Re: [OpenIndiana-discuss] [oi-dev] ANNOUNCE: gcc ported Firefox 37.0b3 now backported to snv_130 x86/x64, Should work on OI/Hipster/S11.[0-3]/S12 etc.

2015-03-26 Thread Apostolos Syropoulos via openindiana-discuss

 I want to be able build FF and and other things too, by myself
 locally, so I can contribute in area I can. So having binary-only does
 not help in that.

 How shall I understand this?
 A) You don't want any bins

I would like to say a few things related to this matter. Binaries
are OK but the problem is that you cannot commit yourself to deliver
binaries for every release. And that is something that people who
maintain OI do not like. They want to be able to build firefox and
deliver binaries to anybody interested.



 B) You want to contribute

What do you mean?

 C) In order to get into a position to contribute FF related work, rather than
 contributing and working on the src until it works, you expect _me_ to provide
 it in fully functional manner!??

 Does that make sense?
 

Nobody expects anything. I had compiled OpenOffice and immediately I released
the patches I had to apply so that others could build it. That's what 

people want. On the other hand if you want to make money out of this, we
can try to make found or something to help you financially. It happened in
the TeX community with luaTeX where people found money and paid someone
to write luaTeX from scratch. The final project is completely Open Source.

A.S.


--

Apostolos Syropoulos
Xanthi, Greece
http://asyropoulos.eu
http://asyropoulos.wordpress.com
http://hypercomputation.blogspot.com/

___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] [oi-dev] ANNOUNCE: gcc ported Firefox 37.0b3 now backported to snv_130 x86/x64, Should work on OI/Hipster/S11.[0-3]/S12 etc.

2015-03-26 Thread Nikola M

On 03/26/15 06:25 AM, Martin Bochnig wrote:
90% of the work I did since 2013 was closely related to amd64 
(x86/x64-only), rather than sparc. But probably it will take 25 
further years until somebody starts to figure this out.
Maybe people would better see things from your perspective if there are 
some code sharing between your project and others.

OI also supports SVR4 .pkg packages so this could be looked at as
rehearsal of making external products to Openindiana/illumos, although
there is also possibility of making IPS repository, accessible using
certificates.

It would have been stupid, had Sun completely burnt all bridges of backwards 
compatibility.
But that was before you started using Solaris. You didn't live through those times of IPS 
democracy in 2007.

I just yesterday found this page about turning SVR4 packages into IPS ones:
http://docs.oracle.com/cd/E26502_01/html/E21383/pkgsvr4.html
haven't tried it yet.

Openindiana also supports installing SVR4 packages (and S10/11)
I was always inclined to see IPS networked way as something more modern 
but maybe your views on SVR4 that it is modernized and fit for a 
purpose, could live inside subset of illumos distributions that use it, 
as an open standard?
Maybe your updates on using SVR4 packaging (is it been updated outside 
illumos?), could end up contributed inside illumos or existing SVR4 in 
illumos is already fit for use that can rival IPS?
I think that exchanging packages between distributions is already done 
with .spec files and processes to make pkgsrc, deb, IPS and possibly 
SVR4 .pkg packages?
Exchanging definitions of packages it through .spec files, to lower 
porting effort between distributions to a minimum. What do you think 
about those ideas?

I want to be able build FF and and other things too, by myself
locally, so I can contribute in area I can. So having binary-only does
not help in that.

How shall I understand this?
A) You don't want any bins
Bins are Ok - that is what ready-made distributions are for, only 
building process needs to be reproducible by users.
(e.g. get source - use same process of making package as in binary 
release, but only for that package - producing package as user).
Anyway that's something what most of free software licenses give as a 
right to every user of software. You know - what is open, stays open, 
and ability to make changes, fork etc.

B) You want to contribute

Of course , everyone to their abilities and level of expertize.
Yet no one can wider their knowledge without doing something and sharing 
ideas through changes and every contribution counts.

C) In order to get into a position to contribute FF related work, rather than 
contributing and working on the src until it works, you expect _me_ to provide 
it in fully functional manner!??

Does that make sense?
I think no one actually expects anything from anyone :) Nor asks. You 
decide it by yourself.

Producing something is never job for one person.
If more people are not involved in a production process organization has 
to be like that to easily include more people in. With more people, also 
burden of making something is lower on developer.


I think that is only expected is to do things in open manner, where 
changes to the code are visible, being able to be audited by users 
themselves, admins and distribution maintainers.


