Language Selection

English French German Italian Portuguese Spanish

openSUSE 10.2 Alpha 3 Report

Filed under
Reviews
SUSE
-s

Well, openSUSE 10.2 Alpha 3 is in our midst and Tuxmachines is here to keep you posted. This release we tested both an upgrade and a fresh install. We found this to be a very interesting release to say the least. It's an alpha to be sure to say the most.

Ok, let's expand on the most. My first interest was in the upgrade procedure - would it complete and provide an equal system to a fresh install? Then one has to decide to update from outside the system through the installer or from within the system with yast2 software manager (system update). Since I had had some questions from users concerning the update feature, I decided to test that method. So from my alpha2 install I defined my new alpha 3 repository and let the system update. The online update didn't show any updates until about 1/2 way through the update through the software manager. The software manager found over 300 packages to update and afterwards, the online update still found 13 more. After all the updates installed with no errors and a reboot, I discovered that didn't work out real well. X wouldn't start at all and yast2 at the commandline seg faulted. There had been some dependency issues and I was asked how to handle them. I used my best judgement, but perhaps this could have contributed to the failure. I decided not to waste much more time here and to test a fresh install.

There were some changes in the installer. I'd forgotten to check the most annoying bugs list before testing, but the graphical installer was working. You might recall my mentioning the new boot screen last report, and it is still that close-up photo of a lizard. The first new element spotted was at the license agreement screen. On the "you know this is beta and we aren't responsible yadda yadda" screen, the new naming convention is apparent. This release was officially tagged as openSUSE 10.1.1 Alpha 3.

There were the now familiar Validation Check Failure errors when the setup system began to start. It also complained about No Checksum Found for every setup file it loaded. Checking "don't show this screen again" checkbox didn't have the desire effect and I had to agree to each and file individually. Thank goodness it was only about a dozen or so of them. I hoped this wouldn't happen on each and every package we were going to install.

The software package selection section has changed this release. Instead of the previous selections categories, we now have "patterns." Similar in nature, but extremely slimmed down, this might make the install a faster process, but it results in a much smaller system. You can still go into the time-consuming Package Groups to choose all your favorites and must-haves such as the kernel-source, but by default we now see a scaled-down setup similar to the following:

  • Base Technologies

    • openSUSE Base System

    • Novell AppArmor
  • Graphical Environments
    • Gnome Desktop Environment

    • KDE Desktop Environment
    • X Window System
  • Primary Functions
    • Graphics (contains only gimp)

    • Gnome Graphics (ex.: eog and f-spot)
    • KDE Graphics (ex.: gimp, kdegraphics-kamera, and gwenview)
    • Print Server (ex.: cups and samba)
    • DHCP and DNS Server
  • Development
    • Basis Development (ex.: gcc, cvs, automake, and ncurses)

    • C/C++ Development (ex.: electricfence, boost, and ltrace)

This yielded a system size less than half of my past SUSE Linux installs. Looking on the good side, it does combat the accusation of being "bloated." This step finished par for the course and we were soon ready for the final configurations.

Hostname, root password, and network configurations are as we remember. Testing the network connection returned a success while the online update setup failed. Next came Users, clean-up, and Release Notes. The Release Notes contained no information.

Backing up to the bootloader configuration for a minute, I had someone ask about auto-detection of other Linux system by the installer. You can about surmize the Linux systems installed on my machine by taking a look at the original content list. I have 22 different various Linux and BSD-clone installs including SLED 10r3 and a couple of older SUSE installs on hdb. The openSUSE bootloader configuration detected four of them. It found Kate OS 3.0b1 which it identified as Debian, Kate OS 3.0 which it identified as Ubuntu, DreamLinux 2.0 which it called Ubuntu, and PCLOS .93a which it called Linux. I don't usually let new systems install a bootloader, and as such makes little difference to me. I just thought this was a bit interesting and worth mentioning. It's possible that the DreamLinux kernel could identify itself as Ubuntu, but I doubt very seriously that the Kate OS developers base anything on Debian or Ubuntu.

Next was the final hardware configuration. This step has either changed some or was having problems. I was used to a long list of hardware to adjust or accept including things such as graphics, tv card, and sound, but this time it only detected and offered my printer and sound. Finishing up the install it started the installed system and X.

Here's where things got real interesting. I have to draw the parallel between some of the "bugs" found in openSUSE as also found in Mandriva 2007 Beta 1. Many if not all of the same X bugs I complained about in my article on Mandriva were also found in openSUSE 10.2 alpha 3. The ugly fonts were present in both, sluggish performance under vesa, and excessive cpu usage wer found in both. openSUSE added limited screen resolution and no nv support. I say no nv support because although I edited my xorg.conf file by hand to make the change to nv, my changes were ignored. I couldn't try to configure X by yast as I was stuck in this "we need to install xorg-x11-server-glx" loop. I suspect this is all related to using Xorg 7.1.1 (7.1.99.2). I did a quick check for Xorg bugs relating to these issues, but all I found were nvidia proprietary driver listings. I might do a more thorough search later. As slow as Xorg moves, I fear continuing problems throughout the 10.2 lifespan for nvidia card users. Hopefully it's just one or two specific chipsets.

