Re: [gentoo-user] Re: [~amd64] Some possibly (?) helpful hints re the big gnome-3.8 update

2013-07-28 Thread Canek Peláez Valdés
On Sat, Jul 27, 2013 at 5:56 PM, walt w41...@gmail.com wrote:
 On 07/27/2013 03:08 PM, Canek Peláez Valdés wrote:
 And actually, the long term plan is for systemd --user to basically
 replace gnome-session-manager,

 Is Lennart part of the gnome project now? ;)

 so just installing systemd is not
 going to work at all in the future, even if it *may* seems to work
 now (which I'm pretty sure it doesn't).

 Maybe I'll be able to confirm or deny tomorrow morning :)  The update
 will take all night (on my slower machine) but I fear that the build
 will quit with an error as soon as my head hits the pillow. I'll let
 you know in the morning.

 As an aside, I got this great idea for automating the update:

 Instead of removing all gnome packages before updating gnome (much too
 much boring work for lazy me) I updated the glib package first because
 All Things Gnome depend on glib.

 Then I started revdep-rebuild, which of course will automatically re-
 install every gnome package on the machine (and update to the latest
 versions of everything in the process :).

 Of course portage refused to do the glib update because of a zillion
 or more package conflicts, so I worked around that by running:

 #ebuild /usr/portage/dev-libs/glib/glib-2.36.3-r1 merge

 Let's see how many hours of pain my laziness will cost me this time :)

I finally updated; I had no problems whatsoever. I also stopped using
my overlay (and will soon discontinue it); with the introduction of
virtual/service-manager, it's not longer necessary to have a
systemd-only (without OpenRC) machine.

The only special thing I'm doing is to mask sys-apps/systemd-204,
since 205 introduced the new cgroups management code (with systemd as
the only writer of the cgroups hierarchy), and it seems to cause some
minor problems with logind. Other than that, it works withouth a
glitch: gnome-base/gnome-3.8.0, sys-apps/systemd-204, no consolekit at
all.

Then again, this is how I've been running my machines since last year.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México



Re: [gentoo-user] Re: [~amd64] Some possibly (?) helpful hints re the big gnome-3.8 update

2013-07-28 Thread Stefan G. Weichinger
Am 28.07.2013 10:04, schrieb Canek Peláez Valdés:

 The only special thing I'm doing is to mask sys-apps/systemd-204,
 since 205 introduced the new cgroups management code (with systemd as
 the only writer of the cgroups hierarchy), and it seems to cause some
 minor problems with logind. Other than that, it works withouth a
 glitch: gnome-base/gnome-3.8.0, sys-apps/systemd-204, no consolekit at
 all.

Same here, yes. I run systemd-206 but I didn't notice an problem(s) yet.
Maybe there are some and I don't get it ;-)

Stefan



[gentoo-user] gentoo-systemd-only deprecation

2013-07-28 Thread Canek Peláez Valdés
William Hubbs closed bug #409385[1] as fixed, introducing
virtual/service-manager and adding it to the @system set, and dropping
OpenRC from baselayout's post dependencies.

Therefore, as of today, anyone can have a Gentoo machine with only
systemd, with no OpenRC installed. Since that was the raison d'être of
the gentoo-systemd-only overlay[2], I'm deprecating it soon.

If you install dracut you will also pull sysvinit (it's needed for
killall5, IIRC), and installing baselayout (instead of
systemd-baselayout) will make orphans of some systemd configuration
files (like /etc/vconsole.conf and /etc/machine-info); but I consider
those only minor problems, and I would strongly recommend to *anyone*
using my gentoo-systemd-only overlay to drop it and use the official
mechanism in the tree to install only systemd, replacing completely
OpenRC.

Also, without OpenRC we don't have /etc/init.d/functions.sh , but you
can use the alternatives provided in my overlay or in bug #373219[3].
I'm pretty sure someone will close that bug pretty soon.

