Guys,
please be reasonable. First of all, let us just stick to the problem at
hand. As I haven't been following it closely, I nevertheless give you some
of my thoughts:
- Let us forget 2.95.*. This for two reasons:
+ we should always target the future, never the past (remember we are
switching to a new branch of development, hopefully to last for some time!)
+ let us not waste effort on two versions.
+ or is there any reason to not use 3.*
+ hint: 3.* is not "new", right, we are at 3.2.* at most systems, maybe
already even 3.3.*
- We must support consCall and deconsCall! And even Alice-ignorant people
should be eager
to support ;-) Why is that?
+ Alice exercises the Mozart engine in many interesting (and quite
different from Mozart)
ways.
+ by having Alice as a user of Mozart, the stability of Mozart itself
increases
+ increased stability should in particular excite Kostja ;-)
+ it is not such a big deal.
And now for the real argument: the Alice people have contributed a large
deal of development
and fixes to Mozart, killing them off is so crazy it almost makes me cry.
I think we
are not exactly in the position to drive away some of the last people
standing! It is
as always: you give, you get!
Just common sense
Christian
"Konstantin Popov" <kost@sics.se> wrote in message
news:200309090717.h897HAG10732@sinuhe.sics.se...
>
> > >> What I'm missing from your list is the issue you brought
> > >> up about the consCall and deconsCall instructions. Their
> > >> presence in official Mozart releases is essential to the
> > >> Alice project. What is the current status of this issue
> > >> and what are your plans?
> > > It is not on my list because currently (a) I was not asked to handle
> > > this issue [when I was drafted for the job], (b) I do not feel that
> > > the Mozart consortium as a whole expects me to do so, and (c) it is
> > > not one of my own priorities.
> >
> [ ]
> > I have been taking care of them since 2001). I hardly think that
> > entitles you to speak on behalf of "the Mozart consortium as a whole"
> > :-) In any case, as a release manager you have to tend to the
> Again, the only thing I said was that I have not received *any*
> pressure to handle the issue. Needless to say, that can of course
> change.
>
> > interests of all parties, whether they relate to your interests or
> > not, and that should include the Alice people.
> of course we should respect interests of everbody. And if the
> majority will tell me to handle this, then I'll proceed with it.
>
> Again, I was asked "what are the plans", and I reported the status,
> trying also to explain why, so it wouldn't sound as harsh as it could.
>
> >
> > > Now, I see no solution (still having invested some rather significant
> > > amount of my *own* time!) and, up until now, did not see anyone who
> > > appeared to care
> >
> > Here, I have two remarks:
> >
> > 1. I cared and I tried to help. you can check the list if you forgot.
> > 2. you never reported your findings to the hackers list. so, you have
> > only yourself to blame for the lack of involvement.
> well, since I got the problem in the form of a CVS-log with the
> accompanying slow-down, I responded the same way.
>
> The summary: the problem relates to register allocation [in gcc 2.95*]
> so the 'PC' gets spilled out of the register. I did not find any
> handles to control that compiler behaviour, nor my attempts to rewrite
> the code of instructions was successful.
>
> >
> > > Also, the issue looks to be non-existing with gcc 3.*
> >
> > In that case, shouldn't the answer be: "yes, consCall and deconsCall
> > will be supported in the 1.3.0 release"?
> not yet:
> (a) we have yet to decide that we switch to 3.* for building the release
> binaries
> (b) we did not decide to drop the 2.95.* support
> (c) formally speaking, I did not see any approval of the stuff by the
> Consortium altogether. I am *investigating* that at the moment.
>
> >
> > > I hope I was objective and informative.
> >
> > Perhaps this was not your intention, but it sounded very arrogant and
> > dismissive. From where I stand, it makes the situation all the more
> > disheartening.
> thank you for sharing your opinion with us.
> I can only confirm that the "arrogant and dismissive" was not the
> intention.
>
> Now, regarding this issue, from now I accept only comments in the form
> "we, partner X, are interested in this stuff". I'll process results
> and deliver the summary.
>
> Cheers,
>
> --- Kostja.
> -
> Please send submissions to hackers@mozart-oz.org
> and administriva mail to hackers-request@mozart-oz.org.
> The Mozart Oz web site is at http://www.mozart-oz.org/.
-
Please send submissions to hackers@mozart-oz.org
and administriva mail to hackers-request@mozart-oz.org.
The Mozart Oz web site is at http://www.mozart-oz.org/.