> --
> Fabian Groffen
> Gentoo on a different level
I filed a bug report on gentoo bugzilla Bug 687286 and posted the problem here:
https://forums.gentoo.org/viewtopic-t-1097582.html
Comments appreciated
Regards
Michael Fothergill
I have a separate gentoo installation on this machine.
Could I compile m4 on it and port it to the debian install and run it
there in some way?
Regards
MF
en fixed in Gentoo with this patch
> >
> > [1]https://gitweb.gentoo.org/repo/gentoo.git/plain/sys-devel/m4/files/m4-1.4.18-glibc228.patch
> >
> > Best,
> >
> > --
> >
> > Olivier Huber
> >
> > On Tue, Jun 4, 2019 at 1:35 PM Michael Fother
ches provided by Debian and/or Ubuntu):
apt-get source m4
cd m4-*
Would a local compilation of m4 this fashion on my debian install help here?
Regards
NF
then compile it as usual
On Tue, 4 Jun 2019 at 12:14, Fabian Groffen wrote:
>
> On 04-06-2019 12:01:42 +0100, Michael Fothergill wrote
why the newer versions
> didn't work.
>
> Fabian
>
>
> On 04-06-2019 11:10:47 +0100, Michael Fothergill wrote:
> > On Tue, 4 Jun 2019 at 11:01, Fabian Groffen wrote:
> > >
> > > Hi,
> > >
> > > Can you confirm you don't have xz tool availab
r versions
> didn't work.
>
> Fabian
>
>
> On 04-06-2019 11:10:47 +0100, Michael Fothergill wrote:
> > On Tue, 4 Jun 2019 at 11:01, Fabian Groffen wrote:
> > >
> > > Hi,
> > >
> > > Can you confirm you don't have xz tool available on your syst
On Tue, 4 Jun 2019 at 11:01, Fabian Groffen wrote:
>
> Hi,
>
> Can you confirm you don't have xz tool available on your system?
Many thanks for the response. I am not sure if xz utils is installed
by default in debian:
https://packages.debian.org/sid/xz-utils
I will log in to the debian
It seems the bug has been reported before:
https://bugs.archlinux.org/task/59562
Oh well.
Regards
MF
On Tue, 4 Jun 2019 at 12:14, Fabian Groffen wrote:
>
> On 04-06-2019 12:01:42 +0100, Michael Fothergill wrote:
> > Many thanks again.
> >
> > I have po
t; bz2.
>
> We'd need a larger part of the stage1 log to see why the newer versions
> didn't work.
>
> Fabian
>
>
> On 04-06-2019 11:10:47 +0100, Michael Fothergill wrote:
> > On Tue, 4 Jun 2019 at 11:01, Fabian Groffen wrote:
> > >
> > > H
ions due to lack of
> xz. Odd though, as the latest m4 (1.4.18) is also available as .gz and
> bz2.
>
> We'd need a larger part of the stage1 log to see why the newer versions
> didn't work.
>
> Fabian
>
>
> On 04-06-2019 11:10:47 +0100, Michael Fothergill wrote:
The installation seems to be running normally.
A large number of packages have been installed including portage which
seems to be compiling gcc 9..
Regards
MF
Dear All,
The installation has succeeded.
I did it in an attempt to try to solve the problems I was having here:
https://forums.gentoo.org/viewtopic-t-1096872.html
Is there some documentation I can read on gentoo prefix?
Regards
MF
Many thanks for your reply.
The reason for compiling a kernel in gentoo tailored for use in debian
is as follows.
A while ago when the first set of kernels were made with the patches
for the meltdown and spectre vulnerabilities, I
installed the latest kernels with these fixes as soon as they
commands if they don't work in gentoo prefix?
Cheers
MF
On Tue, 4 Jun 2019 at 18:24, Fabian Groffen wrote:
>
> On 04-06-2019 17:19:21 +0100, Michael Fothergill wrote:
> > Dear All,
> >
> > The installation is getting near completion. Unfortunately portage
> > seem
I took edited the ebuild file for the nghttp2 package as follows:
multilib_src_configure() {
local myeconfargs=(
--disable-examples
--disable-failmalloc
--disable-python-bindings
--disable-werror
Many thanks for the response.
