The officially official Devuan Forum!

You are not logged in.

#1 Re: Devuan Derivatives » Musl based Devuan derivative? » 2023-02-01 11:11:46

Why?

Thats for customers of virtuozzo, not for end users. My contract is with my provider and not with virtuozzo. So i would have to open a(nother) call with my provider, maybe regrding this (apparantly not possible in an running template even if the conatainer is not running).

Which might lead me to the conclusion that either hacking the glibc version test which should be harmless because the actual kernel api is a lot newer than the apparent one because the kernel is actually a lot newer might be a way out. One that i really do not like at all.

#2 Re: Devuan Derivatives » Musl based Devuan derivative? » 2023-01-31 17:26:23

Mhmm the question was actually not sticking to alpine which i intend to install on a 2nd old machine (replacing chimaera) but that very specific virtuozzo instance which cannot be upgraded beyond ASCII (at least not without dirty tricks). And ASCII has already passed EOL. And that on a production server visible to the world.

#3 Devuan Derivatives » Musl based Devuan derivative? » 2023-01-31 00:22:39

beate
Replies: 5

Hi,

in order to be able upgrade a virtuzzo based vserver running Devuan ASCII 32bit i need something with a different libc than the  glibc provided by the newer Devuan releases.

Reason: the installation is old, the container has been created unter virtuozzo 6. It runsmy main web and email server - so it is mission critical.  After migrating the container to virtuozzo 7 (by my provider), the kernel shows up as 3.10.x. In order to be able to upgrade from ASCII i need a libc working with that given kernel API. However, Beowulf and Chimaera's libc both require at least an api as of kernel version 3.20.x.

AFAIK, musl libc has more relaxed demands on the kernel API. So if there was some Devuan based on musl that would open a possible upgrade path.

THX for any hints.

BTW: i just installed Alpine linux on an old eeePC 1016P with an atom n455 cpu and am impressed on the performance and moderate memory footprint. Even firefox is usable ... thats another reason to switch from glibc to musl libc.

#4 Re: Desktop and Multimedia » Audacity and Musescore Spyware problem » 2021-07-06 01:04:55

I'm aware that this will affect version 3 and not the current one. Nevertheless i set audacity on hold on 4 machines (all running chimaera).

Firejail and the like is nothing for unexperienced users - simply because they do not know or will not dare to use it.

But anyway: the other affected software is musescore. Chimaera provides Musescore 2 and 3, the first due to problems regarding backward compatibility.

I would guess that version 2 is probably "clean" but version 3 ? Or will the new policie affect only recent versions? BTW: the recent android versions of musescore have already been changed - a friend of mine told me that it was unclear wether she could use it free of charge.

IMO version 2 should be prevailed as long as possible. Regardles how long it'll be maintained by Debian.

#5 Desktop and Multimedia » Audacity and Musescore Spyware problem » 2021-07-05 13:14:19

beate
Replies: 33

I just read about a new issue on Audacity and Musescore containing spyware. It looks as if it would be time to supply forks of the packages. I just read about tthat one: https://github.com/cookiengineer/audacity

And for the time being devuan should IMHO be restricted to the latest version of audacity prior to the introduction of spyware elements. Which does not seem much of an issue to me as even old versions of audacity are functional and rock solid.

What do You think?

#6 Re: Freedom Hacks » To permanently block installation of Pulseaudio. » 2021-06-29 15:45:03

... works until the next upgrade to chimaera.

And then murphy will predict that You forget to update the release entry before doing the update.

#7 Re: Desktop and Multimedia » hydrogen problems (beowulf / chimaera) » 2021-04-13 13:47:00

These packages were all installed. I installed the midi apps and soundfonts. This did not remove the warning on missing soundfonts, but starting the mixer will not crash hydrogen anymore. For whatever reason. If i happen have a free moment, i'll check how to setup the midi stuff with hydrogen.

