>
>So, it turns out that I screwed up. You asked if x/m had ever worked,
>and I said "No". This this might perhaps be true, it's sort of misleading.
>I'm afraid the confusion came because I thought it never worked in Mach
>either and I was too tired to go and look it up. More fool I.
>
>Code from Mach follows.

Ok, this looks like the same stuff I have at home. (scary what old code
you can keep on large drives these days).

>
>I don't care whether we use their implementation or not (though it might
>be best), but we should try to maintain UI compatibility unless there's
>a good reason not to.
>

Hmmm...This code pretty much does the same thing but is a bit more overly
complicated for no real gain. Looks like the output is similiar except they 
put a space in between each byte and 2 after every 4. 

i.e. (mach)

0xa1234567: 00 00 00 01  10

vs (mine)

0xa1234567: 00000001 10

I can make the other code space the same way without too much effort but
otherwise it looks ok.

James