On 02/24/15 01:18 AM, ken mays via oi-dev wrote:
Nikolai,
Hi, my name is Nikola (without 'i') ;)
As far as Intel DRM work for illumos, that is proposed for GSoC 2015
development which ends Q4 2015.
It is great to know.
It is then unknown what people with Intel graphics will do with their
hardware, that want to install on and test Hipster in meantime?
Also removing support for older graphics is never discussed anywhere as
I know (we were just hit by update and people started reporting
disastrous effects to their desktops). It is just done without asking.
Maybe reconsider of leaving supported Xorg and drivers in their place or
with an old/new name so it could be auto recognized and used?
As far as Martin, there is Martin's OpenSXCE distro. He currently
mentioned a new release.
There's no source code that follows OpenSXCE releases, correct me if I
am wrong.
Also, there is Tribblix. Both distros are desktop-centric and support
older Intel GPU chipsets (but not the newer ones). There is still the
Hipster-20141010 ISO release you use today.
So you suggest moving from OI Hipster to other distro if wanting to
continue using existing hardware?
It is funny to call Hipster ISOs - releases, because it could be said
that then publishers that are installed with them could receive updates
and support?
Which I am not sure it right, because actual reason ISOs and freezes of
Hipster exist is because Hipster re-build all packages all the time and
because updating takes longer and longer, there is need for starting
from blank history of updates, so that installing packages dont' consume
so much RAM and takes so long, not tend to give support?
Also ISOs are there in Hipster to be able to install it on bare
metal/hardware , because updating from /dev to Hipster last time worked
almost 2 years ago and no one actually care to have release that could
update from /dev.
As far as Hipster, there are the supported ISO releases. The
Hipster-2015 IPS repo is a moving target (considered 'dev', bleeding
edge, or 'hair on fire' development) not for general production or
day-to-day operations. There are other snapshots that may produce
'tested' package releases - which is really what you are discussing.
True development releases are never for the general public (aka users)
- if untested by 'QA'.
We need then something that actually IS for end users as a product of an
effort.
Hipster is not that kind of product, and yes, there is need for tested
product.
As I used Hipster for exactly everyday work for a long time (and
reported all bugs I could on IRC and lists) , I must confess I have seen
accumulation of bugs without fixing.
Is it better "updating everything and not caring for introduced bugs" or
"have working AND development releases separate from each other", with
clear intent that updates work - is maybe better solution?
If you are tracking new releases from Hipster Dev which happens to
plunge your desktop environment into a dark abyss, remember the rule:
Either press the panic button or don't do it!!!
That would not be the case IF Hipster would have "entire" package
numbered on every update,
so that user can update to specific point in time in Hipster updates,
to debug what change break what already working function of an OS.
You always get "latest" Hipster , with no way of turning a step back and
think if update was introducing more bugs of fixing existing ones.
With Boot Environments on top of ZFS, we using and testing OI have best
possible solutions for testing it is just that we don't use what we have
to the full.
The drop of support for the older Intel driver is by design of the
newer X infrastructure - not us. We are moving to the newer Xorg Intel
driver(s) and when Intel DRM for illumos is in place - we will strike
the iron while hot...
So as I understand, Intel support for OI Hipster will never come back,
and there is nothing to do, but to roll separate distro or use other
distro but OI Hipster so that graphics could continue to work? (Or wait
for 2 years for illumos DRM to land...)
I think it is _much_ more important for OI in general to actually
support present and existing users and hardware at a moment.
It is all great to have newer hardware support but I think that removing
driver and Xorg support for existing widespread grahics hardware, will
toss away new users from testing "bleading edge" development releases,
and is disastrous to the future of OI Hipster project.
I suggest there could be solution of leaving Xorg and driver support
where it is for existing intel graphics users. I don't see any benefit
of tossing it away.
Thing is, I have been wrestling with Hipster for a very long time now
and I don't see a point of it anymore.
There are no formed teams, there are no pointed and publicly discussed
goals for a project.
_______________________________________________
oi-dev mailing list
[email protected]
http://openindiana.org/mailman/listinfo/oi-dev