    

The next issue cropped up when I wanted to take screenshots of the new wallpaper. This release brings a lovely variation on the blue wispy wallpapers we've seen in (open)SUSE lately. Did I mention the new KDE starting splash? This too is new this release. It's a really nice royal blue background with the a new openSUSE logo.


There was no ksnapshot in the menu. Knowing I had a scaled-down install, my first instinct was to fire-up yast2 and look for missing KDE packages. But whoops, my root password wouldn't work - several times. I found I could sudo /sbin/yast2 and get the ascii version, but not until after sudo passwd and discovering my password rememberance wasn't the problem.

The graphical yast2 would start from the commandline as well. The software manager in Yast itself seems to be functioning pretty good this release. I couldn't find a scanner config under Hardware and discovered I had to install Yast-scanner. Come on, really. That should not be one of the "extras." However, after install of said package, scanner detection and operation was as desired.

        

In attempting to test Gnome, I found no option for it in the login manager. Using console login, it could be started from the command prompt. Upon start I got two errors. One was "There was an error starting the GNOME Settings Daemon" and "Power Manager did not start." But otherwise we find Gnome 2.12.2 with the signature customized SUSE Gnome menu system.

        

On the desktop we discover that the Firefox icon is inoperative. Trying to start Firefox at the commandline we find out that it's seg faulting.

Some RPM version highlights this release include:

  • OpenOffice_org-2.0.3-3

  • MozillaFirefox-1.5.0.6-2
  • xorg-x11-7.1-11
  • kdebase3-3.5.4-3
  • gnome-desktop-2.12.2-26
  • kernel-source-2.6.18_rc4-2
  • gcc-4.1.3-3
  • Full RPMList


Some Changelog highlights can include:

++++ bitstream-vera:

- install into /usr/share/fonts/truetype

++++ xorg-x11-Xvnc:

- created package

++++ coreutils:

- Move sux to %{_bindir}.

++++ dejavu:

- BuildRequires: xorg-x11-devel is necessary to detect Xorg X11R7.

++++ filesystem:

- Add /usr/share/fonts and remove /usr/X11R6/lib/X11/fonts

++++ kernel-default:

- patches.fixes/kbuild-fix-external-module: kbuild fixes for
2.6.18.
- rpm/kernel-source.spec.in: don't remove include/config/*
for building external modules.
- Update kdb patches.

++++ xorg-x11-driver-video:

- updated i810/intel driver to release 1.6.3

++++ Crystalcursors:

- fix for Xorg 7.1 (move to /usr/share/icons)

++++ gimp:

- Changed branding to SuSE Linux 10.2.

++++ gnome2-SuSE:

- Updated to SuSE Linux 10.2 branding.

++++ kdebase3-SuSE:

- artwork update for openSUSE 10.2

++++ hal:

- disables following patches for STABLE/SL10.2Alpha3, they cause a
segmenatation fault in the STABLE tree:
- hal-performance-properties2.diff
- hal-performance-properties_fix_compiler_warnings.diff
- disabled SLE10 specific patch for DBUS

++++ Full Changelog since Alpha 2.

Well, it was time to check out the Most Annoying Bugs list to see if there was any mention of a workaround for the X issue among others. It came as no real surprize to find most of my issues listed. The most annoying bug list contains:

  • YaST does not allow X11 configuration since it asks for non-existant xorg-x11-server-glx Bug #198250. Note: I could run X11 nevertheless and logged into both KDE and GNOME
  • zen-updater always shows patterns to update Bug #198379
  • f-spot does not work Bug #198377
  • gnome-wm does not handle X11R7 Bug #197093
  • Firefox does not start Bug #197928
  • Registration fails with an internal server error Bug #198381
  • applications using python-gtk are broken, e.g. smart-gui Bug #198391
  • kde su does not accept correct password Bug #198408
  • Most kernel module packages are not build against the new 2.6.18rc4 kernel. If you need them, I advise to wait for their update. The Xen packages are not adjusted either.
  • The change of branding (from "SUSE Linux 10.1" to "openSUSE 10.2") is not complete.
  • X Server fails to start with error message "could not open default font 'fixed'", because SaX2 writes wrong font path entries into /etc/X11/xorg.conf. Replace /usr/lib/X11/fonts with /usr/share/fonts/ as workaround. Bug #198653


I didn't check f-spot, and I don't mess with zen. I didn't get the "X server fails due to fixed font" problem with the fresh install, but perhaps this was the issue with the upgrade. Everything else on the list was spot-on.

I think the biggest issue is with this Xorg version. If developers insist upon using it, there are going to be a lot of unhappy users. I've experienced issues to a smaller degree with a couple of other distros as well using later 7.1 versions. I'm all for bleeding edge and don't mind minor breakage here and there, but the X server is one thing that needs to function fairly properly. I'm afraid like with Mandriva, this issue just spoils the whole experience and labels this release as not download-worthy. I'm not sure what other graphic chipset will have a problem, but nvidia is definitely one of them.

