RE: [MiNT] New kernel features

[joska@nuts.edu (Jo Even Skarstein)]

On Wed, 1 Dec 1999 14:01:23 +0200 (EET), Martin-Eric Racine wrote:

> > ME, you jump to conclusions,
>
> Example?

I don't think that's necessary, just go through your messages the last
two months or so. The "MiNT still needs BigDOS"-thread is a good point
to start.

> > you refuse to clarify your statements (e.g.  you talked about
> > "breaking GEMDOS compatibility", but not been able to tell us
> > exactly *what* does this,
>
> I have mentioned the Compendium several times.  It documents MiNT
> (as a part of the GEMDOS functions) and what is documented there
> should be observed.  This includes signals.  What is so unclear?

When asked what GEMDOS-functions you talked about and what exactly in
the new release-kernels that broke compatibility you only answered
"the Atari Compendium", and nothing more. Fact is that you can't find
any new features that breaks anything - if you do please inform us.

And what about the signals? The signal in question (SIGPWR) didn't
exist in MiNT, it used the unused signal-slot and no existing signals
has been modified or altered in any way. How can that break anything??

> > complained about the new error-codes and claims that they
> > render MiNT useless on your machine, but you've not told us
> > what breaks.)
>
> I have very specificaly said that it rendered ZSH useless,
> because it returns the wrong replies.  Later, someone finally
> explained that a typo in 1.15.4 was at fault.

You specifically said that you couldn't use 1.15.4 *or later* on your
Stacy because of (among other things) "new error messages mapping". You
also said "1.15.4 broke several application's behavior, on every machine.
That's predictable, when return codes are not what's expected
anymore.  I have reverted back to 1.15.3 on the TT for the same
reason.". No mention of any specific problems here either, if you did
in another message then atleast I (and Guido it appears) missed it.

> When too many new fatures are introduced at the same time,
> tracking down _what_ broke in a new beta becomes looking for the
> needle in a haystack.

Not with proper documentation. Reading CHANGES and *understanding what
it means* is always helpful when you debug *beta* kernels.


/*
** Jo Even Skarstein    http://www.stud.ntnu.no/~josk/
**
**    beer - maria mckee - atari falcon - babylon 5
*/