(But first of all i need to deal with the bad news that i will lose my home temporarily for about 1-2 months due to a larger water damage in the building of my appartment. )

#8 Re: Desktop and Multimedia » hydrogen problems (beowulf / chimaera) » 2021-04-12 16:03:35

EDIT: the mixer problem appears to be due to an incompatibility with the rests of the old installation (as far as i remember, it could have been 0.96 or even older). Which looks like a bug in hydrogen itself and not an issue with the packaging.

#9 Desktop and Multimedia » hydrogen problems (beowulf / chimaera) » 2021-04-12 15:46:50

beate
Replies: 4

Hi folks,

i just installed hydrogen on two machines, one running beoulf, the other one running chimaera.
The chimaera installation has (important) rests of a previous installation in local .hydrogen, the one on beowulf was native.

Problem specific to the chimaera install:

crash as soon as the mixer is opened.

Problems to both installations mainly concern filesystem permissions regarding drumkits:

(E) Filesystem::check_permissions Audiophob is not writable
(E) Filesystem::file_copy unable to copy /usr/share/hydrogen/data/drumkits/Audiophob/drumkit.xml to /usr/share/hydrogen/data/drumkits/Audiophob/drumkit.xml.bak, /usr/share/hydrogen/data/drumkits/Audiophob/drumkit.xml.bak is not writable                                                                     
(E) XMLDoc::write Unable to open /usr/share/hydrogen/data/drumkits/Audiophob/drumkit.xml for writing

... and so on for all other drumkits in  /usr/share/hydrogen/data

To me it is asolutely correct that these paths are not writable to unprivileged users !!

Why in hell does hydrogen expect to be allowed to write into folders belonging to the installation ???

Both installations also moan regarding missing midi notes for these drumkits. Where do i find these?  There seems no appropriate package available.

THX

Beate

#10 Re: Off-topic » deb.sury.org now requires systemd » 2021-02-13 16:47:26

As this problem will persist on the future - what is the current status of supporting php within devuan themself? And what about having a closer look to the BSD people, notably those of NetBSD and OpenBSD, who also try hard to remain as systemd-free as possible and thus have to deal with similar problems.

I'm aware that this will be a wide field because maintaining php will also affect web server package maintenance - is there any way to cleanly draw separation lines between these fields (i.e. reduce the number of dependencies)?

I would like to avoid having to go back to compile the whole web server related stuff from sources (which i had to do some years ago due to too many undocumented changes in the setup of all this stuff in debian - a nightmare to people who just want to run their web servers).

#11 Re: Other Issues » php-fpm in chimaera current » 2021-02-13 10:43:06

Gives some insight into the situation - but the alternative packages still seem beowulf only. Which would be an option after reverting the switch to chimaera on the 2nd machine (which is stuck to the 2.6 kernel anyway) bot not for now. But many thanks.

#12 Other Issues » php-fpm in chimaera current » 2021-02-13 02:51:47

beate
Replies: 3

I'm just trying to set up nginx and php-fpm on a machine running a freshly updated chimaera.
PHP version is 7.4.15-1

And this because the recent php-fpm has a dependcy on systemd (which seems to have been correct in a previous version of chimaera still offering php 7.3 - that machine actually runs beowulf and the upgrade failed because the vserver still has a 2.6 kerne (and hence i would like to go back to beowulf)).

Do i need to recompile php or is there any way to get back to the php 7.3-packaged of chimaera (i need that asap, and can live with any version of php since the latest 5.6
...)

#13 Re: Hardware & System Configuration » Old Brother Scanners broken in Beowulf (problem descr. and workaround) » 2020-04-18 11:03:05

Which will not help the users with the old machines. We actually need a(nother) update of the old libs. I also tried to compile the sources of brscan2 - but You have to be relatively deep in sane development to be able to do this (and i didn't have time, needed my scanner urgently to apply for Corona help).

The brscan sources look to me like incomplete sources of libsane utils with the actual driver libs provides as binary libs. So any sane developer should be able to integrate the latter into the new source stream and - if lucky - compile it with those libs or - less luckily - write a wrapper lib in order to match APIs.

