The officially official Devuan Forum!

You are not logged in.

#1 Re: News & Announcements » Forum search tips and tricks » 2024-01-28 19:03:47

Thanks @soren for mentioning duckduckgo. That search engine is rarely used by me, although it ranks 3rd in my firefox profile. (And basically the profile is older then my 1st devuan/jessie install ;-) Usually my order of search engines is startpage, de.wikipedia, duckduckgo, … and others. This setup works well since decades both for me and others! (Read: other peoples computer I'm maintaining.) It provides "every day search findings" in many cases. They are good almost everywhere (at least the last two decades :-)

But: what happens when looking for some devuan issue at dev1galaxy?

Today I have made some investigation to a long outstanding problem:

A) searching the dev1galaxy forum gives an URL like:

    dev1galaxy.org/search.php?action=search … earch+term

Without search restriction (this is the default!) there are many results!

TL;DR: A couple of times I have refined the search: Oh, there are still a lot of other interesting things. But I didn't find what I'm looking for.

B) using duckduckgo according to soren:

Cool. As of this writing duckduckgo provides an javascript free search interface! Append " site:dev1galaxy.org" or prepend "!dev1galaxy " to your search string (note the space before or after string).

    duckduckgo.com/html?q=search+term+site:dev1galaxy.org

or with slightly different results:

    duckduckgo.com/html?q=!dev1galaxy+search+term

C) Using startpage URL:

    www.startpage.com/do/dsearch?q=mouse+fo … galaxy.org

Again: startpage is still my favorite engine for daily usage but not for devuan. The results are very bad, only few to nothing. And they don't look better on a terminal :-(

Summary: As of this writing I'm now starting to search the dev1galaxy with duckduckgo.com/html without javascript.

-- guuml

BTW: some times I like to view some sites with a www-browser like w3m IMHO, others may prefer lynx. Searching from the commandline for "search term" may look like:

www-browser "https://search.engine/options?query=search term"

+1 point for DDG: w3m displays the results in a fancy way.

Update:

Some time ago there was a thread … ah very long time ago … I remember only vaguely … one of the first posts I have read …

Before you ask your question, …

But there are other resources available that might help you to find an answer even before you ask here …

Sometimes I have to read the information twice roll

Update 2: (last of today:)

A search script may look like

#!/bin/sh -xv
# find-dev1galaxy - search in the officially official devuan forum

SEARCH_TERM=`echo "$* site:dev1galaxy.org" | tr ' ' '+'`
exec www-browser "https://duckduckgo.com/html?q=$SEARCH_TERM"

Or if you prefer functions in bashism:

find_dev1galaxy() {
 w3m "https://duckduckgo.com/html?q=${*// /+}+site:dev1galaxy.org"
}

Ok, processing advanced search options isn't that good wink

Have a nice week --guuml

#2 Re: Installation » Updating from ASCII » 2023-07-01 23:01:04

Ok, late to the party.

Yesterday* morning I have decided: (1st) keeping devuan/jessie on dev/sda3 BUT (2nd) apt dist-upgrade from devuan/ascii via devuan/beowulf to devuan/chimaera on /dev/sda1. As of now** this works (not) very well:

  • 1. Update your system with current sources.list to the latest version. See release info for details.

  • 2. Cange your /etc/apt/sources.list to the next release: e.g. s(ascii)(beowulf)

  • 3. Do a

    apt update

    followed by an

    apt dist-upgrade

    You may consider autoremove and autoclean, too.

  • 4. Important are the Release Notes! (Hard to link to each:-( The step from  s(beowulf)(chimaera) has to consider both exim and NetworkManager.

Ok. IMHO was the wicd much better useable then network-manager since devuan/jessie! I like wicd and and I use it. But devuan/chimaera has dropped support for wicd for the right reason (according to the release notes)!

My problem: a dist-upgrade from beowulf to chimaera works well, as an established connection from wicd exists. But after an autoremove there is no wicd anymore. And  a NetworkManager was unknown to the system …

As of this writing the machine has no network at all. But transfering the right package to local file system should help. I will fix this next week.

Update (So 02 Jul 2023 01:09:45 CEST)

That minimal sources.list works for me from beowulf to chimaera works (ascii to beowulf has to be shown):
deb http://deb.devuan.org/merged/ chimaera-security non-free contrib main 
deb http://deb.devuan.org/merged/ chimaera non-free contrib main   
deb http://deb.devuan.org/merged/ chimaera-updates non-free contrib main   

-- guuml

*) Usually known as 2023-06-30
**) 01 Jul 2023 23:32:38

#3 Re: Desktop and Multimedia » [SOLVED] How to set keyboard layout in a remote X session? » 2023-02-07 23:23:09

delgado wrote:

…  And look for another remote X session solution.

Some weeks ago I have read in setxkbmap(1):

USING WITH xkbcomp
If you have an Xserver and a client shell running on different
computers and some XKB configuration files on those machines are
different, you can get problems specifying a keyboard map by model,
layout, and options names. …

For example, the command …

      setxkbmap us -print | xkbcomp - $DISPLAY

makes both steps run on the same (client) machine and loads a keyboard
map into the server.

