Testdriving daily build 24.04 (Noble Numbat)

Thoughts & Feedback
, ,
21

I checked on my three test-laptops with several mice and I can't reproduce what you are experiencing.

Since the USB ports are completely owned by the kernel and the mouse routines by the Xserver I very much doubt that the problem is the MATE-desktop.

Although I strongly suspect a conflict between the newer kernel and your hardware, I advice you to file a bug here anyway:

One thing left to try though: You can completely disable USB powersave in the kernel. The following link explains how to do that:

If this doesn't work then the problem is probably something else, though at the moment I can not imagine what that could be. :slight_smile:

3 Likes
22

Great,
the grub command line entry works.
Now USB and BT works even without ac power.
Thank you

3 Likes
23

Dumb question, but is the UM 24.04 LTS release still scheduled for tomorrow the 25th ?

W

24

Hi, @Watford.

That was NOT a dumb question, at all! :slight_smile: As it turns out, Ubuntu MATE 24.04 LTS ("Noble Numbat") was indeed released today (25th April 2024):

3 Likes
25

More info about the "UBSAN: array-index-out-of-bounds" error

It has been fixed !! :partying_face:

2 Likes
26

Since booting the official 24.04 LTS ISO seems to be not successful for everyone I decided to do a fresh install of Ubuntu-MATE 24.04 LTS on my old Thinkpad T420.

I discovered that it was abit less smooth sailing than the "daily ISO + updates"

Directly after a cold boot, I get:

error: file '/boot/' not found

but it starts anyway :slight_smile:

plymouth changes size during this boot, probably because framebuffer is, as usual during boot, changing resolution.

It reaches the desktop no problem.

entering the installer:
enabling WiFi and logged on to my homenetwork
choosing full install
The installer behaves sometimes suspiciously sluggish, I will have to test this on my Toshiba C670D craptop.

Small visual glitch detected:
choices don't stay highlighted for the first few questions (at least, not visibly) but it works nevertheless.

Once installed everything was ok so after a global update I started exploring the system a bit.
I encountered some new bugs that I haven't encountered before (see end of this post)

I checked the new appcenter, or should I say snapcenter ? I really don't know.
It rubs me the wrong way that it is virtually impossible to discover which of these apps will run as snap and which wont. This information is important to me, and it is missing.

Here is a list of bugs I encountered that I will report in due time (launchpad):

  1. Locale bug 1 (./config/user-dirs.dirs):
    Foldernames in homedirectory do not follow locale translation because xdg-user-dirs-update does not work from autostart. Even manually invoked and with a lot of luck it only does the job partly.

  2. Locale bug 2 (ayatana-date):
    Day and month are swapped in my locale (nl_NL). Luckily I could work around this by enabling custom settings using dconf-editor. (I could also have used the old panel-date instead.)

  3. Appcenter bug:
    visually horribly glitchy, switching compositing off mitigates this.

On the other hand, I experienced no bootproblems at all :slight_smile:
So far so good. :innocent:

1 Like
27

For reasons mentioned above I also decided
to do a fresh install of the official Ubuntu-MATE 24.04 LTS ISO on my old
2011 Toshiba C670D craptop. And oh boy, was I in for a bumpy ride.

booting from USB works OK

wifi took two tries to connect
connecting took a long time.

installer crashes first time
journalctl mentioned [FAILED] to start snapd.service
second time it works

I suspect the networkconnection misfire and the crash were caused by either

  1. not checking for ready but just using a delay time
  2. or timing out on loading.

That, in itself, worries me a lot.

As I suspected (see post above) the installer itself is indeed a bit of a pig:
it took extremely long time, several hours, for a minimal (core) install. It is glitchy, massive and slow as molasses with extreme GUI latencies.

I know my craptop, it is slow, but i know how this hardware behaves under different kinds of loads:
This installer behaves like a badly optimized pythonscript as if it is instantiating and destroying thousands of objects a second willy nilly while running on an emulator or a non-hardware supported virtual machine.
This, also, worries me a lot.

The daily ISO + updates that I used until now were a very smooth experience. This final product, not so much.