Basically, systemd is now a first class citizen in Gentoo (on par with
OpenRC), and therefore there is no need at all for using my overlay.
Thanks to all the people who helped me with pull requests and
comments; the deprecation of the overlay is great news, since now it's
officially possible in Gentoo to ditch OpenRC and switch completely to
systemd.

Regards.

[1] https://bugs.gentoo.org/show_bug.cgi?id=409385
[2] https://github.com/canek-pelaez/gentoo-systemd-only
[3] https://bugs.gentoo.org/show_bug.cgi?id=373219
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México



[gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Alain Didierjean
the above lib won't compile with a emake error. Is it me or anyone else had 
that type of problem ?



Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Alan McKinnon
On 28/07/2013 10:56, Alain Didierjean wrote:
 the above lib won't compile with a emake error. Is it me or anyone else had 
 that type of problem ?
 


It's just you.

Please post emerge output.



-- 
Alan McKinnon
alan.mckin...@gmail.com




Re: [gentoo-user] failed to setup uefi boot

2013-07-28 Thread Stroller

On 28 July 2013, at 03:06, Wankey Cheng wrote:

 Make sure that you're system is booted in EFI mode
 i didn't install a bootloader,so it can be booted in EFI mode only.

Don't some boards have a BIOS compatibility mode in the motherboard settings?

Stroller.


PS: any chance you can set your mailer to send only plain-text messages to the 
list?


Re: [gentoo-user] failed to setup uefi boot

2013-07-28 Thread Michael Hampicke
Am 28.07.2013 04:06, schrieb Wankey Cheng:
 Make sure that you're system is booted in EFI mode
 i didn't install a bootloader,so it can be booted in EFI mode only.
 

Are you sure about that? What method of installation did you choose? The
culprit when installing an EFI system is, that in order to install the
bootmgr an write to the efi nvram, the install system (live cd, usb
stick, etc.) has to be booted in efi mode. Most boards boot in bios
compatible mode by default.

 and also check that the kernel module efivars is loaded.
 do you mean EFI variable Support via sysfs?Yes,it's built-in my kernel
 

Does it work? Try efibootmgr in verbose mode.

# efibootmgr -v
BootCurrent: 
Timeout: 2 seconds
BootOrder: ,0002,0003
Boot* gentoo
HD(1,800,4,2693e7bd-d35e-4a39-9bb9-365ee533d690)File(\EFI\gentoo\grubx64.efi)
Boot0002* Hard DriveBIOS(2,0,00)P3: WDC WD10EALX-229BA1   .
Boot0003* Network Card  BIOS(6,0,00)Realtek PXE B05 D00.





signature.asc
Description: OpenPGP digital signature


Re: [gentoo-user] Re: [~amd64] Some possibly (?) helpful hints re the big gnome-3.8 update

2013-07-28 Thread Michael Hampicke
Am 28.07.2013 10:07, schrieb Stefan G. Weichinger:
 Am 28.07.2013 10:04, schrieb Canek Peláez Valdés:
 
 The only special thing I'm doing is to mask sys-apps/systemd-204,
 since 205 introduced the new cgroups management code (with systemd as
 the only writer of the cgroups hierarchy), and it seems to cause some
 minor problems with logind. Other than that, it works withouth a
 glitch: gnome-base/gnome-3.8.0, sys-apps/systemd-204, no consolekit at
 all.
 
 Same here, yes. I run systemd-206 but I didn't notice an problem(s) yet.
 Maybe there are some and I don't get it ;-)
 

I had one problem, but I am not sure, if it's related to systemd  204,
the removal of consolekit, or gnome at all.

