On Mon, 22 Dec 1997, Not for internal consumption wrote:

:  * That makes significantly less sense than a perl script (which I don't
:  * particularly like, either; there's no guarantee that systems' 'mknod'
:  * binaries won't do something behind your back that you can't correct
:  * for -- and I think i've seen systems do such things in the past 8-).

I didn't think about this.  A properly installed perl binary on any host
system will have a direct mknod() call that just calls the kernel...!  No
ned to call a mknod binary.

: I can't imagine that a mknod(2) call would do such things; such systems
: I would consider BAD (Broken As Designed).

I think cgd was thinking of execing mknod(8)/mknod(1m), not calling
mknod(2).  I think?  :)

: But if you write it in a simplistic way such that it didn't need to
: reference any {machine,platform}-specific #include files, what's the
: problem?  Sure you can't compile OUR mknod, but how about a minimalist
: version which can handle the -n major-bits -w devwidth-bits?

Well, why not a Bourne shell script that _does_ know what a given OS does
with its args to mknod(8)/mknod(1m)?

=====
===== Todd Vierling (Personal tv@pobox.com; Business tv@lucent.com) =====
== "There's a myth that there is a scarcity of justice to go around, so
== that if we extend justice to 'those people,' it will somehow erode the
== quality of justice everyone else receives."  -- Maria Price