Note the "us" difference to an example setting with de_DE.UTF-8:

$ setxkbmap -print
xkb_keymap {
	xkb_keycodes  { include "evdev+aliases(qwertz)"	};
	xkb_types     { include "complete"	};
	xkb_compat    { include "complete"	};
	xkb_symbols   { include "pc+de(nodeadkeys)+inet(evdev)"	};
	xkb_geometry  { include "pc(pc105)"	};
};

Using the command

setxkbmap de nodeadkeys -print | xkbcomp - $DISPLAY

should provide a german keyboard (with no deadkeys) to the server.

#4 Re: News & Announcements » [SOLVED] invalid: EXPKEYSIG BB23C00C61FC752C Devuan Repository » 2022-09-11 23:08:01

Sorry for being late to the party at So 11 Sep 2022 21:44:59 CEST,
date +%c, you know?

Last sunday noon, aka So 04 Sep 12:00, my usual workflow (as root) was disrupted by an EXPKEYSIG error: WTF? Network error caused by provider? No. – DNS problem? No. – What is going on? This looks like a serious problem to my local machine and has to be fixed NOW!
 
Wait, eight days later, I want to make a long story short. Since my 1st beginning with Linux kernels and GNU software on rpm based machines I know a true zen say:

Security is a matter of trust.

Once upon the time an "update" process "distrust his master voice" and therefor ignores my well choosen configuration. I've lost data and recovering from this accident takes some time. The change to debian packaging system, change of distribution, was one of the consequences. And the step from debian/jessie to devuan/jessie some years later wasn't that difficult to continue the way with GNU/Linux. Back to the failing apt update (as root;)

I'm not a C-programmer and I have only a vague understanding of so-called elliptic curves"internet security", but last week there was a urgent need for me to find some thing like a Devuan Cryptographic Key. But where? And HOW TO know that this is the right key?

Thanks to ralph.ronnquist I've found two answers. (There has always to be an alternative to init freedom;-) The 1st alternative is "allow-unauthenticated" and/or "allow-insecure-repositories". Does not sound trustworthy, really? The second way looks better to me:

  • As a "normal user" with UID>=1000 download the new key to a directory of your choice:

    wget http://deb.devuan.org/devuan/pool/main/d/devuan-keyring/devuan-keyring_2022.09.04_all.deb

    Note: Meanwhile https://www.devuan.org/os/keyring states

    apt-get install devuan-keyring

    but I tend to disagree: wget does one thing download one file, but apt installing one file may affect other packages.

  • Verify the checksum by your own, that's to say:

    sha256sum ./devuan-keyring_2022.09.04_all.deb

    has an output of

    96c4a206e8dfdc21138ec619687ef9acf36e1524dd39190c040164f37cc3468d  ./devuan-keyring_2022.09.04_all.deb

    Make sure that's ok!

  • Now inject this proved file to your system:

    # dpkg -i ./devuan-keyring_2022.09.04_all.deb

    Note: sudo want's user's password, but I'll prefer a real root shell.

  • Then update the package information:

    # apt update

Summary:

As far as I can trust myself;-) I have copied that proved deb-file to an USB stick. Using this file with dpkg -i just before any apt update works on every devuan/jessie 'til chimaera I can reach! AND it keeps my last DVD alive: Around easter 2022 I have burned that raw DVD with chimaera to check discless hardware without internet connection. Updating the keyring is just a small step just before getting another host up and running.

last but not least:

apt-cache policy devuan-keyring
devuan-keyring:
  Installiert:           2022.09.04
  Installationskandidat: 2022.09.04
  Versionstabelle:
*** 2022.09.04 500
        500 http://deb.devuan.org/merged chimaera/main amd64 Packages
        100 /var/lib/dpkg/status

apt-key list
Warning: apt-key is deprecated. Manage keyring files in trusted.gpg.d instead (see apt-key(8)).

man apt-key
apt-key(8) will last be available in Debian 11 and Ubuntu 22.04.

Thanks for all your work and a better new week.

#5 Re: Desktop and Multimedia » sometimes network is down after boot » 2022-07-01 22:20:44

The same thing happens to my new installed chimaera. Over weeks every thing works fine, but one day there was no network after login to xfce. For the 1st time this happens I just ignore it. The 2nd time I look to /var/log/syslog:

Jun 24 09:51:58 ... NetworkManager ... dhcp4 (eth0): request timed out
...
Jun 24 09:52:04 ... NetworkManager ... dhcp6 (eth0): request timed out

BTW: I liked it very much not using NetworkManager since devuan/jessie 32bit, but I understand the case against wicd:

Wicd is no longer available in Devuan 4 Chimaera having been removed
from Debian Bullseye because of it's dependency on python2 which is now
obsolete. (from chimeara release notes)

Just before jumping into some rabbit holes there may be a thread [DNG] resolv.conf to read. But for short the following command was helpful to meroot:

echo resolvconf=NO > /etc/resolvconf.conf

Hm, does not not a solution noted during the thread. Being verbose I want to cite
Alessandro Vesely

Some of that arcane is unveiled in Debian wiki:

These packages include various configuration files for other
packages (such as isc-dhcp-client). For example, resolvconf
includes a file which modifies the make_resolv_conf shell
function used by dhclient-script(8).
https://wiki.debian.org/resolv.conf#Con … resolvconf

Thanks a lot to Ale! Here is the note to openresolv:

If openresolv is installed, you can tell it to do nothing
whenever some daemon tries to modify resolv.conf, by putting
resolvconf=NO in the /etc/resolvconf.conf file. (Note: this
is not the /etc/resolv.conf file!)

