[sldev] Packaging the viewer for Linux distributions

[Argent Stonecutter]

On 12-Oct-2007, at 13:33, Ryan Williams (Which) wrote:
> Yeah.  Internally we've asserted that the combination of channel  
> name +
> version is just as unique as the md5, so that might be faster/more
> convenient as an index.

If you're going to binary patch it, I suspect you probably need to  
check the hash of the whole thing anyway. :)