I tried it and it worked.
nghttp2 has compiled.
I am going to try to emerge chromium now.
Regards
MF
On Tue, 25 Jun 2019 at 14:18, Fabian Groffen wrote:
> Hi Mike,
>
> Run:
> ebuild nghttp2-1.39.1.ebuild digest
>
> that would fix it
>
> Fabian
>
>
>
>
MAny thanks again.
node js 999 has compiled and now chromium is slowly compiling.
I should have turned on the jumbo-build flag..
I have a conventional gentoo install on this machine alongside
debian/gentoo prefix.
If I would take e.g. the video card settings etc from the make.conf
files in an out of it from the debian
OS environment?
Regards
Michael Fothergill
rly finds Boost::ASIO, as it can set BOOST_ROOT and
BOOST_LDFLAGS." amongst other things.
What does this refer to and if it would help where does this script live?
Can I edit e.g. with emacs and add this extra line somewhere and it
would then fix the problem?
Comments appreciated.
Regards
MF
PS I forgot to mention that I am running gentoo prefix inside a debian
buster install I have on an amd64 kaveri box.
Regards
MF
On Sun, 23 Jun 2019 at 18:04, Michael Fothergill
wrote:
>
> Dear All,
>
> If I install package nghttp2 v 1.0.36.0-2 in debian buster can gentoo
Dear All,
If I install package nghttp2 v 1.0.36.0-2 in debian buster can gentoo
prefix use it to help emerge chromium to get around the nghttp2 bug in
gentoo prefix?
Comments appreciated
Regards
Michael Fothergill
The bug I am asking about is here:
https://bugs.gentoo.org/647740
Regards
MF
On Sun, 23 Jun 2019 at 18:06, Michael Fothergill
wrote:
>
> PS I forgot to mention that I am running gentoo prefix inside a debian
> buster install I have on an amd64 kaveri box.
>
> Regards
>
>
I have installed nghttp2 on debian.
See here
Get:1 http://ukdebian.mirror.anlx.net/debian buster/main amd64
libjemalloc2 amd64 5.1.0-3 [225 kB]
Get:2 http://ukdebian.mirror.anlx.net/debian buster/main amd64
nghttp2-proxy amd64 1.36.0-2 [389 kB]
Get:3 http://ukdebian.mirror.anlx.net/debian
Many thanks for your advice here.
As it turns out I also have a separate gentoo conventional
installation on the same machine.
If I login to the ordinary gentoo install and then run emerge nghttp2
on it and it works (since I am not running gentoo prefix and its
specific bug problem)
could I then
ib/repos/gentoo) in
> net-libs/nghttp2/nghttp2-1.39.1.ebuild.
>
> Thanks,
> Fabian
>
> On 23-06-2019 11:03:44 +0100, Michael Fothergill wrote:
> > PS
> >
> I ran the installation again and it crashed immediately as follows:
> portage: 'portage' user or group missing.
>
For the defaults, line 1 goes into passwd, and 2 into group.
portage:x:250:250:portage:/var/tmp/portage:/bin/false
portage::250:portage
*** WARNING ***
Many thanks for the advice and help again.
The installation crashed again due to a file ownership issue.
I will start it again and see if this cures the problem.
The relevant output is here:
* Detected file collision(s):
*
*
The install has crashed again in the same way:
>> /home/mikef/gentoo/usr/lib64/libgettextsrc.so -> libgettextsrc-0.20.1.so
!!! Failed to move
/home/mikef/gentoo/var/db/pkg/sys-devel/-MERGING-gettext-0.20.1 to
/home/mikef/gentoo/var/db/pkg/sys-devel/gettext-0.20.1
!!! [Errno 13] Permission denied:
prefix.
>
> On Wed, Jun 26, 2019, 03:53 Fabian Groffen
>> On 25-06-2019 17:37:16 +0100, Michael Fothergill wrote:
>> > MAny thanks again.