This is not the end of the story: I have found a bad hack with a good side effect to thunar: initial startup is fasten by twenty times:

time thunar .

real    0m20,348s
user   0m0,005s
sys     0m0,011s

Now I'm looking for a place to discuss this.

--gü

#6 Re: Installation » [SOLVED] disabling slim » 2019-03-11 23:30:34

ralph.ronnquist wrote:

.... has almost the same effect, except that it doesn't remove slim from the "/etc/init.d/.depend.*" database, which also plays a role for sysvinit.

Thanks a lot to the other side of the world especially to ralph.ronnquist! I have NEVER EVER heard about hidden files in /etc/init.d/ up to now! But they REALLY exist in less /etc/init.d/.depend.* as I see (in devuan/ascii at least).

WTF? As far as I can remember from the solaris 2.x manual: There are no "hidden" files. Every start/stop script is placed in /etc/init.d/ and afterwards the local sysadmin has to place the appropriate symlinks to /etc/rc?.d/ according to the naming convention.

Where does this "hidden database" come from? As far as I know there was no written manual at all-

#7 Re: Installation » [SOLVED] disabling slim » 2019-03-11 21:31:06

The very traditional sysvinit was introduced some decades ago. In an ideal world It has nothing to do with systemd nor upstart! What's about man service? As of this writing I have read that man page the 1st time, and as far as I understand is saying service slim stop exactly the same as /etc/init.d/slim stop in sysvinit. But what is about the next reboot? As man init states there are several run levels in ls -1d /etc/rc*. You can easily adjust the symbolic links in /etc/rc?.d/ by adding or removing the links, but is this your question?

fabien wrote:

Hello,
....
How should I disabling slim, I don't need it.
Fabien

Are you sure you "don't need" to start the Xorg someway automagically? Did you ever tried startx from the console? What's about other "graphical login managers" like lightdm? What's about auto login as a given user?

@Fabien: As root of my own machine I can prevent starting slim with an command like

# rm /etc/rc?.d/S*slim

But there are consequences ...

#8 Re: Installation » Beowulf: Link sbin » 2019-02-20 21:57:24

Just for the records, it may be off topic: I'm watching the testing of beowulf, but in the not to far future I want to distribute a stable operating system to others which I have to maintain!  Quoting my 2nd post to dev1galaxy:

guuml.dev1 wrote:

... minimalistic devuan/jessie with xfce ... You should remember that su - is your friend. BTW: the last few years on LMDE2 betsy I was used to type sudo su. Unfortunately some time ago a guy (new to GNU/Linux but interested to learn) asked me about the differences about su and sudo and why both packages are needed. My answer was not very helpful to him because it was much too detailed ...

#1: Devuan does the right thing: During installation you have to  give two passwords for two accounts. One for root and one for yourself (UID=1000). And there are two home directories, see /etc/passwd.

#2: Since decades i'm used to type su - and the right password to gain a root shell, where I can do everything. YES, including shot me knee and break every toe afterwards. You're warned.

#3: On the other hand sudo ... asks for confirmation with my own password. Just to remember that UID=1000 is able to ruin compromise the system, too.

Back to topic: IMHO there should be a link from /sbin/ to /usr/sbin/ (just in case the merge is done:-(

BTW: Some people are mounting the /usr via network, aka remote. And then mounting /usr/local from local?

PS: Should root be able to interact with user's X-server? I've had never heard about pkexec up to now but looks like policy-kit. I'm going RTFM ...

#9 Re: Devuan Derivatives » Miyo Modern: problem with external monitor » 2019-02-07 02:44:09

In addition to the answers of MiyoLinux and ralph.ronnquist:

w3 wrote:

I have a notebook with the monitor defective, so I use an external VGA monitor.
((snipped))
Finally I created a ~ / .xinitrc file, since the default installation did not create one and put the command.
((snipped))

IMHO the X server does right automagically in most cases (and miyolinux is a well supported distribution of devuan/ascii ;-)

MiyoLinux wrote:

Dual monitors is something I've never dealt with, ...

Now the gory details: According to ralph.ronnquist I don't understand sc -c ..., too. And it is hard to believe in the  need of ~ / .xinitrc. And I've never heard about "the Pekwm window manager" up to now, I believe in slim/xfce :-) Basically slim starts the X server with /var/log/Xorg.0.log, then xfce tooks the windows management, Then you will see errors in tail -f ~/.xsession-errors. To keep a long story short:

xrandr -q -v

is your friend. On MY laptop it looks like:

xrandr program version 1.5.0 devuan/ascii as of this writing
LVDS1 connected ... 1024x768 ... screen of laptop
VGA1 connected ... 1280x1024 ... external monitor

Initially (after 1st install) both screens show the same content.

BTW: I'm using both screens as one "Arbeitsfläche" with
xrandr  --output LVDS1 --primary --auto --rotate normal --pos 0x0 \
        --output VGA1            --auto --rotate normal --left-of LVDS1

According to the initial question: w3 needs to clone "LVDS1" to "VGA1" with one single xrandr-command. May man xrandr helps?

