openSUSE 10.3 alpha 3 Report
openSUSE 10.3 alpha 3 was released yesterday and it seems to be getting quite a lot of press from several sites this round. It's being announced with enthusiam as if it was a release candidate. Some exciting developments are occurring in the openSUSE camp and it's good to see some positive press for them. As usual I test drove alpha 3 and am ready to report my findings.
This release didn't bring any new visuals, other than fonts. It might be my imagination, but the fonts are a bit improved. It could be that all the press on the anti-aliasing has had some positive effects, or perhaps it was my using a different machine. Most likely it was my overactive wishful imagination. In any case, I'm having some issues with the dvd drive in my desktop machine, so this test was run on my HP laptop. I commented in the past, a couple of times, that the fonts were one of the major issues in my using openSUSE only as a backup or second choice to PCLOS on that machine, but testing 10.3 alpha 3 yesterday has me thinking this may no longer be the case. There are actually only a few of fonts installed by default that look exceptionally nice, but one only needs a couple really. Novell's position on Cleartype and the bytecode interpretter are clear, but perhaps they are working on alternative solutions.
This go around I didn't encounter any issues with libata drivers, but I wouldn't on that laptop anyway. During testing for my last report I erroneously thought openSUSE was following suite with Fedora in their handling of ata drives, but what turned out to be a bug was supposedly corrected for this release. And just my luck, I wasn't able to confirm it was all in the past. Hopefully I'll have a new dvd drive for my desktop soon (donations gratefully accepted <hint, hint>).
Last test I had problems with gnome as well. Unfortunately, they weren't corrected for the iso creation this release either. The panel crashes upon logging into Gnome and thus no menu is available. After being bitten on the ego last release, I did check with the Most Annoying Bugs list prior to testing (as opposed to after), so I wasn't surprised and also applied the fix which is to reinstall the gnome-main-menu rpm. There seems to be a problem with the Tango-icons rpm as it fails verification, so I had to use --nodeps as well as --force to reinstall this package. It was also recommended to use the control-center rpm available in the Extra-RPMS directory as well. So, I installed it at the same time. As a result, I had what appeared to be a fully functional gnome this time. The version numbers are still being reported as 2.18 although 2.18.1 was recently released. If openSUSE stays true to form, we will see an upgrade probably once they have all the kinks worked out in their current integration.
Speaking of gnome, this is perhaps the first time I've tested gnome on the laptop. In at least openSUSE's incarnation, we find power management and wireless connection (network manager) applets in the panel similar to found in KDE. They look very similar and have almost equal functionality. Almost. They don't have quite as many options available at click, but the basics are there. Additionally, just as kwallet offers to remember passwords and/or phrases, the keyring manager in gnome does the same. I don't usually opt for these password keepers, but decided to test it this round. It does appear to work well, although it seems just as annoying to input my keyring password each use as it would my wpa passphrase. I don't see the advantage, unless you have several to remember I suppose.
Some of the improvements announced this release include:
- On x86-64: Firefox is now a 64-bit package and uses nspluginwrapper to handle 32-bit i386 plugins if needed.
- AppArmor uses now a new parser. The kernel patches have been reworked completely.
- GNOME 2.18 mostly integrated.
- Update to Kernel 2.6.21 RC5
- New opensuse-updater running natively under GNOME
- Further fixes for using libata by default for IDE devices
- New yast2-ftp-server module.
- The package manager handles more than one CD/DVD drive.
The most annoying bugs list contains:
- Java applications are not working due to implementation errors in SUN Java
- The installation of vim-normal might fail
- Branding and translations are still at the 10.2 level
- The Radeon driver is broken and crashes the X server
- Installation with several CD-ROMs might crash at some point
- gnome-main-menu might crash in some installations
One point of note is the installation of vim-normal might fail. It did indeed here. In fact, I had a 1/2 dozen packages fail either due to what was reported as broken packages or failing verification. None seemed to be really important such as susehelp or tango-icons.