>> >
>> > node js 999 has compiled and now chromium is slowly compiling.
>> >
>> > I should have tur
I ran the emerge command again adding the exclude option for
dev-libs/libinput:
mikef@fart ~/gentoo/etc/portage $ emerge --autounmask-write
--exclude=dev-libs/libinput x11-base/xorg-server
Calculating dependencies... done!
!!! All ebuilds that could satisfy ">=dev-libs/libinput-1.5.0:0=" have
PSmikef@fart ~/gentoo/etc/portage $ startx
/usr/lib/xorg/Xorg.wrap: Only console users are allowed to run the X server
xinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: server error
Couldn't get a file descriptor referring to the console
mikef@fart
Dear All,
I put the line
x11-base/xorg-server -suid
into the package.use file.
I did the source update
I then ran emerge xorg-server again.
It ran OK until it decided it needed to emerge sys-fs/udev-init-scripts
as a result of the -suid change:
>>> Emerging (4 of 4)
I also ran this command:
mikef@fart:~/gentoo$ startx -- vt1
X.Org X Server 1.20.5
X Protocol Version 11, Revision 0
Build Operating System: Linux 4.19.0-5-amd64 x86_64 Gentoo
Current Operating System: Linux fart 4.19.0-5-amd64 #1 SMP Debian 4.19.37-3
(2019-05-15) x86_64
Kernel command line:
Dear All,
If there is no way to fix the input device compilation problem, the next
step
is to try bedrock linux, a meta distribution specifically designed to allow
more than
one linux distro to be combined and run together in a set of layers.
Regards
MF
>
>
>
>
>
>
Many thanks for the advice again.
I ran emerge dev-libs/libinput again.
The output of
emerge --info '=dev-libs/libinput-1.13.2::gentoo'
is here:
http://dpaste.com/2P3MT0E
and
mikef@fart:~/gentoo$ emerge -pqv '=dev-libs/libinput-1.13.2::gentoo'
returns:
[ebuild N]
PS
The build log file has this piece in it just before the crash occurs:
33;01m* [0m QA Notice: the following files are outside of the prefix:
[33;01m* [0m /lib [33;01m* [0m /lib/udev [33;01m* [0m
/lib/udev/libinput-model-quirks [33;01m* [0m
/lib/udev/libinput-device-group [33;01m* [0m
I ran the command you suggested (should it be the chown command for a group
- would the chgrp command be help here?):
root@fart:/home/mikef/gentoo# chown root:
/home/mikef/gentoo/usr/lib64/chromium-browser/chrome-sandbox
root@fart:/home/mikef/gentoo#
It still gives the same error:
mikef@fart:~$
Also, when I was doing the non root xorg stuff I ran the following four
commands:
chown -v :input /usr/bin/Xorg
chmod -v g+s /usr/bin/Xorg
gpasswd -d user input
usermod -a -G video user
I got them from the gentoo non root xorg web page.
I hope they have not contributed to the problem.
Many thanks again for your advice.
I ran the two commands you recommended:
root@fart:/home/mikef/gentoo# chmod 4755
/home/mikef/gentoo/usr/lib64/chromium-browser/chrome-sandbox
root@fart:/home/mikef/gentoo# chown root
/home/mikef/gentoo/usr/lib64/chromium-browser/chrome-sandbox
Many thanks again.
I ran the command to list the chromium sand box file:
mikef@fart:~/gentoo/etc/portage$ ls -la
/home/mikef/gentoo/usr/lib64/chromium-browser/chrome-sandbox
-rwxr-xr-x 1 root root 22384 Jun 26 01:52
/home/mikef/gentoo/usr/lib64/chromium-browser/chrome-sandbox
Many thanks for your response.
I tried running chromium to see if it would work.