#10 Re: Desktop and Multimedia » Launch Xorg works on laptop but not on external display » 2019-02-02 17:25:15

It's common usage on a laptop, but rarely on stationary computers, to connect an external monitor or projector to the running system. On devuan/ascii the xserver was startet by slim on "VT number 7" and less /var/log/Xorg.0.log  shows:

[...] (II) systemd-logind: logind integration requires -keeptty and -keep tty was not provided, disabling logind integration

Nice to see, but note 1: Once upon a time a collegue asks for help about his black screen. My final answer was: NEVER EVER hot plug VGA again with your next monitor! I summe this is true for DVI, but not for HDMI. Nevertheless: On my devuan/ascii-i686 i can simulate connect/disconnect trought power on/off my KVM switch.

The next step was to prevent X from starting by disabling /etc/rc?.d/S04slim. then shutdown, disconnect monitor, boot and  login on tty1 , connect monitor and startx. BTW: the logfiles goes to $HOME/.local/share/xorg/Xorg.0.log:

[...] (++) using VT number 1
[...] (II) systemd-logind: took control of session /org/freedesktop/login1/session/_31
[...] (II) xfree86: Adding drm device (/dev/dri/card0)
[...] (II) systemd-logind: got fd for /dev/dri/card0 226:0 fd 11 paused 0
...

WTF? What tooks control? And why is xfree86 mentioned? Ok, on my hardware X starts despite of this. This may be difference between acer travelmate (ascii 32bit) and thinkpadX60. But see with VGA1 as laptop and LVDS1 as external:

...
[...] (WW) intel(0): Output LVDS1: Strange aspect ratio (304/22800), consider adding a quirk
[...] (II) intel(0): resizing framebuffer to 2304x1024
[...] (II) intel(0): switch to mode 1024x768@60.0 on LVDS1 using pipe 1, position (1280, 256), rotation normal, reflection none
[...] (II) intel(0): switch to mode 1280x1024@60.0 on VGA1 using pipe 0, position (0, 0), rotation normal, reflection none
[...] (**) Option "fd" "22"

The last line "(**) from config file" is repeatet for more "fd" (file descriptors?), log ends up with

[...] (II) UnloadModule: "libinput"
[...] (II) systemd-logind: not releasing fd for 13:71, still in use
[...] (II) UnloadModule: "libinput"
[...] (II) systemd-logind: releasing fd for 13:73
...
[...] (II) UnloadModule: "libinput"
[...] (II) systemd-logind: releasing fd for 13:67
[...] (II) Server terminated successfully (0). Closing log file.

As my X starts there is a $HOME/.xsession-errors with lines:
dbus-update-activation-environment: systemd --user not found, ignoring --systemd argument

Summary: It's not nice to know, that systemd has found an backdoor entry to my current devuan_ascii i386_desktop-live. Although devuan/ascii keeps on running stable, is this problem worth a bug report?

Update: while reviewing lunario's point of view
...
[    76.733] (II) Using input driver 'wacom' for 'Serial Wacom Tablet WACf004'
[    76.734] (EE) systemd-logind: failed to take device /dev/ttyS0: No such device
[    76.734] (**) Serial Wacom Tablet WACf004: always reports core events
...

there is a bug report 232 "xserver-xorg-input-wacom: starting X session results in permanent black screen" already, dated 8 Aug 2018 16:03:01 UTC. IMHO hard to resolve.

#11 Re: Hardware & System Configuration » trouble with installing new package for epson printers » 2019-02-02 16:55:52

Sorry but there is one driver package for epson, see  epson-inkjet-printer-escpr, that can't be used:

holgerw wrote:

I can't use epson-inkjet-printer-escpr, which can be found in Devuan Repo, because this package does not contain a driver for this new printer model XP-6000 from Epson.

But there is another package epson-inkjet-printer-escpr2_1.0.30-1lsb3.2_amd64.deb from epson. That package looks like having problems on debian/jessie based distros: two examples for others:

Coming from both ubuntu 12.04 and LMDE 2 I have seen changes with the lsb (linux standard base) in devuan/jessie and ascii, but they worked for me.

Update #1: notable commands according to https://forums.linuxmint.com/viewtopic. … 4#p1395991:

# find installed packages for "Linux Standard Base"
dpkg -l | grep lsb

# find installed epson print drivers 
dpkg -l epson-inkjet-printer*

As Devuan Package Search for lsb stats, adding lsb_4.1+devuan2  with additional packages may be helpful.

apt install lsb --dry-run

Note: The Option "--dry-run" is stongly recommended! Think about it.

@holgerw: I strongly support the approach to keep working hardware running;-) Hopefully you get some glue to looking for.

Good luck and kind regards

#12 Re: Installation » Dual-boot with Windows 10 from separate drives » 2019-01-30 00:18:40

CTAKAH wrote:

last time I dealt with grub/mbr/.. was years ago and now it seems to be all different.
... (snipped)
Alexander

Welcome and thank you for the question, but as far as I know there are some points to consider:

In 2015(!) it was the last time (for me;-) to install a dualboot of a debian/jessie besides "Win XP SP4" on the same boot device. Since then some things has changed, but some basics remain the same:

1.) Microsoft likes it to be "the one and only" on your boot device: they call it "drive C:", i'll prefer "/dev/sda". And the partition the BIOS want's to boot the operating system from has to be marked as "bootable".

