The manual page for crypt(3) indicates that it supports an extended
"setting" format with longer keys, etc.

In the version of crypt I'm working on outside the U.S. I've implemented
this as described, and now I'm writing verification programs to check
everything.  However, when I import my test program to sun-lamp and link
with its libcrypt, crypt() locks up.

	crypt("foob", "ar") produces "arlEKn0OzVJn.", as expected.

	crypt("holyhooplasbatman!", "_arararar") hangs.  No response.
	I expect it to produce "_ararararNGMzvpNjeCc".

Does anything actually use the extended crypt() functions?  Has anyone
attempted to verify its operation before?

I'm forwarding the query to the list in the hope of getting a response
from a U.S. person.  I _could_ go and delve in the source for sun-lamp's
libcrypt, but that would be crossing a legal boundary I'd rather leave
alone.  I'm trying to keep my code clear of such political problems...

Thanks

- David B.

& can you say "Controlled munition"? &

------------------------------------------------------------------------------