I did as user and then as root.
mikef@fart:~$ /home/mikef/gentoo/usr/bin/chromium
[1459:1459:0628/081848.057654:FATAL:setuid_sandbox_host.cc(157)] The SUID
sandbox helper binary was found, but is not configured
egards
MF
On Fri, 28 Jun 2019 at 11:19, Fabian Groffen wrote:
> chmod 4755 the file again, and check ls -la
>
> On 28-06-2019 10:11:06 +0100, Michael Fothergill wrote:
> > Many thanks again.
> >
> > I ran the command to list the chromium sand box file:
> >
Chromium seems to be running normally.
Google searches work OK on it.
The output to the terminal before it runs is as follows:
mikef@fart:~/gentoo/etc/portage$ /home/mikef/gentoo/usr/bin/chromium
/etc/chromium/master_preferences: line 2: distribution:: command not found
Chromium is running!
A great many thanks.
Regards
MF
On Fri, 28 Jun 2019 at 11:22, Michael Fothergill <
michael.fotherg...@gmail.com> wrote:
> I ran the chmod 4755 command again and now there is an "s" present..
>
> mikef@fart:~/gentoo/etc/portage$ !472
> l
I have started the installation again.
This time it seems to have started on the same package where the crash
occurred.
It doesn't seem to be compiling gcc 9 and binutils all over again for hours
until it gets past the package where it died over night
and goes on and dies on a later package in
I thought to post a few links to people having problems with EPERM in WSL 2.
Maybe they could be helpful at little bit:
https://github.com/microsoft/WSL/issues/4255
https://github.com/Microsoft/WSL/issues
d
>
> I remember this problem from Ubuntu bootstraps, you need to apt-get g++.
>
> Fabian
>
>
> On 02-07-2019 12:45:11 +0100, Michael Fothergill wrote:
> > This early piece of output does not look encouraging to me:
> >
> > gcc version 6.3.0 20170516 (Debian 6.3.0-1
3-07-2019 19:17:20 +0100, Michael Fothergill wrote:
> > Yes.
> >
> > The parent debian OS lives on the ntfs partition alongside Windows 10.
> >
> > WSL 2 is Microsoft''s latest foray into Linux.
> >
> > I will see what I can find out about this EPERM thin
Yes.
The parent debian OS lives on the ntfs partition alongside Windows 10.
WSL 2 is Microsoft''s latest foray into Linux.
I will see what I can find out about this EPERM thing.
Regards
MF
The founder of Bedrock linux says that running gentoo and debian together
in it allows the dpkg-buildpackage of debian to be used by gentoo...
This was the main problem I had with the out of tree kernels I tried making
in gentoo for the debian install.
See here:
Dear All,
The installation crashed in stage 3.
It seems to be a file permission problem.
The stage3 log file is too large for pastebinit.
I have posted the last 200 odd lines here:
https://paste.debian.net/1090102/
This looks to be the problem:
>>>
Dear All,
I have posted the build log here:
http://paste.debian.net/
Regards
MF
>
>
>
>
>
>
>
>
>>>
It's OK.
I was sailing a bit too close to the wind there I think.
I have recently installed Clear Linux on the Fx 8350 box I have.
For a bit of fun I tried installing Gentoo Prefix on it.
The bootstrap script crashed at the outset - it is unhappy with the
environment variables for CCFLAGS and
Many thanks again.
OK I went into the ebuild directory for libinput.
I then edited the ebuild file and added the code you gave at the end of the
file:
src_install() {
meson_src_install
if use doc ; then
docinto html
dodoc -r
I did it and I ran emerge libinput and it has compiled and installed.
What do I run the diff command on?
I don't even know if I have diff installed.
Regards
MF
Many thanks again.
I have now compiled and installed xorg server and all the video libraries
and driver packages.
I will see what happens if I try running startx.
Regards
MF
I ran startx.
I am in /home/mikef/gentoo - that is the $PREFIX directory.
The error I get is
Only console users are allowed to run the X server
See here:
mikef@fart:~/gentoo$ startx
/usr/lib/xorg/Xorg.wrap: Only console users are allowed to run the X server
xinit: giving up
xinit: unable to
Looking at your patch code based on what you have just explained,
I could go back and re-edit the build file and remove the extra code I
added at the end restoring the file to its original state.