2) The UEFI-BIOS of your own! hardware may not cooperate: In one single case I found one specific BIOS does not want to boot from my! USB-Stick because of "checksum error". Please check this step! Note: I was not able to reproduce this specific error the 1st time and there was no 2nd chance (for me).

3) Download devuan_ascii_2.0.0_{amd64,i386}_desktop-live.iso and check the sha256sum! Then create your! boot media and proceed with step #2.

And last but not least: Backup your data to a save place you can restore from with low effort (IMHO m$-backup is not the 1st choice, consider GNU-tar for mounted ntfs partitions). And really cool is sfdisk for saving/restoring partition tables, lsblk -f for current UUIDs and fdisk -l from util-linux 2.25.2 that handles "Disklabel type(s)" dos or gpt.

Update: IMHO the 1st problem may be caused by UEFI, see e.g. zdnet.com - MINIX: Intel's hidden in-chip operating system or other sites. Just for the records: I have never ever done any "firmware upgrade" on any machine up to now. In each and every case (except one) I was able to boot the devuan from a trustable medium and install it to /dev/sda3. The step from jessie to ascii  (on /dev/sda1) isn't that hard.

CTAKAH wrote:

... I'd like to insert another ssd to install Devuan to. ...

Next step after BIOS is the MBR. As far as I know does  the devuan installer his (or her?) aim well: it places GRUB to the (1) MBR (say /dev/sda) or to your (2) devuan partition (say /dev/sda3). (3) Your Wiń10 doesn't boot from /dev/sda1 anymore? Look at answers.microsoft forums "windows boot loader in Windows 10" or  "How to restore missing dual boot menu in latest Windows 10 build" for example.

The choice between (1), (2) and (3) is somewhat exclusive in a strict sense: You should know what your doing before installation, but there is a lack of documentation.

#13 Re: Hardware & System Configuration » shutdown does not poweroff nor restart » 2019-01-24 00:23:18

According to (sda3)/etc/apt/sources.list this thread is about "Debian Devuan GNU/Linux 1.0 _Jessie_ - Official Release Candidate i386 CD Binary-1 20170504-15:34", a somewhat minimalistic operating system, with entries like:

deb http://auto.mirror.devuan.org/merged/ jessie main contrib non-free
...

Some days ago I've checked "devuan_ascii_2.0.0_i386_desktop-live.iso" from an usb stick: this problem doesn't exist any more. Some hours ago I've placed this on (sda1)/etc/apt/sources.list, and that system can poweroff (at least), reboot (yes) or hibernate (not thoroughly tested) as espected.

deb http://pkgmaster.devuan.org/merged ascii main 
...

Conclusion: this thread is CLOSED.

#14 Re: Desktop and Multimedia » Launch Xorg works on laptop but not on external display » 2019-01-23 23:29:19

Attaching an external screen to a laptop with Xorg may cause strange results. As far as I know have learned the last decade, any problem of this kind can be solved despite there is a lack of documentation. Let's try:

lunario wrote:

... launch X with "startx" ... while I am on the tty (and attach external monitor afterwards) ... Xserver briefly tries to launch but then closes again. ...

It's hard to figure out when (at which point) things went wrong between case #1 "X starts normally" and case #2 "when it does not work". Looking at the beginning of both logs slightly stripped:

lunario wrote:

X.Org X Server 1.19.2
... Markers:
... (--) probed, (**) from config file, (==) default setting,
... (++) from command line, (!!) notice, (II) informational,
... (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
.... (==) Log file: "/home/.../.local/share/xorg/Xorg.0.log", Time: ...

Then there are a lot of lines about what Xorg automagically guesses about keyboard, video and mouse. (KVM you know?). Now read the log of case #2 backwards:

lunario wrote:

... ((snipped)
[  9300.888] (++) using VT number 1

[  9300.891] (II) systemd-logind: took control of session /org/freedesktop/login1/session/_31
... ((snipped)
[  9301.270] (II) Using input driver 'wacom' for 'Serial Wacom Tablet WACf004'
[  9301.271] (EE) systemd-logind: failed to take device /dev/ttyS0: No such device
... ((snipped)
[  9301.271] (EE) xf86OpenSerial: Cannot open device /dev/ttyS0
    Permission denied.
[  9301.271] (EE) Serial Wacom Tablet WACf004: Error opening /dev/ttyS0 (Permission denied)
[  9301.271] (EE) PreInit returned 8 for "Serial Wacom Tablet WACf004"
... ((snipped)
[  9302.040] (II) systemd-logind: releasing fd for 13:70
[  9302.147] (II) Server terminated successfully (0). Closing log file.

Ah, huu ... is there a need to mark the relevant string? IMHO this problem looks like the (really stable:-) devuan/ascii was bitten by some "scope creep". Further investigation needed.

#15 Re: Hardware & System Configuration » shutdown does not poweroff nor restart » 2019-01-12 16:01:46

Thanks for the hint to

ralph.ronnquist wrote:

... the next-to-final step in the shutdown sequence ...

On my system I have found something interesting about differences in /sbin/halt between devuan/jessie and the debian/jessie system used by the LMDE2 (Linuxmint Debian Edition called "betsy"). Some details just for the records:

  • The minimalistic devuan/jessie uses /etc/rc0.d/K09halt or /etc/rc6.d/K09reboot

  • Installing live-tools move this scripts to /etc/rc0.d/K10halt and /etc/rc6.d/K10reboot. BTW: the /etc/rc?.d/K09live-tools are used for /bin/live-medium-eject.

