[Opensim-dev] SQLite doen't need Community.CsharpSqlite.dll?

[Jim Williams]

I left OSG first because the grid became unstable for people using the
current code as 0.8 nodes started being used.  Has this problem been
addressed?

I left Metro and returned to OSG (rc2) because after jumping into an 0.8
upon returning you became a permanent cloud and had to relog.  Has this
problem been addressed?

I finally left OSG and put up my own grid on the current release because my
scripts will not work under 0.8, and I'm not going to waste time trying to
write to a moving target.  I figured I could jump to places and then relog
to get home.  Now I'm learning that you are planning on releasing crap I
won't even be able to jump into.

What you should do is stop 0.8 dead in its tracks and produce an
intermediate release that can survive the 0.8 protocol changes but which
does not actually introduce any of them.  This intermediate release should
fix enough bugs to induce people to move to it, but introduce no new
functionality.  Then, once people have generally moved to a release that
won't be badly effected by 0.8 you can try introducing it.  What you are
doing is subjecting people to your changes without any regard for their
needs or desires.  People cannot ignore your changes simply by not using
the new release, therefore you are obligated to make those changes in a
manner that does not negatively impact people who are slow to update to
your latest and greatest (for whatever reason).

You have no migration planning at all, and therefore effectively no
planning at all.  (And your QA sucks too.)

(I'm not trying to catch flies, I'm trying to kill cockroaches.  Enough
people are playing nice-nice, and the results are not good.  I'd like to 1)
see some of the programmers leave, and 2) see the procedures changed so
that it isn't the people who want to make new toys who set the development
agenda, such that 3) Opensim finally gets some QA.)



On Sun, Jun 8, 2014 at 8:40 AM, Shaun T. Erickson <ste at smxy.org> wrote:

> Jim,
>
> It could, until just a couple days ago, but that's been lessened
> significantly. Now, it won't let you teleport to it, if you're coming from
> older code, if the destination is larger than 256x256. That eliminates a
> huge problem, right there. You get a nice, graceful refusal of the teleport
> and no crash.
>
> Now, it is still true, that if you wish to teleport from a 0.8, 256x256
> region to a region larger than 256x256, you must be using a viewer that
> supports large regions. I don't think that is unreasonable.
>
> However, if it were possible for the large region to determine whether the
> incoming viewer is new enough to support it or not, then I would be in
> favor of it gracefully refusing teleports from viewers that do not. That
> would close the other vector for crashes due to old code.
>
> -ste
>
> P.S.: As the saying goes, "you catch more flies with honey, than you do
> with vinegar." Something to keep in mind.
>
> On 6/8/14, 7:56 AM, Jim Williams wrote:
>
>> What I do know is that 0.8 is giving people who do not use it grief.
>>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at opensimulator.org
> http://opensimulator.org/cgi-bin/mailman/listinfo/opensim-dev
>



-- 
No essence.  No permanence.  No perfection.  Only action.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://opensimulator.org/pipermail/opensim-dev/attachments/20140608/effb836a/attachment.html>