August 2022 - openSUSE Users - openSUSE Mailing Lists

Tensorflow
by Roger Oberholtzer 24 Aug '22

24 Aug '22

Is anyone using tensorflow/tensorflow2 on Leap 15.3? I can only find tensorflow-lite, which is a scaled down version for things like mobile devices. Has it dropped off the radar? In OBS, it seems to be locked for all targets, and I don't see any previous builds anywhere. I thought maybe it was only available for SLE15. But I don't see those packages. Perhaps I don't know where to look for SLE15 packages that might be compatible with 15.3. IIRC, on a recent Leap release (15.0?), tensorflow and such tools were listed as a big part of what the release added. Is that no longer the case? I have looked at contacting the maintainer (https://build.opensuse.org/package/users/openSUSE:Leap:15.3/tensorflow), but last time I contacted someone listed as a maintainer, I got a surprisingly rude response that they were in fact not the maintainer. Despite what was listed in the project. So I'm asking here first... -- Roger Oberholtzer

2 4

Tumbleweed: prevent SSH client from sending environment variables
by Nicolas Kovacs 22 Aug '22

22 Aug '22

Hi, Up until a few weeks ago, I ran OpenSUSE Leap as my daily driver. I only recently moved to Tumbleweed, which seems to get more attention by the OpenSUSE developers. One of the things I did on my Leap system was reconfigure the SSH client in order to *not* send any environment variables. To do this, I opened /etc/ssh/ssh_config and found the relevant stanza: SendEnv LANG LC_CTYPE LC_NUMERIC LC_TIME LC_COLLATE LC_MONETARY LC_MESSAGES SendEnv LC_PAPER LC_NAME LC_ADDRESS LC_TELEPHONE LC_MEASUREMENT SendEnv LC_IDENTIFICATION LC_ALL I simply commented this out and reloaded SSH: # systemctl reload sshd I tried to do this with Tumbleweed, but I don't have an /etc/ssh/ssh_config file. Instead, I only have an empty /etc/ssh/ssh_config.d directory. So I guess my SSH client has all default values defined at build time. What would be an orthodox approach to prevent it from sending any environment variables ? Thanks & cheers from Austria, Niki -- Microlinux - Solutions informatiques durables 7, place de l'église - 30730 Montpezat Site : https://www.microlinux.fr Blog : https://blog.microlinux.fr Mail : info(a)microlinux.fr Tél. : 04 66 63 10 32 Mob. : 06 51 80 12 12

6 11

TW: boot problem kernel 5.18.9-2 and higher
by Martin /Nightowl/ Byttebier 21 Aug '22

