Re: kernel-2.2.1-undefined references.

[Benjamin Scherrey <scherrey at gte dot net>]

    Geez people lighten up! This is worse than the C++-in-kernel thread that
occurs in the linux list. I think we've identified what the issue is here
(right?). Now - who's the guy who makes the decision (Jeff?) and what's it going
to be?

    I think Linus and company are slightly aware of what the egcs team has
provided and I don't think your recitation was intended to clarify his mind. This
"discussion" is only wasting my time and every one else's who attempts to keep up
with this list.

    If the issue isn't completely understood by both sides yet perhaps Linus could
provide a couple of source examples (if he hasn't already - I haven't bothered to
read every tirade, err message in this thread) who's behavior has changed and an
egcs maintainer could either say "it will be fixed" or, at least, explain how one
might predict whether or not the code will inline as expected. The last I heard,
these computer things were still deterministic and truly random results were still
hard to generate...

    flame out,

        Ben Scherrey

craig@jcb-sc.com wrote:

> Which means, what, exactly?  That you therefore can claim some kind
> of *ownership* of the exact behavior of gcc in the presence of
> all sorts of extensions, regardless of the degree to which they're
> clearly specified and documented?  Sorry, you *should* know better.
>
> >There were tons of code generation bugs over time, but hey, I _expect_
> >bugs, and while compiler bugs are sometimes really hard and nasty to find,
> >I don't complain too much: I report them, and I tell people not to use
> >certain versions. I may be grumpy for a while because I spent ages looking
> >at a bug that didn't turn out to be mine, but I never get the feeling that
> >I should really dislike the compiler people.
>
> Given how you've treated the egcs developers lately, for daring to:
>
>   -  Offer you a free compiler
>
>   -  Offer you free upgrades
>
>   -  Offer you free responses to bug reports
>
>   -  Offer you free bug fixes
>
>   -  Offer you free improvements to unclear documentation
>
>   -  Offer you free advice as to how to improve your code to make
>      it more immune to likely changes to egcs code development
>
> I think we're *all* grateful you don't feel you should actually
> *dislike* us.