But when logging into my gnome session, /usr/libexec/gvfsd-fuse can not
be started, because the permissions of /dev/fuse are rw-- root:root

Other distros like ubuntu have a fuse group for that, which does not
exist on gentoo. So I assume the default permissions for /dev/fuse on
gentoo machines should be rw-rw-rw- root:root?



signature.asc
Description: OpenPGP digital signature


Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Alain Didierjean


- Mail original -
 De: Alan McKinnon alan.mckin...@gmail.com
 À: gentoo-user@lists.gentoo.org
 Envoyé: Dimanche 28 Juillet 2013 10:59:37
 Objet: Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1
 
 On 28/07/2013 10:56, Alain Didierjean wrote:
  the above lib won't compile with a emake error. Is it me or
  anyone else had that type of problem ?
  
 
 
 It's just you.
 
 Please post emerge output.
 

Thanks. Emerge output: emake error.
I could'nt find an emake executable. Which package contains such a beast ?



Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Neil Bothwick
On Sun, 28 Jul 2013 17:59:49 +0200 (CEST), Alain Didierjean wrote:

   the above lib won't compile with a emake error. Is it me or
   anyone else had that type of problem ?
 
  
  
  It's just you.
  
  Please post emerge output.

 
 Thanks. Emerge output: emake error.

Your portage installation is seriously screwed if that is all the output
you get.

Don't try to guess which parts of the output are relevant, if you could do
that you could probably work out the cause of the problem. Post the full
output plus any other information that output tells you to (and it
does).

Remember, you are the one sat in front of your terminal, the rest of us
only know what you told us, and so far your two mails have said.

It doesn't work.
It doesn't work because of an error.

There are some clever people on this list but, AFAIK, no magicians or
mind readers.

 I could'nt find an emake executable. Which package contains such a
 beast ?

emake is a portage function, not a standalone executable.


-- 
Neil Bothwick

QOTD:
The only easy way to tell a hamster from a gerbil is that the
gerbil has more dark meat.


signature.asc
Description: PGP signature


Re: [gentoo-user] Re: bash-completion change?

2013-07-28 Thread Neil Bothwick
On Sat, 27 Jul 2013 10:49:46 -0400, gottl...@nyu.edu wrote:

  emerge -av1 \$(qfile -q -S -C /usr/share/bash-completion)
  
  little syntax help:
  
  emerge -av1 $(qfile -q -S -C /usr/share/bash-completion)  
 
  emerge -1a /usr/share/bash-completion  
 
 Yes we learned this trick a month or two ago.

Apparently, not all of us.


-- 
Neil Bothwick

Remember that the Titanic was built by experts, and the Ark by a newbie


signature.asc
Description: PGP signature


Re: [gentoo-user] Re: [~amd64] Some possibly (?) helpful hints re the big gnome-3.8 update

2013-07-28 Thread Canek Peláez Valdés
On Sun, Jul 28, 2013 at 8:23 AM, Michael Hampicke m...@hadt.biz wrote:
 Am 28.07.2013 10:07, schrieb Stefan G. Weichinger:
 Am 28.07.2013 10:04, schrieb Canek Peláez Valdés:

 The only special thing I'm doing is to mask sys-apps/systemd-204,
 since 205 introduced the new cgroups management code (with systemd as
 the only writer of the cgroups hierarchy), and it seems to cause some
 minor problems with logind. Other than that, it works withouth a
 glitch: gnome-base/gnome-3.8.0, sys-apps/systemd-204, no consolekit at
 all.

 Same here, yes. I run systemd-206 but I didn't notice an problem(s) yet.
 Maybe there are some and I don't get it ;-)


 I had one problem, but I am not sure, if it's related to systemd  204,
 the removal of consolekit, or gnome at all.

 But when logging into my gnome session, /usr/libexec/gvfsd-fuse can not
 be started, because the permissions of /dev/fuse are rw-- root:root

 Other distros like ubuntu have a fuse group for that, which does not
 exist on gentoo. So I assume the default permissions for /dev/fuse on
 gentoo machines should be rw-rw-rw- root:root?

My problem was that *sometimes* (not always) I was unable to unlock my
session after suspending my laptop or desktop. Reverting back to
systemd-204 solved it, so I'm assuming that's the problem, although I
didn't really investigated the issue.

