Subject: RE: SIGSEGV error when using big float matrix under pvm
To: 'Peter Seebach' <seebs@solon.com>
From: Graham, James <James.Graham@Schwab.COM>
List: current-users
Date: 05/19/1997 08:34:40
That's how C was written and implemented most places, though; 
there has never been anything which was strictly an N-dimensional
array!  At least this is how it was explained to me by my teacher
and countless compiler hackers when I was taking a C course
(oh, to be young and foolish again):  When an N-dimensional
array is allocated, it starts at address X and proceeds to allocate
space up to 
X + (numelem(D1) * numelem(D2) * numelem(D3) * ... * numelem(DN)) -1.

If ANSI is going to proceed to completely rewrite all the semantics
of C, they might as well call it Pascal, rewrite it as such, and declare
_that_ the standard and leave C alone.

ANSI/POSIX/C9X/whatever-flavour-du-jour of C these days
is all well and good, but I think they're trying to change too much
too fast.

> ----------
> From: 	Peter Seebach[SMTP:seebs@solon.com]
> Sent: 	Saturday, May 17, 1997 9:37 AM
> To: 	current-users@NetBSD.ORG
> Subject: 	Re: SIGSEGV error when using big float matrix under pvm
> 
> 
> Just a quick note, the trick with allocating a 1D array and passing
> it to a function expecting a 2D array is *NOT* strictly conforming,
> and is generally well into "bad mojo".  Ditto the trick with
> treating a 2D array as a 1D array.
> 
> -s
>