In my opinion the installer has to mature a bit.
I've also read somewhere in this forum that this installer is a snap.
I understand the design decision but practically speaking this can turn out a bit troublesome in some cases.

Anyway, after a bumpy install it runs smoothly. :slight_smile:
I think I also will try this on my Asus F3T (build in 2005). :innocent:

2 Likes
28

Fresh install on Asus F3T (build in 2005)

perfect boot
Needed to activate WiFi twice before it worked.
started the installer -> installer crashed

I discovered that the installer was checking the ISO.
This process continued after the installer crash, I had to kill it in order to start the installer again.

Second time installer worked. It also reacted a bit sluggish at times. Nevertheless, it was finished within an hour :slight_smile:

Language settings were not respected. Had to install local language manually (which the installer forgot to do).

After running update and a reboot: language settings were ok and the foldernames in $HOME were changed to the intended foldernames according to the locale. :slight_smile:

No real probles so far, and it runs like a dream :innocent:

2 Likes
29

:partying_face: Pointrelease is released, yay !!:partying_face:

So I decided today to upgrade my mediacenter from 22.04 to 24.04:

Machine:
  Type: Desktop Mobo: ASRock model: J4205-ITX serial: <superuser required>
    UEFI: American Megatrends v: P1.80 date: 05/21/2018
CPU:
  quad core Intel Pentium J4205 (-MCP-) speed/min/max: 1035/800/2600 MHz
  TDP: < 10 Watt / Fanless
Kernel:
  6.8.0-41-generic x86_64 Up: 5m Mem: 868.8 MiB/7.17 GiB (11.8%)
Graphics:
  Device-1: Intel Apollo Lake [HD Graphics 505] driver: i915 v: kernel
  Display: server: X.org v: 1.21.1.11 driver: X: loaded: modesetting unloaded: fbdev,vesa
    dri: iris gpu: i915 tty: 104x55 resolution: 1: 1920x1200 2: 1280x1024
  API: EGL v: 1.5 drivers: iris,swrast platforms: surfaceless,device
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: mesa v: 24.0.9-0ubuntu0.1 note: console (EGL sourced)
    renderer: Mesa Intel HD Graphics 505 (APL 3), llvmpipe (LLVM 17.0.6 128 bits)

I had my fingers crossed as usual because I use VAAPI hardware acceleration for videoplayback and that means manual intervention after every release upgrade because sometimes libva / libmesa related stuff is missing/not-installed/version mismatch and sometimes mpv changes implementation of options which forces me to modify my mpv.config c.q. config-generator.

The Upgrade

screenlocking

Before the upgrade process really started, the system complained that it couldn't deactivate screenlocking without so much as a hint how that should be done.

I still haven't the slightest idea what the installer wanted me to do, but since I do not use a screensaver or a display time-out on this home brew appliance I deemed it OK to continue the upgrade.

cryptsetup ?!?

During the upgrade, the installer complained about not being able to install cryptsetup ( which I don't use and don't care about anyway ) but it finished beautifully nevertheless.

The first update afterwards, the update-manager complained about cryptsetup again and wanted a partial upgrade.

Big Fun :partying_face:, because as usual, the update-manager starts happily doing that and is completely oblivious that it is always failing to succeed in the end :wink:
I ended up doing it from the commandline which worked flawlessly.

settings were preserved

All my settings were preserved i.e. it was exactly the same as before ( except for some minor panel-app related options ).
( @mickee , you are good to go :wink: )

core task: mediacenter

VAAPI (hardware assisted video decoding on intel)

Did not work immediately because allegedly some codecs moved from the intel-media-va-driver to the intel-media-va-driver-non-free which took me some time to discover.

After installing the intel-media-va-driver-non-free everything video worked as expected.

audio

I was prepared to face some pipewire troubles but all audio worked out of the box, including S/PDIF and multichannel DTS and AC3 over optical fiber (toslink/lightpipe).

conclusion

This was seriously quite a painless upgrade. :slight_smile:

3 Likes
30

So, was it this hint from Thom that was the specific fix? or was it something else?