Signature mismatch, thanks

[Wechsler]

OK, so it's looking very much like my HMAC_SHA1 implementation is at 
fault. Slightly surprising as (as I mentioned) it passes all the RFC 
2202 tests, but there's obviously something 'special' here.

The code's in PHP, which appears to *lack* an HMAC function, so I had to 
write my own - it's at http://www.ch3.org.uk/hmac.phps, with all the 
tests. I'm gonna go throw it at various PHP groups to see if they can 
figure it out.

Feel free to run the tests through your own HMAC clients, just to 
sanity-check the value that both Adam and the LJ server give me for the 
last test.

Thanks to all who helped,
	Wechsler