[DGD] objregwatch flagged

[Raymond Jennings]

On Fri, Jul 5, 2013 at 6:30 PM, Noah Gibbs <noah_gibbs at yahoo.com> wrote:

> He's right on standard etiquette.  Similarly, you asking Felix for advice
> isn't a problem, but it's useful to make it clear that you're asking for
> advice,


seeking advice on what is normal with the klib was indeed my intention.  I
also figured felix was the best one to ask because most of the code is
still his work.  Any deviation from normal would be a bug to chase down.

incidentally, at first I thought this might be to my 128 byte sector
experiment.  I had a similiar crash from that, and put it back to 512.
 However, this one happened on 512 byte sectors and after a cold boot
starting with 512, so the root cause isn't as obvious.

At any rate, based on the klib's code and the fact that nothing outside the
klib is even allowed to tamper with objregd's information, I'm convinced
that there is a bug.  The code seems solid, the objreg linking is guarded
by infinite rlimits which is usually a good sign of "must not fail".  The
fact that the klib does not provide a method to hotfix a broken list is
also hintful.

I'd like to thank everyone for teaching me something about etiqueete that I
was unaware of.  Social experienc eis hard to come by for me, sadly :<



approximately as a favor, rather than him being normally expected to do
> that.
>
> It's not bad to inform and see if he offers, but it's presumptuous to say
> or imply that he should be the one doing it, since ordinarily you would be.
>
>
> ________________________________
>  From: Raymond Jennings <shentino at gmail.com>
> To: All about DGD and Hydra <dgd at dworkin.nl>
> Sent: Friday, July 5, 2013 5:01 PM
> Subject: Re: [DGD] objregwatch flagged
>
>
> On Fri, Jul 5, 2013 at 4:13 PM, Kris Van Hees <aedil at alchar.org> wrote:
>
> > At the risk of stepping in the middle of an issue...
>
>
> You are, but that doesn't mean you are unwelcome.  As far as I know this is
> a public list and the issue isn't exactly a personal one that I know of.
>
> As I detail below you've actually touched on an area of developership I am
> rather naive about.
>
>
> > It is slightly
> > unreasonable to request the author of some code to guarantee correctness
> > so you can rule out any potential issues in the code you forked from
> > prior to ensuring your own code is correct.
>
>
> I did not know this.
>
> It is customary that if
> > you find a problem, you perform the analysis,
>
>
> This was a no brainer when the code is original.  When you start with
> someone else's code, it got fuzzy for me and I didn't see a problem with
> poking him for advice first, especially since the changes I made didn't
> seem to have anything to do with the problem I encountered.
>
>
> > and if the problem is in
> > your code, you fix it, and if you find credible evidence that the problem
> > is in the original code, then you report that problem (and often, still
> > try to provide a fix also).
> >
>
> Believe it or not I actually appreciate this kind of advice.  Collaboration
> with other coders is not really something I have a lot of experience with.
>
>         JMHO,
> >         Kris
> >
> > On Fri, Jul 05, 2013 at 03:37:15PM -0700, Raymond Jennings wrote:
> > > I was asking about your version that I forked from, which I assumed you
> > > were still responsible for.
> > >
> > > If your version is guaranteed not to do this, then I can narrow the bug
> > > down to either my own changes or a glitch in DGD :P.
> > >
> > > I do apologize though for not being specific.
> > >
> > >
> > > On Fri, Jul 5, 2013 at 3:16 PM, Felix A. Croes <felix at dworkin.nl>
> wrote:
> > >
> > > > Raymond Jennings <shentino at gmail.com> wrote:
> > > >
> > > > > Is there a case where the objregd list might legitimately be
> > corrupted?
> > > > >
> > > > > As in due to misbehavior/errors in other code, and not due to a
> klib
> > > > > problem or dgd problem?
> > > >
> > > > You forked the kernel library, you put out the announcement:
> > > >
> > > >     "After some wrangling with git, I have volunteered to take over
> > > >     maintenance of the kernel library."
> > > >
> > > > Now you're going to have to do your own debugging.
> > > >
> > > > Regards,
> > > > Felix Croes
> > > > ____________________________________________
> > > > https://mail.dworkin.nl/mailman/listinfo/dgd
> > > >
> > > ____________________________________________
> > > https://mail.dworkin.nl/mailman/listinfo/dgd
> >
> > --
> > Never underestimate a Mage with:
> >  - the Intelligence to cast Magic Missile,
> >  - the Constitution to survive the first hit, and
> >  - the Dexterity to run fast enough to avoid being hit a second time.
> > ____________________________________________
> > https://mail.dworkin.nl/mailman/listinfo/dgd
> >
> ____________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
> ____________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd
>