#14 Hardware & System Configuration » Old Brother Scanners broken in Beowulf (problem descr. and workaround) » 2020-04-18 00:03:11

beate
Replies: 4

Hi there, i just want to notify You on a problem with many older Brother scanners, noably those depending on the brscan2 and brscan3 drivers. These drivers cannot be used with the sane version provided by beowulf any morel - libsane 1.0.27. There seems to be an API change from sane 1.0.25 to 1.0.27 which will lead to surprising "undefined symbol" errors in the sane tools.

THIS IS A PROBLEM WITH SANE, NOT WITH BEOWULF. You should and do see this bugs in other Linuxes unsing sane 1.0.27. See for example here

https://forum.ubuntuusers.de/post/9011933/

and here

https://forum.ubuntuusers.de/post/8938890/

I had exactly the same error message in my Beowulf system after installing the brscan2 "driver" from Brother.

A little bit of background (which i found out when i debugged that situation): bracan2 and brscan3 as well are nut just scanner drivers but actually full replacements for libsane.

You see it using ldd on a sane application - here on my system with the workaround: libsane is taken from the brscan2 place and not from the installation path of sane itself. This will only work if libsane coming with the brother driver is the same as the API from the system provided libsane. As soon as these don't match You'll see strange messages like in the links above.

(Sidemark: this also explains why brscan2 and brscan3 or scanners of different brands cannot be used together in a single system; brscan2  simply appears to be a dedicated version of libsane set up with just the set of supported scanners.)

 ldd /usr/bin/scanimage
        linux-vdso.so.1 (0x00007ffc955cc000)
        libsane.so.1 => /usr/lib64/sane/libsane.so.1 (0x00007f9bf1daa000)
        libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f9bf1da5000)[/b]
        libv4l1.so.0 => /usr/lib/x86_64-linux-gnu/libv4l1.so.0 (0x00007f9bf1d9e000)
        libieee1284.so.3 => /usr/lib/x86_64-linux-gnu/libieee1284.so.3 (0x00007f9bf1b92000)

WORKAROUND:
my scanner had worked perfectly in ASCII. So i purged the sane stuff altogether and built a suitable older sane version from the raw sources, specifically 1.0.24 as used in ASCII. I installed and set up the brscan2 package from Brother as usual (udev rule, DLL search path) - et voila, it works!

In order to avoid this to be overridden by future updates i also made dummy packages for libsane, libsane-common, sane-utils providing a large version number. The Beowulf packages sane and xsane are still compatible with sane-1.0.24, so i installed them as usual but set them on hold. (which means that in future i'll also have to build them manually...)

Again: this will hurt any user of older Brother scanners in ANY linux distro, not just Devuan. But I'd like to explain it here before it beats others.

#15 Re: Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-03-03 04:34:34

I checked, apt-get full-upgrade did nothing additional, which means that the upgrade was complete.
On machine b) - a notebook - the problem disappeared after i deinstalled laptop-mode-tools. And with that, some other problems. On machine a) the problem also disappeared but i don't know why. I temporarily ran the box from its old disk which carries an i386-ASCII upgraded from Jessie installation. After reinserting the new drive with the Beowulf, the problem was gone.

#16 Re: Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-02-26 16:28:01

Update: upgrades the i386 machine to beowulf and installed kernel 5.4.0 from the backports. Problem reproducable. I wonder if this is a problem with Devuan or with the underlying Debian.

Noch ein update: mit meiner ebenfalls älteren USB-Soundkarte tritt das Problem nicht auf.

#17 Re: Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-02-26 15:20:37

This is a fairly old device, and i have installed it successfully many times... and it was working before the upgrade.

There is actually an easy workaround - i still have  my old ASCII installation on an older disk drive (which i replaced two days ago). But it does not feel good knowing it will not be upgradable.