21 Aug '22

Hardware: Dell XPS 8940 11th Gen Intel i7-11700 (16) @ 4.800GHz NVIDIA GeForce RTX 3060 Ti I'm using latest TW-release with kernel 5.19.0-1-default and NVIDIA driver 515.65.01 installed the hard way. kernel 5.18.9-2 or higher seems to give a (plymouth) problem. The problem doesn't occur with kernel 5.18.9-1-default or lower. Booting up with one of these kernels the plymouth boot screen stops displaying after this line: I915 0000:00:02.0: enabling device (0006 -> 0007) Eventually I get to the KDE login screen. The problem is that I'm using the hard way to install NVIDIA. This makes I can't boot when a new kernel is installed as the boot process ends up with a black screen. Luckely I can access the faulty pc from another pc and install the driver from there. journalctl show me this: Aug 12 18:17:32 dell-desk sddm[1381]: Initializing... Aug 12 18:17:32 dell-desk sddm[1381]: Starting... Aug 12 18:17:32 dell-desk sddm[1381]: Logind interface found Aug 12 18:17:32 dell-desk sddm[1381]: Adding new display on vt 7 ... Aug 12 18:17:32 dell-desk sddm[1381]: Loading theme configuration from "" Aug 12 18:17:32 dell-desk sddm[1381]: Display server starting... Aug 12 18:17:32 dell-desk sddm[1381]: Running: /usr/bin/X -nolisten tcp -auth /run/sddm/{765f2352-11a7-428b-9a7a-55f8fef85402} -background none -noreset -displayfd 17 -seat seat0 vt7 Aug 12 18:17:33 dell-desk sddm[1381]: Failed to read display number from pipe Aug 12 18:17:33 dell-desk sddm[1381]: Attempt 1 starting the Display server on vt 7 failed Aug 12 18:17:35 dell-desk sddm[1381]: Display server starting... Aug 12 18:17:35 dell-desk sddm[1381]: Running: /usr/bin/X -nolisten tcp -auth /run/sddm/{765f2352-11a7-428b-9a7a-55f8fef85402} -background none -noreset -displayfd 17 -seat seat0 vt7 Aug 12 18:17:35 dell-desk sddm[1381]: Failed to read display number from pipe Aug 12 18:17:35 dell-desk sddm[1381]: Attempt 2 starting the Display server on vt 7 failed Aug 12 18:17:37 dell-desk sddm[1381]: Display server starting... Aug 12 18:17:37 dell-desk sddm[1381]: Running: /usr/bin/X -nolisten tcp -auth /run/sddm/{765f2352-11a7-428b-9a7a-55f8fef85402} -background none -noreset -displayfd 17 -seat seat0 vt7 Aug 12 18:17:37 dell-desk sddm[1381]: Failed to read display number from pipe Aug 12 18:17:37 dell-desk sddm[1381]: Attempt 3 starting the Display server on vt 7 failed Aug 12 18:17:37 dell-desk sddm[1381]: Could not start Display server on vt 7 Any idea how I can fix this? Martin

10 49

VirtualBox Puel Crashes After Latest Updates on 15.4 Qt related.
by David C. Rankin 18 Aug '22

18 Aug '22

All, I am running Virtualbox 6.1.36 PUEL version (I need the USB). It has run fine for years and years (since version 3.X). Today I went to start it and it would not load. Starting from the command line as my user, I get: VirtualBox Qt FATAL: FATAL: The application binary appears to be running setuid, this is a security hole. Aborted Huh? WTF? Obviously it is unhappy with Qt or vice versa. This worked fine a couple of weeks ago. What to check to narrow down the issue? I can su and then VirtualBox runs, but since it is running under root, none of my VM's are found. Running with sudo -E I receive: No protocol specified No protocol specified Qt CRITICAL: This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem. Available platform plugins are: eglfs, linuxfb, minimal, minimalegl, offscreen, vnc, xcb. Now running su'ed to root, Virtualbox starts and the dialog appears, but no vms are shown. The following warning is shown on the command line: Qt WARNING: QStandardPaths: runtime directory '/run/user/1000' is not owned by UID 0, but a directory permissions 0700 owned by UID 1000 GID 1060 What has changed with Qt these past two weeks that is causing VirtualBox to freak out? -- David C. Rankin, J.D.,P.E.

1 1

Permissions of /usr/local/bin
by Siard 17 Aug '22

17 Aug '22

/usr/local/bin is owned by root with permissions 755, so as a user I cannot put a script in there. I can change the ownership, but every time after a reboot it is owned by root again, with permissions 755. Does anyone know where this is set, and/or how I can make /usr/local/bin permanently make writable for a user?

12 25

missing kernel source for 5.18.12-lp154.5-default
by Boyd Lynn Gerber 16 Aug '22

16 Aug '22

Hello, I'm trying to get my VM Workstation 16 Pro working. I need the source to compile vmware-host-modules-workstation-16.2.4. I booted this kernel. Where may I get the source? All I have is 5.14.21-150400.24.18-default Regards, -- Boyd Gerber <gerberb(a)zenez.com> 801 849-0213 ZENEZ 1042 East Fort Union #135, Midvale Utah 84047

