Re: [gentoo-user] gentoo alternatives

[Marat BN]

Chrome OS is made by Google to run specifically on the Chromebooks.  I 
don't think it is intended for general computing and there is no 
enthusiast community around it like around other distros.


The closest cousin to Gentoo would be Funtoo.  It used to be that Gentoo 
Portage could only use rsync, while Funtoo Portage could use git which 
is much faster, but since then Gentoo Portage has also gained the 
functionality to use git for this purpose.


My biggest problem with Gentoo was not so much the time needed to 
compile huge ebuilds like Firefox, Thunderbird, or Chromium, but that 
say if you neglected doing updates for a while and then decided to start 
again, then you'd have serious problems.  This is because, at least the 
way I understood it, after some time old ebuilds would get deleted from 
the Portage servers to conserve space there, but some of those now 
deleted ebuilds would still be needed as dependencies to do iterative 
updates.  The sure-way to resolve this problem would be to re-emerge the 
whole @world set, which of course would take way-longer than just 
Firefox, and might work differently because the '/etc/' configuration 
schema might have changed.


In my case I had some weird problem either emerging some ebuild or 
keeping an old version of an ebuild to keep the functionality or the 
'/etc/' schema removed in the new versions.  I just let things sit, and 
moved on to other projects.  But when later on I tried to go back to the 
original issue, I had even more trouble because now I was even further 
behind @world, and more ebuilds would not upgrade because of deleted 
dependencies.


So to sum it up, my problem with Gentoo was that you could not just do 
iterative updates after long periods of inactivity.  You pretty much had 
to emerge daily and if you had some problem then drop everything and fix 
it right away, or else you'll fall even further behind and eventually 
might have to rebuild @world.  And so because constant attention 
intervention and trial and error was required you could not just compile 
huge ebuilds overnight and go about your life during the day.


The distro I would recommend to look at now is NixOS -- it is also 
source-based, but if you have problems with one package that will not 
prevent you from keeping the rest of the system up to date.  Upstream 
changes are pulled pretty regularly.  And even though it is 
source-based, you download most packages pre-compiled.  However if you 
want to you can tweak the source and re-compile locally.  You can also 
keep multiple versions of the same package.  You also do not mess 
directly with the '/etc/' files for individual packages, instead you 
specify a global configuration "recipe" in 
'/etc/nixos/configuration.nix', which is used to generate the 
package-specific '/etc/' files.  This layer of abstraction improves 
reliability and allows easy config cloning across machines.


The down-side is that NixOS has a radically-different paradigm that 
takes a while to wrap your head around, requires learning the Nix 
Expression Language (which is radically-different too), and is not yet 
that "mature" so theoretically things can break, but I would still 
recommend it over Chrome OS.



-- Marat


On 6/7/21 1:10 AM, n952162 wrote:
I'm looking for a gentoo alternative and am surprised to see that google 
chrome os is based on gentoo.


Does anybody have any experience with this?

Do they support multi-media and basic modern desktop capabilities?  I 
see that there's some concentration on a special browser, but I'd be 
running Firefox and FVWM anyway.


Do they use /portage/ and source packages?

Do they push down every single upstream modification, like gentoo does, 
or maybe have a bit of hysteresis?


I updated on May first and built firefox 78.10.*0*.  2+ days of 
building.  I updated on June first and built firefox  78.10.*1*. and 
spent 2+ days building.  I updated today because of the same old slot 
collision problems I've run into over a year


dev-python/setuptools:0
dev-python/setuptools_scm:0
dev-python/toml:0
dev-python/certifi:0
dev-python/jinja:0
dev-python/markupsafe:0

and now, on the 7th, I'm building firefox 78.11.   I just don't have the 
time for this.  It impacts my machines too much.


Yes, I know, there are binary versions, but if I wanted to use binary, I 
wouldn't use gentoo.  And anyway, there's always rust and gcc and ...

Re: [gentoo-user] HTML5 player (YouTube) is a pain!...Alternatives?