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México



Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Alan McKinnon
On 28/07/2013 17:59, Alain Didierjean wrote:
 
 
 - Mail original -
 De: Alan McKinnon alan.mckin...@gmail.com
 À: gentoo-user@lists.gentoo.org
 Envoyé: Dimanche 28 Juillet 2013 10:59:37
 Objet: Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

 On 28/07/2013 10:56, Alain Didierjean wrote:
 the above lib won't compile with a emake error. Is it me or
 anyone else had that type of problem ?



 It's just you.

 Please post emerge output.

 
 Thanks. Emerge output: emake error.
 I could'nt find an emake executable. Which package contains such a beast ?
 


No package explicity claims it:

$ ls -al /usr/lib64/portage/bin/ebuild-helpers/emake
-rwxr-xr-x 1 root root 922 Jul 26 16:16
/usr/lib64/portage/bin/ebuild-helpers/emake

$ equery belongs /usr/lib64/portage/bin/ebuild-helpers/emake
 * Searching for /usr/lib64/portage/bin/ebuild-helpers/emake ...
alanm@khamul ~


genlop -l shows the package emerged at that time is portage. I don't
quite figure yet how the file got created and not recorded as belonging
to portage.

Maybe your's has the wrong permissions or ownership?


Here's the content of the file just in case:

#!/bin/bash
# Copyright 1999-2010 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
#
# emake: Supplies some default parameters to GNU make. At the moment the
#only parameter supplied is -jN, where N is a number of
#parallel processes that should be ideal for the running host
#(e.g. on a single-CPU machine, N=2).  The MAKEOPTS variable
#is set in make.globals. We don't source make.globals
#here because emake is only called from an ebuild.

source ${PORTAGE_BIN_PATH:-/usr/lib/portage/bin}/isolated-functions.sh

if [[ $PORTAGE_QUIET != 1 ]] ; then
(
for arg in ${MAKE:-make} $MAKEOPTS $EXTRA_EMAKE $@ ; do
[[ ${arg} == * * ]] \
 printf '%s'  ${arg} \
|| printf %s  ${arg}
done
printf \n
) 2
fi

${MAKE:-make} ${MAKEOPTS} ${EXTRA_EMAKE} $@
ret=$?
[[ $ret -ne 0 ]]  __helpers_die ${0##*/} failed
exit $ret




-- 
Alan McKinnon
alan.mckin...@gmail.com




[gentoo-user] Re: bash-completion change?

2013-07-28 Thread Nikos Chantziaras

On 28/07/13 19:08, Neil Bothwick wrote:

On Sat, 27 Jul 2013 10:49:46 -0400, gottl...@nyu.edu wrote:


emerge -av1 \$(qfile -q -S -C /usr/share/bash-completion)


little syntax help:

emerge -av1 $(qfile -q -S -C /usr/share/bash-completion)


emerge -1a /usr/share/bash-completion


Yes we learned this trick a month or two ago.


Apparently, not all of us.


And apparently not even the Gentoo dev who put that message in the 
ebuild :-)





Re: [gentoo-user] Re: bash-completion change?

2013-07-28 Thread Alan McKinnon
On 28/07/2013 18:47, Nikos Chantziaras wrote:
 On 28/07/13 19:08, Neil Bothwick wrote:
 On Sat, 27 Jul 2013 10:49:46 -0400, gottl...@nyu.edu wrote:

 emerge -av1 \$(qfile -q -S -C /usr/share/bash-completion)

 little syntax help:

 emerge -av1 $(qfile -q -S -C /usr/share/bash-completion)

 emerge -1a /usr/share/bash-completion

 Yes we learned this trick a month or two ago.

 Apparently, not all of us.
 
 And apparently not even the Gentoo dev who put that message in the
 ebuild :-)
 
 


I'm still stumped about what this change is for or what it does.
And I've read all the bugs, all the posts here - still stumped.

