Maybe I'm missing the obvious but...

[hutch--]

oex,

Thanks for finding this one, it is a fault in to library algo. Attached is a replacement that has exhaustive proof of byte range 0 to 255. I would be inclined to have a look at the thread JJ started in the Lab as it has a number of faster algorithms. I would make sure they handle the byte range exhaustively but you will get faster results using one of them if they test up reliable.

Dave,

looking at the code - that thing seems a little silly to me - lol
but, the -1 is returned as an error

I imagine it would if you are not familiar with constructing finite state machines. When I put this algo together some years back there was nothing as fast as it but its basically a technique for word testing rather that two character numeric representation as in a binary string.

[oex]

 :bg

[dedndave]

lol - sorry, Hutch - didn't know you were the author
ya hafta stick your name in there so we know who wrote it

[hutch--]

Thats easy, if it hasn't got a name on it, I wrote it.  :P

[oex]

Is there a speed difference in jmp @F vs jmp Label?

[dedndave]

no - it just tells masm to use the next "anonymous label" instead of a specific one

[oex]

awesome ty :)