#18 Re: Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-02-26 14:39:37

I just tried on the oldie: result negative:

[  140.602090] usb 2-1: new full-speed USB device number 5 using uhci_hcd
[  140.795381] usb 2-1: New USB device found, idVendor=04f9, idProduct=0180, bcdDevice= 1.00
[  140.795392] usb 2-1: New USB device strings: Mfr=0, Product=0, SerialNumber=3
[  140.795399] usb 2-1: SerialNumber: 000J7J739526
[  140.802725] usblp 2-1:1.0: usblp0: USB Bidirectional printer dev 5 if 0 alt 0 proto 2 vid 0x04F9 pid 0x0180
[  140.809535] udevd[1913]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1 2 5': No such file or directory
beate@knuell:~$ uname -a
Linux knuell 4.19.0-0.bpo.6-686 #1 SMP Debian 4.19.67-2+deb10u2~bpo9+1 (2019-11-12) i686 GNU/Linux

which means i would have go back to jessie at least temporarily on one machine in order to be able to use my printer and scanner (or give NetBSD a try - i hate that - inoperable printers cost an extreme amount of time...)

#19 Re: Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-02-26 14:25:35

xinomilo wrote:

try installing libmtp-runtime and check again..

has been installed.

I just did an upgrade to an old i386 machine (ASCII+backports) and installed xsane+cups there. I noticed that udev had been replaced by eudev. Same problem after the upgrade.

regarding the linux image: on another machine (notebook)  the proposed 4.9.18-06  kernel lead to an unbootable system unless i also installed linux-firmware-free. Which is not a dependency. But i'll give it a try

#20 Hardware & System Configuration » UDEV problem - Brother MFC7420 » 2020-02-26 13:27:15

beate
Replies: 10

Hello,  i reinstalled my desktop machine with Beowulf (as an upgrade from ASCII 2.1)

Everything seems to work fine, with one severe exception: my MFP printer is not accessible, neither as a printer nor as a scanner.

The following error message does occur after the device is detected normally:

[72188.051622] usb 1-1.3: new full-speed USB device number 11 using ehci-pci
[72188.165032] usb 1-1.3: New USB device found, idVendor=04f9, idProduct=0180
[72188.165036] usb 1-1.3: New USB device strings: Mfr=0, Product=0, SerialNumber=3
[72188.165038] usb 1-1.3: SerialNumber: 000J7J739526
[72188.167058] usblp 1-1.3:1.0: usblp0: USB Bidirectional printer dev 11 if 0 alt 0 proto 2 vid 0x04F9 pid 0x0180
[72188.168512] udevd[10218]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.3 1 11': No such file or directory

and here some contents of /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1 :

drwxr-xr-x 9 root root     0 Feb 26 13:51 1-1.1
drwxr-xr-x 8 root root     0 Feb 26 13:51 1-1:1.0
drwxr-xr-x 7 root root     0 Feb 26 13:51 1-1.3
-rw-r--r-- 1 root root  4096 Feb 26 13:51 authorized
-rw-r--r-- 1 root root  4096 Feb 26 13:51 avoid_reset_quirk
...

I verified that this is actually the MFC7429.
As .../1.1.3 appears to contain all the subdevices of the printer. So my question: did the device naming scheme change recently? Should i switch back to an older kernel until this is represented in udev/mtp ?

BTW: here the output of uname -a

Linux merapi 4.9.0-11-amd64 #1 SMP Debian 4.9.189-3+deb9u2 (2019-11-11) x86_64 GNU/Linux

#21 Re: Off-topic » Life after Firefox » 2020-02-25 23:55:26

beate wrote:

On my very new notebook i (64bit ASCII) i tried Iridium, but failed due to a library version mismatch - the binary is made for Ubuntu.

Today i gave it a try after upgrading to beowulf. Well, to make it short: it crashed.

reminds me of the shortest version of ancient Netscape 2 ...

#!/bin/sh

touch core
/bin/echo "Segmentation fault (core dumped)" >&2
exit -1

Back to the audio problem: on my systems (both beowulf as well as ascii) the browsers work directly on card0 - and they mix: chromium, palemoon, qupzilla, vlc

No additional config required. libpulse is pulled in by the dependencies and surely used (but vlc was a self made ALSA version without anything pulse).