I get a sense that it somehow has something to do with

make the ebuild $MORE_CLOSELY resemble the $BEHAVIOUR of $UPSTREAM_DEFAULT

Note the variables :-)



-- 
Alan McKinnon
alan.mckin...@gmail.com




Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Neil Bothwick
On Sun, 28 Jul 2013 18:35:07 +0200, Alan McKinnon wrote:

  I could'nt find an emake executable. Which package contains such a
  beast ? 
 
 
 No package explicity claims it:
 
 $ ls -al /usr/lib64/portage/bin/ebuild-helpers/emake
 -rwxr-xr-x 1 root root 922 Jul 26 16:16
 /usr/lib64/portage/bin/ebuild-helpers/emake
 
 $ equery belongs /usr/lib64/portage/bin/ebuild-helpers/emake
  * Searching for /usr/lib64/portage/bin/ebuild-helpers/emake ...
 alanm@khamul ~
 
 
 genlop -l shows the package emerged at that time is portage. I don't
 quite figure yet how the file got created and not recorded as belonging
 to portage.

That's the old lib vs lib64 thing

]% qfile /usr/lib64/portage/bin/ebuild-helpers/emake
returns nothing except an RC of 1

% qfile /usr/lib/portage/bin/ebuild-helpers/emake 
sys-apps/portage (/usr/lib/portage/bin/ebuild-helpers/emake)


-- 
Neil Bothwick

Top Oxymorons Number 18: Taped live


signature.asc
Description: PGP signature


Re: [gentoo-user] Re: bash-completion change?

2013-07-28 Thread Neil Bothwick
On Sun, 28 Jul 2013 19:47:38 +0300, Nikos Chantziaras wrote:

  Yes we learned this trick a month or two ago.  
 
  Apparently, not all of us.  
 
 And apparently not even the Gentoo dev who put that message in the 
 ebuild :-)

Is is portage 2.2 only? I don't have anything with portage 2.1 installed
to check.


-- 
Neil Bothwick

Top Oxymorons Number 38: Government organization


signature.asc
Description: PGP signature


[gentoo-user] Re: [~amd64] Some possibly (?) helpful hints re the big gnome-3.8 update

2013-07-28 Thread walt
On 07/28/2013 01:04 AM, Canek Peláez Valdés wrote:
 I finally updated; I had no problems whatsoever.

The only way I could solve all of the conflicts on my openrc machine
was to enable the systemd useflag globally (to prevent consolekit and
udev being pulled in).  It's updating to gnome-3.8 now and I'm sure
it will take all night.

With systemd being used by several critical gnome packages I'm thinking
that systemd will indeed need to be running for all those packages to
work correctly.  But I'll find out tomorrow :)



Re: [gentoo-user] fail: kde-base/kdelibs-4.10.5-r1

2013-07-28 Thread Alan McKinnon
On 28/07/2013 20:23, Neil Bothwick wrote:
 On Sun, 28 Jul 2013 18:35:07 +0200, Alan McKinnon wrote:
 
 I could'nt find an emake executable. Which package contains such a
 beast ? 


 No package explicity claims it:

 $ ls -al /usr/lib64/portage/bin/ebuild-helpers/emake
 -rwxr-xr-x 1 root root 922 Jul 26 16:16
 /usr/lib64/portage/bin/ebuild-helpers/emake

 $ equery belongs /usr/lib64/portage/bin/ebuild-helpers/emake
  * Searching for /usr/lib64/portage/bin/ebuild-helpers/emake ...
 alanm@khamul ~


 genlop -l shows the package emerged at that time is portage. I don't
 quite figure yet how the file got created and not recorded as belonging
 to portage.
 
