent, I wonder?


As for the email addresses: They are encoded. If you believe the  
encoding is not safe enough, we can improve it. But yes, I do think  
that it's absolutely vital to keep that information in a central  
place were all of us can access it.


Bye,
Max