[Marat BN]

 Original Message 
Subject: Re: [gentoo-user] HTML5 player (YouTube) is a pain!...Alternatives?
From: Alarig Le Lay 
To: gentoo-user@lists.gentoo.org
Date: 11/29/2016 01:36 AM
> On Tue Nov 29 10:19:15 2016, meino.cra...@gmx.de wrote:
>> Hi Marat,
>>
>> thanks for your help and link!
>>
>> ...if I install that addon I only get:
>> "Palyback isn't supported on this device"
>> (using Firefox 50.0)
> Mozilla began to drop flash support since firefox 50.
>
Hmmm all in the name of progress...

I just checked my own FF version and I'm still running 45.4.0.  So
perhaps the solution for having a decent YouTube experience would be to
run a < 50 version of FF with the youtube-flash-player addon in a
virtual machine or in a Docker container.


-- Marat

Re: [gentoo-user] HTML5 player (YouTube) is a pain!...Alternatives?

[Marat BN]

I've been running the youtube-flash-player Firefox addon
https://addons.mozilla.org/en-US/firefox/addon/youtube-flash-player/ for
several months now with decent results.  It just reverts YouTube back to
the old Flash video player.  It's been working well for me.  I'm not
affiliated with it and I recommend it.


-- Marat


 Original Message 
Subject: [gentoo-user] HTML5 player (YouTube) is a pain!...Alternatives?
From: meino.cra...@gmx.de
To: Gentoo 
Date: 11/28/2016 10:42 PM
> Hi,
>
> I get sick of this [CESNORED] HTML5 stuff coming from YouTube!
> Most of the videos while playayaingnaying arereare stututterutering
> liiklike helellell. 
>
> I tested a lot of HTML5 addons to fix that problem and googled
> around the world but nothing helps. The only thing I found is an
> iconic sentence, which is true in this case also:
> "You are not alone...!".
>
> Is there any way to feed the stream from youtube direktly into
> a not blown player like mpv/mplayer or such (I would prefer
> not to load a complete gui (vlc) with any click again and again.)?
>
> Thanks for any help in advance!
> Cheers
> Meino
>
> PS: For the completeness:
> I have tried to fix that problem by:
> Restarting firefox
> Disabling hardware acceleration
> Disabling webgl
> Checking, whether the YouTube-testpage shows "all ok" for HTML5 (it does)
> Checking various about:config-setting for usefulness
> De-/Installing various addons, which all promise the heaven on earth
>
>

Re: [gentoo-user] Wireless DHCP takes over resolv.conf

[Marat BN]

The solution I use when dealing with the problem of network software
overwriting '/etc/resolf.conf' is to make that file immutable with
'chattr +i'.

Not quite an answer to your question on nameserver prioritization, but
could be useful to prevent your nameservers from being changed.


-- Marat



On 07/09/2016 07:53 AM, Mick wrote:
> Hi All,
>
> I just noticed my resolv.conf is topped up with the nameservers of the 
> wireless LAN I happen to be associated at the time and my wired ethernet 
> nameserver(s) are pushed further down.  This happens despite the fact that I 
> have configured my wired ethernet to have a lesser priority than the wired 
> NIC.
>
> For example:
>
> less /etc/resolv.conf 
> # Generated by dhcpcd from wlan0.dhcp, enp11s0.dhcp
> # /etc/resolv.conf.head can replace this line
> domain openwifi
> nameserver 192.168.22.22
> nameserver 192.168.22.23
> nameserver 10.10.10.254
> # /etc/resolv.conf.tail can replace this line
>
> The first 3 non-commented entries were produced by wlan0, demoting my wired 
> ethernet nameserver.
>
> ip route show
> default via 10.10.10.254 dev enp11s0  metric 10 
> default via 10.160.95.1 dev wlan0  metric 20 
> 10.10.10.0/24 dev enp11s0  proto kernel  scope link  src 10.10.10.7  metric 
> 10 
> 10.160.95.0/29 dev wlan0  proto kernel  scope link  src 10.160.95.2  metric 
> 20 
> 127.0.0.0/8 dev lo  scope host 
>
> If I am associated, but not authenticated with the wireless LAN, any URLs I 
> try to visit will be queried with the 192.168.22.2* nameserver, before it 
> times out and 10.10.10.254 takes over.
>
> Waiting for URLs to resolve becomes tedious after a while.  Is there a way to 
> prioritise nameservers so that the NIC metric is respected, whenever the 
> resolv.conf content is updated?
>

[gentoo-user] 'emerge --update @world -pv' keeps trying to upgrade to a masked package

[Marat BN]

Hello there,

I'm having a curious problem trying to update my system.