*There is much more then producer - consumer relationship in free 
software and open source.*


Valuating one's contribution through money is not directly connected 
with the level of involvement, it is more of a business model and 
marketing and position on the market, but that is clearly not the role 
of developer potentials of an Operating system/software distribution to 
think about those things.

Maintaining good name and good image is maybe good enough.

I think that sales and marketing are very much also needed in free 
software projects, to provide incentives (monetary support) for 
developers to do their work.
Also developers are not the only ones using final product, but people 
that are using it in the wild, so listening to their needs is crucial to 
supporting developers needs (of course with developers freedom of making 
new great things).
Also it should be noted that code and packaging and maintaining and 
support and testing and bug reporting contributions, also value , and in 
many cases are more valuable then money.
Money influx is in direct correlation not only to the greatness of the 
products, but also as the image outside world has it to it. (Are people 
positive and smiley, does process of including new people is open, is 
doing things transparent, does it fit with competition.)

That said, without source, you can not be sure what is inside, like it
was for chromium port we already had here.

How ridiulous is that very argument, and why won't you stop repeating it time 
and again??
Because it is right. it's how free software work. You get sources from 
somewhere, why people after you 

[OpenIndiana-discuss] rsyncd configuration

2015-03-26 Thread Sebastian Gabler

Hi,

I am trying to solve a problem that i have ignored for quite a long 
time. The issue is that messages are flooded with rsync permission 
errors, and that some files are not backed up properly. What I have 
found so far is the following:

- rsyncd is running as root
-the issue is the same when rsync is invoked locally, or from a remote 
host (I am running OI as a backup server calling other machines, OI and 
Ubuntu as clients). The call usually is rsync -azt --numeric-ids 
--timeout=600 --port ...
- the failing files all seem to have in common that there are no read 
rights on the o- bits. (i.e. 2.5K -rw---  1 root root 
ssh_host_rsa_key file is faling consistently)


I seem to be missing something really basic with the access rights here. 
My understanding is that the access rights of the user running rsyncd on 
the client will count.  Naively, why would rsync running as root not 
read a file root owns? Why is this going per o- permissions?
If I could make it go by the group at least, then still how could I 
solve the other aspects, i.e. the rsync user reading files that are 
owned by others like dladm:netadm? I understand as well that at least 
for the above RSA key I don't want to have anybody else but owner read 
the file. Is there still some RBAC trick or ACLs that could help me back 
up everything?
Once solved for OI, I will probably have to address the same on the 
linux machines.


Thanks for any hints, and sorry for the noob aspects of this matter if 
there are any.


Regards,

Sebastian


___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] rsyncd configuration

2015-03-26 Thread Bob Friesenhahn

On Thu, 26 Mar 2015, Sebastian Gabler wrote:


Hi,

I am trying to solve a problem that i have ignored for quite a long time. The 
issue is that messages are flooded with rsync permission errors, and that 
some files are not backed up properly. What I have found so far is the 
following:

- rsyncd is running as root


Check your rsyncd.conf file.  For example, one of my rsyncd.conf files 
starts with:


uid = nobody
gid = nobody


so that rsync changes its effective uid to 'nobody' before doing 
anything.  This is pretty common since rsyncd is often used in 
evironments with untrusted users.


On another rsyncd.conf file which is used to successfully transfer ssh 
private key files, I am using


uid = root
gid = root

Bob
--
Bob Friesenhahn
bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,http://www.GraphicsMagick.org/

___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] Hipster upgrade 18-feb-2015

2015-03-26 Thread Jonathan Adams
I got fed up one evening, and then I remembered working with Intel on my
old Solaris 10 desktop, and needing to set the resolution with
915resolution ...

I went back to my old files and found it, downloaded the newer version
from 2007, made some changes to it and got a working 915resolution for
Illumos that allows the modes to be set.

I then de-installed the intel driver from hipster, set the resolution and
have vesa working at 1280x800.

obviously not the perfect answer, but since the hardware acceleration on
the old intel driver (on the old intel card) wasn't the best, I have a
working, updated version of OI on my laptop, with no sticky's.

Jon

On 20 February 2015 at 09:14, Nikola M minik...@gmail.com wrote:


 On 02/18/15 05:03 PM, Jonathan Adams wrote:

 root@jadlaptop:~# pkg list -af | grep xorg-video-intel
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.99.917-2015.0.0.0---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.18.0-2015.0.0.0  ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.18.0-2015.0.0.0  ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.9.1-2015.0.0.0   ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.6.3-2015.0.0.1   ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.6.3-2014.1.3.0   i--

 hmm ... I do appear to be stuck at an old version of the intel driver ...