  • The LMDE2 uses /etc/rc0.d/K12halt and /etc/rc6.d/K12reboot instead. And the system does poweroff or reboot as requested!

Due ro my special configuration I can do: (Note: /srv/mnt/hda1/ is mount point for the LMDE2 partition, and /srv/mnt/hda3/ is just a symlink to the current devuan root partition)

  • $ LANG=C diff -s /srv/mnt/hda?/etc/init.d/halt 
    Files /srv/mnt/hda1/etc/init.d/halt and /srv/mnt/hda3/etc/init.d/halt are identical

    Looking into /etc/init.d/halt reveals:

    log_action_msg "Will now halt"
    halt -d -f $netdown $poweroff $hddown
  • $ LANG=C diff -s /srv/mnt/hda?/etc/init.d/halt 
    Files /srv/mnt/hda1/etc/init.d/halt and /srv/mnt/hda3/etc/init.d/halt are identical

    Looking into /etc/init.d/reboot reveals:

    log_action_msg "Will now restart"
    halt -d -f -i

Trying some options according to man halt doesn't help but I found:

$ ls -lgG /srv/mnt/hda?/sbin/halt
-rwxr-xr-x 1 13820 Apr  6  2015 /srv/mnt/hda1/sbin/halt
-rwxr-xr-x 1 13820 Mai 29  2015 /srv/mnt/hda3/sbin/halt

Ok, the devuan /sbin/halt is slightly newer then LMDE2 /sbin/halt. And both are the same size. This is not surprising because devuan/jessie ought to be "compatible" with debian/jessie. But now:

$ LANG=C cmp -l /srv/mnt/hda?/sbin/halt
  409 337 271
((snipped))
  428 362   2

This looks like "recompiling during package building". My /sbin/halt comes from:
Package: sysvinit-core
Version: 2.88dsf-59.2+devuan2
Maintainer: Roger Leigh <rleigh@debian.org>
Homepage: https://git.devuan.org/packages-base/sysvinit

Just for the records: As of this writing I don't how and when this differences between devuan and debian (or lmde?) are introduced. Should I consider this as a "missing feature" of the now oldstable devuan?

#16 Re: Hardware & System Configuration » shutdown does not poweroff nor restart » 2019-01-05 00:25:58

Thank you for the answers and a happy new year for all!

The first glue I've got is to explain the question more briefly: I'm talking about a minimalistic devuan/jessie+xfce on an acer travelmate 2410. In mai 2017 it was straightforward to place the system in /dev/sda3 (20GB) with LANG=de_DE.UTF-8. As times goes by the debian/jessie-derived system on /dev/sda1 (20GB, too) has to be replaced with devuan/ascii+mate as soon as possible. There are two independent problems that I'm trying to understand:

  • (Problem #1): No xfce-user with UID>=1000 is able to halt, restart, hibernate or standby the machine without xfce4-session and upower, the last is introducing 14 more packages. (problem #1 solved by questioning problem #2 ;-)

  • (Problem #2): After the lines "system halted" or "system will restart" nothing more happens! The only thing remaining is press the power button long enough until the machine powers off.

On the second approach I want to reply to each provided answer:

1. bring the system down

Although I'm not a native speaker of english I am aware of reading LANG=C man shutdown, I am able to create an /etc/shutdown.allow for myself, and there are no obvious flaws in /etcinittab for me. The problem just still persists.

2. BIOS and syslog

In response to chris2be8:

How long did you wait before pressing the power button? ...

The system writes "halted" or "restart" and stops working afterwards. Then I have to press the power button long enough for powering the machine off. And later then I have to press the button again to restart again.

Look in the BIOS to see if there are any options that might affect shutdown processing ...

It's hard to believe that's the case: Since 2014 this machine runs "linuxmint debian edition MATE" and does poweroff/reboot/hibernate with sysV init as expected. This problem looks like a difference between debian/jessie (with linuxmint additions?) and devuan/jessie.

Look in syslog etc to see if any interesting messages are there.

Nice point, but which pattern I should grep in /var/log/*?

3. live-tools

ralph.ronnquist wrote:

Basically, live-tools inserts a next-to-final step in the shutdown sequence ...

Yup, there may be missings links in ls -l /etc/rc{0,6}.d/* ... at least in devuan/jessie.

Conclusion

Thanks for your suggestions and updating my knowledge about sysvinit. Up to now this problem is not solved in the oldstable jessie, but there is a good chance the problem does not exists in stable ascii.

#17 Hardware & System Configuration » shutdown does not poweroff nor restart » 2018-12-29 14:44:07

guuml.dev1
Replies: 6

Sitting at a laptop with devuan/jessie both shutdown -h now nor shutdown -r now as root doesn't poweroff the "halted" system nor "restart". The system is just "going down for shutdown", shows the appropriate messages and stops working, but hardware is still running. I have to press the power key additionally long enough to switch off power, and press again for reboot. According to an answer to an old question elsewhere (see
How to make XFCE show the Shutdown Menu?) I have found:

Short solution: Make sure that xfce4-session[1] and upower[2] are installed, and ...