Some changelog highlights include:
- opensuse-updater-gnome:
- fixed some memory management problems
- v 0.2.1
- changing applet icon according to current operation/status
- various gnome components & applications updated to 2.18.0
- nautilus:
- Update to version 2.18.0.1
- nano:
- update to version GNU nano 2.0.4
- helix-banshee:
- Update to version 0.12.1
- Added more translations from SLED SP1
- control-center2:
- Added "Use default web browser" option
- kdebase3:
- simplify resize and rotate settings dialog
- update 3_5_BRANCH.diff
- update kickoff.diff from SVN
- k3b:
- fix detection of "sox"
- kernel:
- update to 2.6.21-rc5-git13
- libata fixes
- NetworkManager-kde:
- add patch knetworkmanager-wpaeap_storage-hschaa-01.patch
- gnome-session:
- Session start script (gnome) cleanup
- powertweak:
- Fix cpuid check
- xorg-x11-driver-video:
- updated "nvrandr12" driver to current git
- intel driver release 1.9.94
- xorg-x11-driver-input:
- updated joystick driver to release 1.2.0
- OpenOffice_org:
- updated ooo-build to 2.1.10
- cups:
- Upgrade to cups-1.2.10
- Full Changelog Here
Some RPM version highlighs are:
- kernel-default-2.6.21_rc5_git13-2
- xorg-x11-7.2-62
- gcc-4.1.3-40
- MozillaFirefox-2.0.0.2-10
- OpenOffice_org-2.1-18
- kdebase3-3.5.6-26
- gnome-desktop-2.18.0-6
- qt3-3.3.8-12
- gtk-1.2.10-948
- gimp-2.2.13-53
- freetype2-2.3.1-7
- Full RPM List
So all in all, they are moving right along. There wasn't any new eye candy or major changes on the surface, but there was a lot of bug fixes and feature updates under the hood. One thing I noticed was that the new logout dialog boxes that we discovered in alpha 1 are now gone again. These options have reverted back to the originals.
It's always amazing to me, even after all this time testing their alphas and betas, that openSUSE releases are so functional and stable even during developmental cycles. It's particularly amazing when you look at all the customization, the sheer volume of packages, and size of system. I hope to start seeing some new eye candy in the upcoming releases, although we are usually teased through the beta portion and early release candidates while final window dressings are saved for the last release candidate or final.
The roadmap is still not fully planned out, but final is tentatively scheduled for the end of September.
-
- Login or register to post comments
Printer-friendly version- 40598 reads
PDF version
More in Tux Machines
- Highlights
- Front Page
- Latest Headlines
- Archive
- Recent comments
- All-Time Popular Stories
- Hot Topics
- New Members
today's howtos
| Red Hat Hires a Blind Software Engineer to Improve Accessibility on Linux Desktop
Accessibility on a Linux desktop is not one of the strongest points to highlight. However, GNOME, one of the best desktop environments, has managed to do better comparatively (I think).
In a blog post by Christian Fredrik Schaller (Director for Desktop/Graphics, Red Hat), he mentions that they are making serious efforts to improve accessibility.
Starting with Red Hat hiring Lukas Tyrychtr, who is a blind software engineer to lead the effort in improving Red Hat Enterprise Linux, and Fedora Workstation in terms of accessibility.
|
Today in Techrights
| Android Leftovers |
.svg_.png)
Content (where original) is available under CC-BY-SA, copyrighted by original author/s.
ATI(fglrx) + 2.6.21-rc5 + 64Bit
There is another change in the latest kernel code which means that if you tried to compile and install fglrx from the latest ati driver (8.35.5) then you won't be able to suceed.
I know from experience that this is the case at least in 64bit (not tried 32bit). There is however a fix out there which allows for this to work.
