Python and Industry, IBM I'm afraid

[Aahz Maruch]

In article <946evt02mr at news2.newsguy.com>,
Alex Martelli <aleaxit at yahoo.com> wrote:
>
>  [AIX 4]
>
>As I recall, its main
>"here's our horrible surprise of the day for you" feature was a malloc
>that would never return 0 even if you asked it for FAR more memory
>than you had around -- rather, the program died horribly later when
>it actually tried to USE the memory it THOUGHT it had allocated... a
>"designed-in feature", they adamantly insisted, NOT a bug (later I did
>meet some BSD variants with this wonderful 'feature' too -- never did I
>really understand WHY one would play such a trick to poor application
>programs...!); that accounted for well over half of our porting woes to
>that specific box/OS.  Later, they did relent and allow some ways to
>bypass the 'malloc never fails' feature.

You're not making it sound as horrible as it actually was: the OS would
actually start feeding you memory from the malloc'd chunk and your
program would run.  At some (possibly long-distant) point in the future,
your program would suddenly die a horrible death, possibly crashing the
rest of the machine with it if it had consumed all of virtual memory.
More than that, the memory allocation was *sparse*, so you could malloc
a gigabyte, try to use memory at the top and bottom of the range, and
everything would keep chugging along until you died.

I-had-no-direct-experience-with-it-but-lots-of-customer-complaints-ly y'rs
-- 
                      --- Aahz (Copyright 2001 by aahz at pobox.com)

Androgynous poly kinky vanilla queer het    <*>     http://www.rahul.net/aahz/
Hugs and backrubs -- I break Rule 6

A house is a machine to keep your cat dry.  --Aahz