[1] Package: xfce4-session 4.10.1-10+devuan1 comes right out of the box

[2]  but

apt install upower
Paketlisten werden gelesen... Fertig
Abhängigkeitsbaum wird aufgebaut.       
Statusinformationen werden eingelesen.... Fertig
Die folgenden zusätzlichen Pakete werden installiert:
  ethtool hdparm libimobiledevice4 libplist2 libpolkit-agent-1-0
  libpolkit-backend-1-0 libupower-glib1 libusbmuxd2 libx86-1 pm-utils
  policykit-1 powermgmt-base usbmuxd vbetool
Vorgeschlagene Pakete:
  apmd libusbmuxd-tools cpufrequtils wireless-tools radeontool
Die folgenden NEUEN Pakete werden installiert:
  ethtool hdparm libimobiledevice4 libplist2 libpolkit-agent-1-0
  libpolkit-backend-1-0 libupower-glib1 libusbmuxd2 libx86-1 pm-utils
  policykit-1 powermgmt-base upower usbmuxd vbetool
0 aktualisiert, 15 neu installiert, 0 zu entfernen und 0 nicht aktualisiert.
 ... (snipped)
usbmuxd (1.0.8+git20140527.e72f2f7-2) wird eingerichtet ...
Warnung: Auf das von Ihnen angegebene Home-Verzeichnis /var/lib/usbmux kann nicht zugegriffen werden: Datei oder Verzeichnis nicht gefunden
Lege Systembenutzer »usbmux« (UID 103) an ...
Lege neuen Benutzer »usbmux« (UID 103) mit Gruppe »plugdev« an ...
Erstelle Home-Verzeichnis »/var/lib/usbmux« nicht.

It's somewhat obvious that I have missed the point, the problem still persists: As root of my laptop there should be a way to poweroff or reboot via command line.

Please help me with some glue (links to further reading are ok).

#18 Re: Desktop and Multimedia » [SOLVED] Firefox 60 (ESR) poor print quality » 2018-12-16 13:49:56

Starting again with the first assumption:

pcalvert wrote:

...
Note: The problem may possibly be related to this bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=932289

There you found some "Background on pdf.js printing". IMHO there is a bug with javascript in firefox (and thunderbird), but it's hard to track down. Back in 2016 a problem occured under LMDE2 with mate desktop, based on debian/jessie but no ubuntu packages. The problem can be ignored on i386, but is fatal on amd_64. Just to say "Thunderbird kills betsy on the second shot".  Why and how does this affect devuan? Starting from commandline:

firefox 2>firefox.errmsg &
tail -f firefox.errmsg

With devuan/jessie-i386, xfce and firefox-esr (60.4.0esr-1~deb8u1) there are many repeating messages during startup of firefox:

(/usr/lib/firefox-esr/firefox-esr:3466): GLib-GObject-CRITICAL **: g_object_ref: assertion 'object->ref_count > 0' failed
(/usr/lib/firefox-esr/firefox-esr:3466): GLib-GObject-CRITICAL **: g_object_unref: assertion 'object->ref_count > 0' failed

Opening a new tab in firefox there are additional lines:

(/usr/lib/firefox-esr/firefox-esr:3414): GLib-GObject-CRITICAL **: g_object_unref: assertion 'object->ref_count > 0' failed
alloc factor 0,900000 0,900000
alloc factor 0,900000 0,900000

Selecting menu "Print" but not "Print Preview":

** (firefox-esr:2865): WARNING **: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.Avahi was not provided by any .service files
** (firefox-esr:2865): WARNING **: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.Avahi was not provided by any .service files

Visiting both examples (1) and (2) with firefox shows no problems to me. Trying the URL given by pcalvert i get;

Thanks for visiting! GoodRx is not available outside of the United States.

Conclusion: There is an annoying problem with javascript in firefox or in GLib during transition form Gnome2 to Gnome3. This may be fixed in devuan/ascii desktop-live, I will doing a fresh install soon.

#19 Re: Desktop and Multimedia » [SOLVED] Firefox 60 (ESR) poor print quality » 2018-12-14 20:13:29

Under some circumstances there may be a PDF created with "poor print quality". The usage of a file contaning "Portable Document Format" is wellknown, therefore problems belong to three different domains:

  • 1. Creating the file real file with valid PDF-content.

  • 2. Verifying the file content with a previewer.

  • 3. Printing the file to real sheets of paper.

The 1st step: Create your file from your browser not with "Save as ...", but with "Print ...". Then select "print to file" ... (Nowadays every x-window-manager supports this, other os's I don't care). Then choose path and filename at your will.

The 2nd step is to verify the file content with a previewer of your choice.

I'm just assuming, you have saved a file named "pfd.js" instead of printing to a file "pdf.pdf". Realiter the first file is some javascript that looks good to google and to your browser, the latter you can copy to an usb stick and print out anywhere.

#20 Re: DIY » first steps with devuan/jessie on a laptop » 2018-12-12 10:03:58

Update 2018-12-12: basics about xfce and firefox
Update 2018-12-29: changed markup and inserted links to other threads

As noted in my previous post I have made some experiences with GNU/Linux both on servers (even headless) and on desktop/laptop with older hardware. And I have had to learn some lessions about details I was not interested previously. But to keep a long story short: In the not to far future there should be some desktop/laptops booting a devuan/ascii with MATE desktop which are somewhat "easy and straightforward" to maintain the next couple of years.