  Same (1024X768 fixed screen resolution) happens with intel 945 graphics,
 that worked OK till 2015.
 I think it also worked in 2015 before last update.
 Only, Hipster does not update 'entire' package on every update, so I can
 not update from 20141010 to 2015 prior to latest update of 2015. (destroyed
 it before another 2015 update). So 2015 is not usable currently for desktop.

 Also , managing screen brightness from GNOME applet fails in today's
 hipster-2015 , this is second blow to Hipster Laptop use since standby
 stopped working in Spring 2014 (also destroyed old Hipster BE where it last
 worked so can not go back to it to see what changed)

 When updating, even when /opt is cleaned and mounted and work like that
 right after that in 2014.1 (separate issuue on freezing display on logging
 as another user - I should report user's settings maybe),
 after update to 2015 it needs also cleaning /opt and rebooting again to
 boot right.
 Obviously update mess with installing something in /opt under
 /opt/install-test
 and leaves dirs of sub-datasets in there. (That is problem from before
 20141010)
 It is suggested it is illumos problem and to cry there for leaving /opt in
 piece... Let's cry.



 ___
 openindiana-discuss mailing list
 openindiana-discuss@openindiana.org
 http://openindiana.org/mailman/listinfo/openindiana-discuss

___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss


Re: [OpenIndiana-discuss] Hipster upgrade 18-feb-2015

2015-03-26 Thread ken mays via openindiana-discuss
i915resolution is a nice utility.
I had built out the Gallium drivers awhile ago so I think we could do a lot of 
supportfrom the 'userland' perspective. Just need to get the kernel DRM pieces 
updated for some efforts.
~K 


 On Thursday, March 26, 2015 5:40 AM, Jonathan Adams 
t12nsloo...@gmail.com wrote:
   

 I got fed up one evening, and then I remembered working with Intel on my
old Solaris 10 desktop, and needing to set the resolution with
915resolution ...

I went back to my old files and found it, downloaded the newer version
from 2007, made some changes to it and got a working 915resolution for
Illumos that allows the modes to be set.

I then de-installed the intel driver from hipster, set the resolution and
have vesa working at 1280x800.

obviously not the perfect answer, but since the hardware acceleration on
the old intel driver (on the old intel card) wasn't the best, I have a
working, updated version of OI on my laptop, with no sticky's.

Jon

On 20 February 2015 at 09:14, Nikola M minik...@gmail.com wrote:


 On 02/18/15 05:03 PM, Jonathan Adams wrote:

 root@jadlaptop:~# pkg list -af | grep xorg-video-intel
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.99.917-2015.0.0.0        ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.18.0-2015.0.0.0          ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.18.0-2015.0.0.0          ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.9.1-2015.0.0.0          ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.6.3-2015.0.0.1          ---
 x11/server/xorg/driver/xorg-video-intel (openindiana.org)
 2.6.3-2014.1.3.0          i--

 hmm ... I do appear to be stuck at an old version of the intel driver ...

  Same (1024X768 fixed screen resolution) happens with intel 945 graphics,
 that worked OK till 2015.
 I think it also worked in 2015 before last update.
 Only, Hipster does not update 'entire' package on every update, so I can
 not update from 20141010 to 2015 prior to latest update of 2015. (destroyed
 it before another 2015 update). So 2015 is not usable currently for desktop.

 Also , managing screen brightness from GNOME applet fails in today's
 hipster-2015 , this is second blow to Hipster Laptop use since standby
 stopped working in Spring 2014 (also destroyed old Hipster BE where it last
 worked so can not go back to it to see what changed)

 When updating, even when /opt is cleaned and mounted and work like that
 right after that in 2014.1 (separate issuue on freezing display on logging
 as another user - I should report user's settings maybe),
 after update to 2015 it needs also cleaning /opt and rebooting again to
 boot right.
 Obviously update mess with installing something in /opt under
 /opt/install-test
 and leaves dirs of sub-datasets in there. (That is problem from before
 20141010)
 It is suggested it is illumos problem and to cry there for leaving /opt in
 piece... Let's cry.



 ___
 openindiana-discuss mailing list
 openindiana-discuss@openindiana.org
 http://openindiana.org/mailman/listinfo/openindiana-discuss

___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss


  
___
openindiana-discuss mailing list
openindiana-discuss@openindiana.org
http://openindiana.org/mailman/listinfo/openindiana-discuss