I issue the command:

emerge --update @world -pv


Portage comes back with the following:

!!! Multiple package instances within a single package slot have been pulled
!!! into the dependency graph, resulting in a slot conflict:

net-wireless/bluez:0

   (net-wireless/bluez-4.101-r9:0/0::gentoo, installed) pulled in by
   net-wireless/bluez:0/0= required by
(media-sound/pulseaudio-5.0-r7:0/0::gentoo, installed)
   ^   
   
   

   =net-wireless/bluez-5 required by
(net-misc/networkmanager-1.0.6:0/0::gentoo, ebuild scheduled for merge)
   ^^ ^ 
   
   
   
  


But I have the following line in my '/etc/portage/package.mask':

=net-misc/networkmanager-1.0.6:0/0::gentoo


So doesn't this tell portage not to upgrade to it?   Why does it still
schedule it for merge?


Thanks a lot,
Marat

[gentoo-user] Re: 'emerge --update @world -pv' keeps trying to upgrade to a masked package

[Marat BN]

Ah, I think I figured out what was the problem.  Looked down further in
the Portage output, and saw this:

The following mask changes are necessary to proceed:
 (see "package.unmask" in the portage(5) man page for more details)
# required by gnome-extra/nm-applet-1.0.6::gentoo
# required by net-misc/networkmanager-pptp-1.0.6::gentoo[gtk]
# required by @selected
# required by @world (argument)
# /etc/portage/package.mask:
# 2015-09-19--01
=net-misc/networkmanager-1.0.6
# required by net-misc/networkmanager-pptp-1.0.6::gentoo[gtk]
# required by @selected
# required by @world (argument)
# /etc/portage/package.mask:
# 2015-09-14--01
=gnome-extra/nm-applet-1.0.6
# required by net-misc/networkmanager-1.0.6::gentoo[bluetooth]
# required by gnome-extra/nm-applet-1.0.6::gentoo
# required by net-misc/networkmanager-pptp-1.0.6::gentoo[gtk]
# required by @selected
# required by @world (argument)
# /etc/portage/package.mask:
# 2015-06-22--01
=net-wireless/bluez-5.33


It appears that 'net-misc/networkmanager-1.0.6' was being pulled in by
'gnome-extra/nm-applet-1.0.6::gentoo',
'net-misc/networkmanager-pptp-1.0.6::gentoo[gtk]', and so on, and these
were just being reported further down below.

So I added them to '/etc/portage/package,mask' as well and the problem
went away.


-- Marat

 Original Message 
Subject: 'emerge --update @world -pv' keeps trying to upgrade to a
masked package
From: Marat BN <mara...@gmail.com>
To: gentoo-user@lists.gentoo.org
Date: 10/13/2015 05:53 PM

> Hello there,
> 
> I'm having a curious problem trying to update my system.
> 
> I issue the command:
> 
> emerge --update @world -pv
> 
> 
> Portage comes back with the following:
> 
> !!! Multiple package instances within a single package slot have been pulled
> !!! into the dependency graph, resulting in a slot conflict:
> 
> net-wireless/bluez:0
> 
>(net-wireless/bluez-4.101-r9:0/0::gentoo, installed) pulled in by
>net-wireless/bluez:0/0= required by
> (media-sound/pulseaudio-5.0-r7:0/0::gentoo, installed)
>^   
>
>
> 
> (net-wireless/blueman-1.23-r2:0/0::maratbn, installed)
>^ ^ 
>
>
> 
> 
>(net-wireless/bluez-5.33:0/3::gentoo, ebuild scheduled for merge)
> pulled in by
>>=net-wireless/bluez-5 required by
> (net-misc/networkmanager-1.0.6:0/0::gentoo, ebuild scheduled for merge)
>^^ ^ 
>
>
>
>   
> 
> 
> But I have the following line in my '/etc/portage/package.mask':
> 
> =net-misc/networkmanager-1.0.6:0/0::gentoo
> 
> 
> So doesn't this tell portage not to upgrade to it?   Why does it still
> schedule it for merge?
> 
> 
> Thanks a lot,
> Marat
>