As of this writing I am using an rather minimalistic devuan/jessie with xfce on 32bit: It starts and runs very fast although there are some details to adjust. You should remember that su - is your friend. BTW: the last few years on LMDE2 betsy I was used to type sudo su. Unfortunately some time ago a guy (new to GNU/Linux but interested to learn) asked me about the differences about su and sudo and why both packages are needed. My answer was not very helpful to him because it was much too detailed ...

According to devuan-cd/devuan_ascii/README.txt there are different flavours of devuan for me to consider:

desktop-live/: A live isohybrid image featuring the standard desktop. This is the recommended choice for desktop users wanting to try Devuan out. It also allows to install the system on your hard-disk using refractainstaller-gui.

The next upcoming step is to replace "the other os" (on /dev/sda1) with "devuan_ascii_2.0.0_i386_desktop-live.iso" with the only help from devuan/jessie on /dev/sda3. This approach helps me to learn more about the "scope creep" introduced some years ago. But there is a drawback:

As far as I remember there was no difference in devuan/jessie between installer-iso/ and minimal-live/. There's a good chance that I found some issues in oldstable that are fixed in stable. FYI some examples:

1. Calling uxterm the first time after installing jessie it comes up with a much to small fontsize for me to read easily. No big thing to fix, just remember the basics:

! 2017-11-20 fontsize increase for UXterm
! afterwards do
! xrdb -merge ~/.Xresources
xterm*font:     *-fixed-*-*-*-18-*

2. A more interesting issue are some messages foung in less +F ~/.xsession-errors:

(xfce4-*): Gtk-WARNING **: Im Modulpfad »clearlooks« konnte keine Themen-Engine gefunden werden,
... repeatet many times

To get rid of messages like this use:

apt install gtk2-engines

Some links worth to be mentioned (at least for me):

3. Is there any reason for looking into ~/.xsession-error? Just in case you're using firefox, firefox-esr or thunderbird you will find many pairing messages:

(firefox-esr:3721): GLib-GObject-CRITICAL **: g_object_ref: assertion 'object->ref_count > 0' failed
(firefox-esr:3721): GLib-GObject-CRITICAL **: g_object_unref: assertion 'object->ref_count > 0' failed

Looks like on 32bit devuan/jessie with xfce (and LMDE2 with MATE) this can be ignored, say:

firefox --Profilemanager 2>~/firefox.errmsg &
less +F ~/firefox.errmsg

But at least on an specific amd_64 with LMDE2 MATE the failed assertions causes heavy havoc: "thunderbird kills betsy on the second shot." -- Ok, as devuan/ascii shall replace the LMDE2 as soon as possible I just ignore this  problem for now. here some may-be-related threads:

#21 DIY » first steps with devuan/jessie on a laptop » 2018-12-10 01:33:10

guuml.dev1
Replies: 1

Hello world,

because this is my first post to the dev1galaxy, my 1st aim is to send thanks to all the people, who have decided to fork debian. THANK YOU for the work you've done over the last four years!  And it's on my personal wishlist in 2018 that devuan continues longer than the next four years (at least;-)

In this post I want to note some topics which are important to me:

  • I'll prefer the debian way to packaging software for many years, it's just rock solid. (No extra server load neccessary during recompiling with important patches.)

  • At the end of the last century there was an request for "*nix to the desktop", long before Mac OS X becomes available.  On behalf to GNU/Linux I've found that Ubuntu has completed the job around 2008. At that time I was able to install and maintain an alternative operating system for others(!) without too much hazzle. There's nobody to blame.

  • In 2014 I was looking for another desktop distribution, just because the

    apt-get dist-upgrade

    (BTW: from Ubuntu 10.04 LTS to Ubuntu 12.04 LTS) does not work like suspected. NOTE: Two years are a very long time for any desktop operating system. You should consider an fresh installation after that period.

  • In Nov 2014 I took my freedom to choose the "LinuxMint Debian Edition with MATE-Desktop" as desktop system not only for myself. This "LMDE 2 Betsy with MATE" works (nearly:-( well up to now. But it reaches "end of life" at 2018-12-31

Most of the above relates only loosly to devuan, but just to say: This post is written with help of

  • acer laptop: an "Intel(R) Celeron(R) M processor , 1.50GHz" with 1GB RAM and (around) 74GB hard disk.

  • an "unknown linux" ... aah ... "unknown" to grub (which misses lsb_release;-). It's a minimalistic devuan/jessie with xfce installed in mai 2017 and up-to-date until now.

  • a shell (in an adjusted xterm), an editor (not pico but nano) and a long-time-used x-www-browser.

And last but not least: I've managed to post here.

Greetings and best wishes to all of you
--guuml

PS at 2018-12-14: Today I found an interesting discussion Devuan/A philosophical diatribe: ... (started at 2018-12-07 02:04:09). Just to say: I'm not a native speaker of english nor an real programmer. But to set a counterpoint with acronyms: I'm a VUGL or at least a VCU in terms of PDP-7. That means "veteran computer user" or "veteran user of GNU/Linux". And this thread is just about how to get "devuan/ascii with mate desktop" running on several machines just with one devuan/jessie currently running. I want to note my first steps.

Board footer

Forum Software