Re: AIX I/O failures

[Mark Mitchell <mark at codesourcery dot com>]

>>>>> "Phil" == Phil Edwards <pedwards@disaster.jaj.com> writes:

    Phil> The major difference from this example and what you first
    Phil> posted, and what my problem is with the originally proposed
    Phil> solution -- and maybe what Gaby's disagreement is too,
    Phil> although of course I cannot speak for him -- is that here
    Phil> "place" is a character array, and everywhere else "place" is
    Phil> a Something. 

Right -- that's why my code isn't standard-conforming.

    >> The only difference is that there's not the subsequent ODR
    >> violations that emerges in my code.

    Phil> You'll have to remind me what ODR stands for, though.

One Definition Rule.  It's the thing that says "you can't say `cout'
is a character array over here and then say it's an `ostream' over
here".  It's why I'm being a Bad Boy. :-)

However, the important thing is not that we deliver code that is
*written* in a standards-conforming manner -- it's that it *work* in a
standards-conforming manner.  You already use GNUish extensions in the
source, for example, which makes the code nonconforming, too.  But,
users can't detect the use of those extensions when they use G++, so
by the as-if rule, y'all have written a conforming library.  Yay!

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com