Re: [DotGNU]Platform testers wanted

[Glenn Chambers]

On Thursday, January 9, 2003, at 01:08 AM, Dawkins, David wrote:

> The only thing I can see is to check that AddString() mallocs
> its own copy of "str", so that it's safe to do that "free(str)" call
> in AddString2().
>
> DavidD

Ding, ding, ding.  We have a winner!  As with all such errors, I didn't
realize this until I was getting dressed the next morning.  And, having
applied the correction, I get further.

This time, I get 'macro or #include recursion too deep' on lines 30-33
of include/malloc.h

The relevant lines are the declarations of malloc, calloc, realloc, and
free.  The common element is 'size_t'.

Reading include/stddef.h, I find the oddly compelling line:

#define size_t  size_t

which is not 'safe' in the GCC version used by my 1.5.5 NetBSD.

I commented that out, and proceeded.

I had also tried a variation of Gopal's double-quote suggestion, only
to discover that I botched it.  A second attempt is in progress as I
type; we'll see if it works.  <wait...wait...wait..>  Nope, dies the
true death.

Oh, well, one last try tomorrow.  Or, I might just bag it, and you
can document that "CPP=/usr/libexec/cpp gmake" is the 'right' way
to go...