10.2 Alpha 2 Report.


More in Tux Machines

GNOME and Fedora

  • RFC: Integrating rsvg-rs into librsvg
    I have started an RFC to integrate rsvg-rs into librsvg. rsvg-rs is the Rust binding to librsvg. Like the gtk-rs bindings, it gets generated from a pre-built GIR file.
  • 1+ year of Fedora and GNOME hardware enablement
    A year and a couple of months ago, Christian Schaller asked me to pivot a little bit from working full time on Fleet Commander to manage a new team we were building to work on client hardware enablement for Fedora and GNOME with an emphasis on upstream. The idea was to fill the gap in the organization where nobody really owned the problem of bringing up new client hardware features vertically across the stack (from shell down to the kernel), or rather, ensure Fedora and GNOME both work great on modern laptops. Part of that deal was to take over the bootloader and start working closer to customers and hardware manufacturing parnters.
  • Fedora Atomic Workstation: Works on the beach
    My trip is getting really close, so I decided to upgrade my system to rawhide. Wait, what ? That is usually what everybody would tell you not to do. Rawhide has this reputation for frequent breakage, and who knows if my apps will work any given day. Not something you want to deal with while traveling.
  • 4 cool new projects to try in COPR for February

Why You Shouldn’t Use Firefox Forks (and Proprietary Opera)

  • Why You Shouldn’t Use Firefox Forks Like Waterfox, Pale Moon, or Basilisk
    Mozilla Firefox is an open source project, so anyone can take its code, modify it, and release a new browser. That’s what Waterfox, Pale Moon, and Basilisk are—alternative browsers based on the Firefox code. But we recommend against using any of them.
  • Opera Says Its Next Opera Release Will Have the Fastest Ad Blocker on the Block
    Opera Software promoted today its upcoming Opera 52 web browser to the beta channel claiming that it has the faster ad blocker on the market compared to previous Opera release and Google Chrome. One of the key highlights of the Opera 52 release will be the improved performance of the built-in ad blocker as Opera claims to have enhanced the string matching algorithm of the ad blocker to make it open web pages that contain ads much faster than before, and, apparently than other web browsers, such as Chrome.

Graphics: Glxinfo, ANV, SPIR-V

  • Glxinfo Gets Updated With OpenGL 4.6 Support, More vRAM Reporting
    The glxinfo utility is handy for Linux users in checking on their OpenGL driver in use by their system and related information. But it's not often that glxinfo itself gets updated, except that changed today with the release of mesa-demos-8.4.0 as the package providing this information utility. Mesa-demos is the collection of glxinfo, eglinfo, glxgears, and utilities related to Mesa. With the Mesa-demos 8.4.0 it is predominantly glxinfo updates.
  • Intel ANV Getting VK_KHR_16bit_storage Support Wrapped Up
    Igalia's Jose Maria Casanova Crespo sent out a set of patches today for fixes that allow for the enabling of the VK_KHR_16bit_storage extension within Intel's ANV Vulkan driver. The patches are here for those interested in 16-bit storage support in Vulkan. This flips on the features for storageBuffer16BitAccess, uniformAndStorageBuffer16BitAccess, storagePushConstant16 and the VK_KHR_16bit_storage extension. This support is present for Intel "Gen 8" Broadwell graphics and newer. Hopefully the work will be landing in Mesa Git soon.
  • SPIR-V Support For Gallium3D's Clover Is Closer To Reality
    It's been a busy past week for open-source GPU compute with Intel opening up their new NEO OpenCL stack, Karol Herbst at Red Hat posting the latest on Nouveau NIR support for SPIR-V compute, and now longtime Nouveau contributor Pierre Moreau has presented his latest for SPIR-V Clover support. Pierre has been spending about the past year adding SPIR-V support to Gallium3D's "Clover" OpenCL state tracker. SPIR-V, of course, is the intermediate representation used now by OpenCL and Vulkan.

Security: Updates, Tinder, FUD and KPTI Meltdown Mitigation

  • Security updates for Friday
  • Tinder vulnerability let hackers [sic] take over accounts with just a phone number

    The attack worked by exploiting two separate vulnerabilities: one in Tinder and another in Facebook’s Account Kit system, which Tinder uses to manage logins. The Account Kit vulnerability exposed users’ access tokens (also called an “aks” token), making them accessible through a simple API request with an associated phone number.

  • PSA: Improperly Secured Linux Servers Targeted with Chaos Backdoor [Ed: Drama queen once again (second time in a week almost) compares compromised GNU/Linux boxes to "back doors"]
    Hackers are using SSH brute-force attacks to take over Linux systems secured with weak passwords and are deploying a backdoor named Chaos. Attacks with this malware have been spotted since June, last year. They have been recently documented and broken down in a GoSecure report.
  • Another Potential Performance Optimization For KPTI Meltdown Mitigation
    Now that the dust is beginning to settle around the Meltdown and Spectre mitigation techniques on the major operating systems, in the weeks and months ahead we are likely to see more performance optimizations come to help offset the performance penalties incurred by mitigations like kernel page table isolation (KPTI) and Retpolines. This week a new patch series was published that may help with KPTI performance.