🚀 go-pugleaf

I'm not sure I made myself clear.  Let me put this in terms of UNIX:

- Have you ever thought about what to do when calloc() returns an error
because it can't find more memory?  Does calloc() even return, or does
it crash your program on UNIX?  

- Have you ever thought about what to do when your program needs to
write a temporary file, but there's no space in /tmp?  No space
ANYWHERE on the disk?

- Have you ever tried to write a UNIX program that *didn't* wire
constants into the size of the C data structures?  This is generally a
no-no if you want your program to work in unusual situations.

- Have you ever tried to use a UNIX command for something it wasn't
intended for?  Very often, it will crash because some static array in
the C code overflows.  And to handle the "most cases", many programs
size their data structures far larger than necessary.  The result is
wasted memory and still the software can be flakey.


You may argue that UNIX is easier to program.  That's because you're
probably writing lousy code.  Sure, you can write lousy code on
the macintosh and life is just as easy.  Never assume you run out
of anything, and your code will be truly *simple*.

The Macintosh is built so that you implement contingencies when
resources get low.  I believe that many parts of UNIX (stdio) are not.
They just assume it's fine to crash the process.

Remember: EVERY computer system, EVEN WITH VIRTUAL MEMORY runs out of
EVERY resource.  If you didn't realize this, now you know.

Don Gillies {ihnp4!uiucdcs!gillies} U of Illinois
            {gillies@p.cs.uiuc.edu}