Then I could move the cursor up to the following section of the build file:
src_configure() {
A great many thanks for the help here.
There are quite a few different write ups on making patches in gentoo.
I am using this one:
https://wiki.gentoo.org/wiki//etc/portage/patches
I am going to do the following things:
1. I am going to create a directory to put the patch in:
2. mkdir -p
PS I forgot to add in step 3 that I will create a file to put the helpful
patch code in using emacs.
I will also replace the $EPREFIX symbol with /home/mikef/gentoo in each
instance in the patch code.
Regards
MF
On Thu, 27 Jun 2019 at 12:03, Michael Fothergill <
michael.foth
There is something here:
https://wiki.gentoo.org/wiki/Non_root_Xorg
that might be helpful. I will read through it and try it out.
Regards
MF
>
>
>
>
>
>>
Many thanks for your help again.
I did the export command as you suggested and then ran the bootstrap script
again.
It started at stage 2 and ran through and crashed again exactly as before.
No cure yet.
I can't emerge wgetpaste yet.
So I have posted as much of the build log as the terminal
$? = 1
configure:4295: gcc -qversion >&5
gcc: error: unrecognized command line option '-qversion'; did you mean
'--version'?
gcc: fatal error: no input files
compilation terminated.
configure:4306: $? = 1
Comments appreciated.
Regards
MF
On Tue, 2 Jul 2019 at 12:40, Michael Fothergill <
I tried installing emacs but it is awkward to use in the windows terminal.
Regards
MF
>
>
>
>>
>>
>>
Many thanks again.
I have posted as much of the config.log file as the terminal buffer retains
here:
https://paste.debian.net/1089991/
Comments appreciated.
Regards
Michael
>
>
>
bound of lines around that are the interesting ones.
>
> On 02-07-2019 10:56:32 +0100, Michael Fothergill wrote:
> > Many thanks again.
> >
> > I have posted as much of the config.log file as the terminal buffer
> retains
> > here:
> >
> > [1]https://pas
.
Regards
Michael Fothergill
It has Ubuntu, ArchLinux and so on but no Gentoo.
>
> On Sat, Jul 20, 2019, 10:06 AM Michael Fothergill <
> michael.fotherg...@gmail.com> wrote:
>
>> Dear All,
>>
>> I notice there is a discussion about installing gentoo prefix on a mobile
>> phone:
&g
on mobile phones or is that not
quite possible yet?
Regards
Michael Fothergill
If I would install e.g. Debian on the phone using userland would I then be
able to install gentoo prefix alongside it as I have done my my amd64
kaveri box pc debian buster install?
Comments appreciated.
Regards
Michael
Climostat Ltd
I have been googling around and I think it would be better to root the
phone and install linux directly on it.
Then I can install gentoo directly...
Regards
MF
On Sun, 21 Jul 2019 at 11:09, Michael Fothergill <
michael.fotherg...@gmail.com> wrote:
> If I would install e.
Dear All,
The third ponderous compilation of gcc 9.1 is running.
Nearly there now.
Many thanks.
Regards
MF
>
>
>
Many thanks.
Gentoo prefix installed successfully on clear linux.
But something is wrong with the path to start the prefix working properly.
emerge and eselect news read commands are not found.
I changed directory to /home/mikef/gentoo/usr/bin and found emerge.
I ran
Many thanks for your email and advice.
On Thu, 18 Jul 2019 at 08:22, Fabian Groffen wrote:
> Your path is zapped by your .bashrc or something. Did the bootstrap
> script give a message/warning about this near the end of its execution?
> It should have.
>
I remember a grumble about the path
Dear All,
I have edited the .bashrc file and commented out the offending lines.
See here:
more ../.bashrc
# Use global profile when available
#if [ -f /usr/share/defaults/etc/profile ]; then
# . /usr/share/defaults/etc/profile
#fi
# allow admin overrides
if [ -f /etc/profile ]; then
.
alongside it
and then chroot into the clear linux install and fix the .bashrc file and
rescue clear linux and prefix together.