 That's the old lib vs lib64 thing
 
 ]% qfile /usr/lib64/portage/bin/ebuild-helpers/emake
 returns nothing except an RC of 1
 
 % qfile /usr/lib/portage/bin/ebuild-helpers/emake 
 sys-apps/portage (/usr/lib/portage/bin/ebuild-helpers/emake)
 
 


Good catch, I missed that

$ ls -ald /usr/lib*
lrwxrwxrwx   1 root root  5 Jan  5  2012 /usr/lib - lib64
drwxr-xr-x  50 root root  53248 Jul 23 10:00 /usr/lib32
drwxr-xr-x 163 root root 163840 Jul 27 01:50 /usr/lib64
drwxr-xr-x  13 root root   4096 Jul 27 01:50 /usr/libexec


-- 
Alan McKinnon
alan.mckin...@gmail.com




[gentoo-user] Re: bash-completion change?

2013-07-28 Thread Nikos Chantziaras

On 28/07/13 21:25, Neil Bothwick wrote:

On Sun, 28 Jul 2013 19:47:38 +0300, Nikos Chantziaras wrote:


Yes we learned this trick a month or two ago.


Apparently, not all of us.


And apparently not even the Gentoo dev who put that message in the
ebuild :-)


Is is portage 2.2 only? I don't have anything with portage 2.1 installed
to check.


Works with 2.1 from ~arch too. And it has for a while now; it's just 
something I used only once or twice and tend to forget about it.





Re: [gentoo-user] failed to setup uefi boot

2013-07-28 Thread Wankey Cheng
I following the step from handbook and
http://wiki.gentoo.org/wiki/UEFI_Gentoo_Quick_Install_Guide.

and sometime after i boot from usb(efi mode),i will see the message this is
none ...when i login with root ,i found that (none)#...what's wrong with it
?


The last,I have give up boot from uefi now ,maybe i should use grub for
easy,i hope i can success after i learn more.


2013/7/28 Michael Hampicke m...@hadt.biz

 Am 28.07.2013 04:06, schrieb Wankey Cheng:
  Make sure that you're system is booted in EFI mode
  i didn't install a bootloader,so it can be booted in EFI mode only.
 

 Are you sure about that? What method of installation did you choose? The
 culprit when installing an EFI system is, that in order to install the
 bootmgr an write to the efi nvram, the install system (live cd, usb
 stick, etc.) has to be booted in efi mode. Most boards boot in bios
 compatible mode by default.

  and also check that the kernel module efivars is loaded.
  do you mean EFI variable Support via sysfs?Yes,it's built-in my kernel
 

 Does it work? Try efibootmgr in verbose mode.

 # efibootmgr -v
 BootCurrent: 
 Timeout: 2 seconds
 BootOrder: ,0002,0003
 Boot* gentoo

 HD(1,800,4,2693e7bd-d35e-4a39-9bb9-365ee533d690)File(\EFI\gentoo\grubx64.efi)
 Boot0002* Hard DriveBIOS(2,0,00)P3: WDC WD10EALX-229BA1   .
 Boot0003* Network Card  BIOS(6,0,00)Realtek PXE B05 D00.






[gentoo-user] Re: gentoo-systemd-only deprecation

2013-07-28 Thread »Q«
On Sun, 28 Jul 2013 03:22:02 -0500
Canek Peláez Valdés can...@gmail.com wrote:

 Basically, systemd is now a first class citizen in Gentoo (on par with
 OpenRC), and therefore there is no need at all for using my overlay.
 Thanks to all the people who helped me with pull requests and
 comments; the deprecation of the overlay is great news, since now it's
 officially possible in Gentoo to ditch OpenRC and switch completely to
 systemd.

Thanks for all the work everybody's done on this, and thanks in advance
for all the work to come.  The emphasis on choice is one of the main
reasons I'm a Gentoo user, and it's good to see you care about
providing choice for such an important thing as system manager.  

And thanks for ensuring that no disruption was caused to those of us
who continue to choose OpenRC.  (Or IMO trivial disruption at worst,
like the little bits of cruft that can be install-masked.)

Another of the main reasons I use Gentoo is (well, was) the emphasis on
proving thorough, clear information.  I don't know enough about init
systems to really judge, but https://wiki.gentoo.org/wiki/Systemd
looks very thorough, and it's definitely clear, so thanks for that too.