The only thing that must be ensured if an additional high quality sound card is used: it must be assigned by the kernel as card0. Even if the onboard sound is disabled in BIOS (my desktop plays through a  mixing console attached through USB), additional cards are by default not mapped as card0.

(i could not check the behavior with multichannel devices).

#22 Installation » (ASCII) Vserver hangs after upgrade of sysvinit » 2019-11-04 15:37:22

beate
Replies: 1

environments

machine a:

Linux v165062 2.6.32-042stab140.4 #1 SMP Fri Oct 11 11:36:17 MSK 2019 i686 GNU/Linux

sysvinit and related programs: 2.88dsf-59.9+devuan2_i386

machine b:

Linux v165051 2.6.32-042stab140.4 #1 SMP Fri Oct 11 11:36:17 MSK 2019 i686 GNU/Linux

sysvinit and related programs: 2.88dsf-59.2+devuan2_i386
dependencies all on latest patch level like machine a, well except the selinux lib.

You see two vservers in an old OpenVZ environment hosted on the same physical hardware. Both started as Debian 7 installations and were upgraded through Debian/Devuan lenny.

Machine a(!) hangs during init, machine b boots just fine. During the upgrade of machine a, sysvinit-core did not install correctly. I reinstalled manually from the system console.
Strange enough, sysvinit-core was simply not upgraded on machine b - apt's che provides the younger binary, not the installed version.

I found a dirty workaround to get machine a up and running:

/etc/init.d/networking stop
/etc/init.d/networking start
then start the contents of /etc/rc3.d in a for loop.

I am looking for a hint to make machine a a fully bootable again. Downgrade sysvinit-core and all its dependencies? How?
Better suggestions?

BTW: both machines are productive servers, so rebooting is a problem. Especially on machine b.

#23 Re: Off-topic » Life after Firefox » 2019-03-26 00:05:59

yeti wrote:

This one?

https://web.archive.org/web/20190325194 … evuan2.png

Debian/Sid has it and a quick and dirty rebuild on ASCII is doable.

Yes, that one.

Falkon is going to be integrated into KDE, and hence Qupzilla will die,  at least as a standalone browser.

There is a significant community outside the KDE/GNOME/XFCE folks, and these (or maybe especially these?) might be the main community for the smaller browser projects independent on any desktop.

Anyway, regardless if someone likes midori or not - it is part of debian and hence devuan. In a meanwhile old version. A reasonably well tested port might be a serious candidate for the backports, won't it?

But anyway, i think i'll try a build and check the sid patches before, as soon as i have the time.

#24 Re: Off-topic » Show your desktop (rebooted) » 2019-03-25 19:02:11

No picture but a description. Windowmaker. Style "Golden". Background suitable image or just "wheat shades". Terminal urxvt with X11 font color "gold", background a really dark gray. wmtime as the only dockapp, well on the notebook wmbattery, too.

Still clean and elegant after all these years (i am using wmaker since version 0.17)

#25 Re: Off-topic » Life after Firefox » 2019-03-25 18:42:01

Well, we all have our preferences. Here on my large ASCII desktop (64bit hardware with 32 bit OS) i am running qupzilla (pretty unstable...), chromium and palemoon more or less permanently and midori for specific tasks where i want to ensure that everything is cleared after the session (banking...).

In NetBSD/i386 midori is the only acceptable browser - native alternative is just an old and slow version of firefox... but anyway, midori is an interesting project i am following (and using) for quite a while. And midory 7 uses a much cleaner GUI that 0.5.11 ... so it would be really nice if there was at least a debian package..

On my very new notebook i (64bit ASCII) i tried Iridium, but failed due to a library version mismatch - the binary is made for Ubuntu.

So in conclusion, there seems to be the need of a free, not too feature overloaded web browser, ideally independent on any specific GUI environment.

Board footer

Forum Software