3 2

NetworkManager
by Wolfgang Mueller 16 Aug '22

16 Aug '22

I have to computers, - "hewpac" with Leap 15.2 installed and - "azure" with Leap 15.4, both working under KDE and connected to the router via WiFi. For a long time, both worked under the so-called Wicked connection. But yesterday in the morning, this type of connection failed. Consequently I had to select NetworkManager for having them connected. After an hour, the NetworkManager connection of "azure" (Leap 15.4) failed, and since then it cannot be re-established any more. When i click the "Networks" mini-icon in the panel, four routers in the neighborhood are listed under "Available connections", but mine is not. What can I do for making it find my router again? Thank you in advance for your help. Best regards, Wolfgang

8 13

Build libheif to support heif-convert of Apple .heic files to .jpg/.png (and enable open by Gimp)
by David C. Rankin 12 Aug '22

12 Aug '22

All, If you have ever been stuck with a native iPhone .heic file on Linux it is a pain in the .... However, you can simply rebuild the current source package for 15.4 (or 15.3, etc..): http://download.opensuse.org/source/distribution/leap/15.4/repo/oss/src/lib… using rpmbuild and the "--with x265" option to enable building with libde265-0 and libx265-199 (both from packman) which will build the heif-examples package that provides: /usr/bin/heif-convert /usr/bin/heif-enc /usr/bin/heif-info /usr/bin/heif-thumbnailer You will also need to install the corresponding -devel packages as well along with librav1e0 and rav1e-devel, and libaom3 and libaom-devel. The list of subpackages created using the x265 option are: gdk-pixbuf-loader-libheif heif-examples libheif-devel libheif1 The available gimp-plugin-heif, allows Gimp to open .heic files directly after you have build and installed the updated version. You will also need to create a *.heic *.HEIC file association for whatever desktop you are using in order to enable the rt-click context menu "Open with Gimp". Primary use I have is simply to convert the same .heic files I get sent. After the updated package is installed, it's as simple as: $ heif-convert file.heic file.jpg I hacked a spec together which worked swimmingly, to ask on the packaging list about handing a couple of conflicts only to find the conditional build in the current openSUSE srpm already provided a path. Hopefully this will save some other soul the same grief. -- David C. Rankin, J.D.,P.E.

2 6

[oS-en] Recovering dead text file
by Carlos E. R. 11 Aug '22

11 Aug '22

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, Yesterday I powered down the computer (not hibernated as usual). I was editing a text file with jstar (aka joe) and told it to save the file. But I failed somehow, because yesterday edit is not there, is lost. Not a big deal, I know what it was and can redo. If I try to edit the file again, it knows: Locked by cer(a)Telcontar.403. (S)teal lock, (I) edit anyway, (Q) cancel edit? I could just steal lock, but I'm curious whether I might recover the lost edit. Where is it, if it is somewhere? The edit command line is called: jstar howto I see in the last files of the home dir: lrwxrwxrwx 1 cer users 17 Aug 9 19:55 .#howto -> cer(a)Telcontar.403 - -rw------- 1 cer users 4707 Aug 10 12:01 .joe_state But those are not the edited file copy. And "cer(a)Telcontar.403" does not exist. It is not in tmp: cer@Telcontar:~> ls /tmp/j* ls: cannot access '/tmp/j*': No such file or directory cer@Telcontar:~> So, does jstar, aka joe, save a temporary file with the edit somewhere? There is a "JOE" file in home dir with old aborted edits, but it is dated 2007. - -- Cheers Carlos E. R. (from 15.3 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCYvOF4Rwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfV18oAnAxanSh2BLy7rvV009og iZ/apGGZAJ9qDr/ZMjfGIOgBqn7YFZlBbwWBZg== =ik6S -----END PGP SIGNATURE-----

5 10

logrotate, systemd, stopped rotating a log
by joe a 11 Aug '22

11 Aug '22

Leap 15.3. Some time back got logrotate working to rotate fetchmail logs. Found that it seems to be no longer rotating that file. I noticed the logrotate.conf file is empty. Before delving any further I wanted to assure or disabuse myself of the notion that it matters not, as long as the proper configurations are in /etc/logrotate.d/def_for_my_log.

5 13