Jesper Harder <harder@ifa.au.dk> writes:

> Here's a patch.

Great, thanks!  Added.  I guess you have signed papers?

> It has some limitations -- it doesn't:
>
> * handle multipart binaries, 
> * check the crc32 checksum (it didn't seem very straight-forward, when
>   you've only got 28 bit ints), 
> * support an external decoder (the elisp decoder wasn't too slow, so I
>   didn't bother)

I expect that committing this will make these limitations magicly
vanish over time.  Perhaps you can find out if the attached crc32.el
will work.  I guess the main problem is getting it assigned to FSF.

>> but supporting encoding is a bad idea IMHO.
>
> I agree.

Except a better, MIME-aware, version of yEnc.