Comments appreciated.
Regards
MF
On Thu, 18 Jul 2019 at 10:05, Fabian Groffen wrote:
> On 18-07-2019 09:02:49 +0000, Michael Fothergill wrote:
> >
I have managed to hack my way back into clear linux.
It now doesn't find the profile file as expected.
bash: /etc/profile: No such file or directory
Both emerge and eselect news read run OK now.
Regards
MF
>
Many thanks again.
I will try this out.
There was a choice of window manager enviornments you could select from
when logging into clear linux.
By selecting the Gnome on Wayland option, for some reason I do not
understand, problem of my username
not being recognised was cured and I could log
I have posted a query on the clear linux forums about the CFLAGS CCXFLAGS
variable problem.
I will see what they say.
Maybe a compromise fix is possible and I will able to install gentoo prefix
on clear linux OS.
Cheers
MF
>>
>>
>>
The installation is proceeding...
The discussion on the clear linux community page here
https://community.clearlinux.org/t/installing-gentoo-prefix-in-clear-linux-system/1118/2
says not to worry too much about what I did but that I should move the
change to another location/file and
do it
Many thanks again.
The suggestion about adding unset CFLAGS CXXFLAGS to the profile file has
worked.
Gentoo prefix is being installed.
Clear Linux is installing it at a rocket like pace..
I am running thermal daemon in clear linux.
This seems to be helping avoid the cpu shutting down.
The installation of gcc 9 for the second time seems to be taking a lot
longer than the first time round.
Most distros are slower than clear linux. The Phoronix people have tested
this many times including on AMD machines not just Intel ones.
Gentoo was not compared with clear linux so much I
Many thanks for your advice.
I will try this and see what happens.
The profile flle is here:
https://paste.debian.net/1091866/
The .bashrc file that calls it is here:
mikef@clr-0fdd1f831fb14378ac28883a158cf04d~ $ more .bashrc
# Use global profile when available
if [ -f
Sadly the cpu shut down due to overheating.
I need to get the new cooler device and install it.
The bootstrap tried detected 8 cores and ran with 5 threads or whatever it
said.
If I enter 1 instead of 5 at the start of the bootstrap run surely that
would cool things down?
Comments
Dear All,
I selected 1 parallel process in the bootstrap script when it started this
time.
The temperatures are lower now.
Clear Linux is eating gcc 9..
If I fix the cooling problem it will eat everything I throw at it for its
breakfast I think.
Regards
MF
>
>
Dear All,
What do think about the procedure recommended here:
https://sourceforge.net/p/gentooandroid/wiki/InstallingInstructions/
?
Apparently it will work on an unrooted phone.
Comments appreciated.
Regards
MF
If I would make a tutorial, would it be a discussion on the gentoo talk web
page or a youtube video or something else?
What is the most interesting thing for other users here?
Is it the compilation of chromium and the fixing of the nghttp2 problem or
is the installation of gentoo prefix on
Could I write it here?
https://wiki.gentoo.org/index.php?title=Project_Talk:Prefix/Use_cases=edit=1
Regards
MF
>>>
> I ran the installation again .
After 10 hours it crashed at a different point but with a "spawn-esque"
permission problem:
!!! Failed to move
/home/mikef/gentoo/var/db/pkg/sys-apps/-MERGING-coreutils-8.31 to
/home/mikef/gentoo/var/db/pkg/sys-apps/coreutils-8.31
!!! [Errno 13] Permission
Many thanks for all the help.
I think that WSL2 is not quite ready for gentoo prefix at present.
I tried several repeat tries with the installer.
It is stuck in a loop of permission errors.
I will try again Microsoft puts out another significant update.
The WSL team they have will be aware of
window fires up.
The history command still shows the commands I ran to install it.
Cheers
MF
On Wed, 13 May 2020 at 01:34, Benda Xu wrote:
> Hi Michael,
>
> How was this project going? Have you succeeded in compiling chromium in
> Gentoo Prefix?
>
> Benda
>
> Mi
92 matches
Mail list logo
