[darcs-devel] performance measurements of "darcs get", "hg clone", and "rsync -r"

[zooko]

On Jan 25, 2008, at 4:08 PM, David Roundy wrote:

> It also would require that we develop an entire ftp protocol including
> transmition of errors.

This part I don't understand.  You just define a protocol, like:

I send you a stream of patchids (patchids have a known length, so  
there doesn't need to be any terminating character or length  
encoding), and you send me a stream of patches, with each patch  
prefixed by its length in bytes expressed as a big-endian, unsigned 8  
byte integer.

This would be a rich enough protocol to do the job of an FTP server,  
but it doesn't need to be.

>   I think this would definitely be harder than
> learning haskell.  Which isn't to say it's not a reasonable problem to
> tackle, but rather that your lack of Haskell knowledge isn't an  
> excuse!

Heh heh.  To me, implementing network protocols is easy, and learning  
Haskell is hard.

Regards,

Zooko