You could extract the driver from the install package and use the patch from: http://www.phoronix.net/forums/showpost.php?p=5204&postcount=38
This patch will allow you to build a working module for your system and although it was released before the 8.35.5 driver this was not included?? shame, but oh well it's available at least
I tried building and installing the rpm, then after it bailed out used the code not removed from: /usr/src/kernel-modules/fglrx
This works for me on openSuSE 10.2 with a more upto date kernel of the day. (2.6.21-rc4-git3 if i remember correctly)
I haven't tried this on the latest openSuSE alpha but as this new alpha is based on a kernel that won't cause my laptop to melt (laptop hp:nx6125 and kernel 2.6.21-rc* )
I will happily migrate this alpha soon som nice cutting edge versions of some progs. I can't wait for openSuSE 10.3 full and lets hope that it has early inclusion of KDE4 stuffs
Just in case the patch shouldn't be available it is below (hope it displays correctly):
--- firegl_public.c-orig 2007-02-21 19:56:45.000000000 -0500
+++ firegl_public.c 2007-02-25 12:31:59.000000000 -0500
@@ -119,6 +119,7 @@
#include
#include
+#include
#include
//#include
@@ -204,6 +205,70 @@
int errno;
#endif // __ia64__
+#if defined(__i386__)
+#define __syscall_return(type, res) \
+do { \
+ if ((unsigned long)(res) >= (unsigned long)(-(128 + 1))) { \
+ errno = -(res); \
+ res = -1; \
+ } \
+ return (type) (res); \
+} while (0)
+#define _syscall2(type,name,type1,arg1,type2,arg2) \
+type name(type1 arg1,type2 arg2) \
+{ \
+long __res; \
+__asm__ volatile ("push %%ebx ; movl %2,%%ebx ; int $0x80 ; pop %%ebx" \
+ : "=a" (__res) \
+ : "0" (__NR_##name),"ri" ((long)(arg1)),"c" ((long)(arg2)) \
+ : "memory"); \
+__syscall_return(type,__res); \
+}
+
+#define _syscall3(type,name,type1,arg1,type2,arg2,type3,arg3) \
+type name(type1 arg1,type2 arg2,type3 arg3) \
+{ \
+long __res; \
+__asm__ volatile ("push %%ebx ; movl %2,%%ebx ; int $0x80 ; pop %%ebx" \
+ : "=a" (__res) \
+ : "0" (__NR_##name),"ri" ((long)(arg1)),"c" ((long)(arg2)), \
+ "d" ((long)(arg3)) : "memory"); \
+__syscall_return(type,__res); \
+}
+#elif defined(__x86_64__)
+#define __syscall_clobber "r11","rcx","memory"
+#define __syscall "syscall"
+
+#define __syscall_return(type, res) \
+do { \
+ if ((unsigned long)(res) >= (unsigned long)(-127)) { \
+ errno = -(res); \
+ res = -1; \
+ } \
+ return (type) (res); \
+} while (0)
+#define _syscall2(type,name,type1,arg1,type2,arg2) \
+type name(type1 arg1,type2 arg2) \
+{ \
+long __res; \
+__asm__ volatile (__syscall \
+ : "=a" (__res) \
+ : "0" (__NR_##name),"D" ((long)(arg1)),"S" ((long)(arg2)) : __syscall_clobber ); \
+__syscall_return(type,__res); \
+}
+
+#define _syscall3(type,name,type1,arg1,type2,arg2,type3,arg3) \
+type name(type1 arg1,type2 arg2,type3 arg3) \
+{ \
+long __res; \
+__asm__ volatile (__syscall \
+ : "=a" (__res) \
+ : "0" (__NR_##name),"D" ((long)(arg1)),"S" ((long)(arg2)), \
+ "d" ((long)(arg3)) : __syscall_clobber); \
+__syscall_return(type,__res); \
+}
+#endif
+
// int mlock(const void *addr, size_t len);
_syscall2(int, mlock, const void *, addr, size_t, len )
// int munlock(const void *addr, size_t len);
@@ -4341,24 +4406,6 @@
#endif
}
-/** \brief Freeze the thread if kernel requested so because of going to suspend
- *
- * \return Nonzero if freeze has been performed, zero otherwise
- *
- */
-unsigned int kas_try_to_freeze(void)
-{
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,10)
- return 0;
-#else
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,12)
- return try_to_freeze(PF_FREEZE);
-#else
- return try_to_freeze();
-#endif
-#endif
-}
-
/** \brief Storage for execution level(s) */
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0)
/* SMP support for 2.6.0 and higher */
@@ -4985,28 +5032,13 @@
unsigned int ret = 0;
kasSlabCache_t* slabcache_obj = (kasSlabCache_t*)hSlabCache;
- DBG_ENTER("0x%08X", hSlabCache);
-
- if (!(slabcache_obj->cache))
- {
- DBG_ERROR("slab object '%s' is not initialized");
- DBG_LEAVE("0");
- return 0;
- }
-
- DBG_TRACE("destroying slab object '%s'", slabcache_obj->name);
-
- if (kmem_cache_destroy(slabcache_obj->cache) == 0)
+ if (slabcache_obj->cache)
{
- ret = 1;
+ kmem_cache_destroy(slabcache_obj->cache);
slabcache_obj->cache = NULL;
- }
- else
- {
- DBG_ERROR("destroying failed");
+ ret = 1;
}
- DBG_LEAVE("%d", ret);
return ret;
}
@@ -5190,6 +5222,24 @@
return 1;
}
+/** \brief Freeze the thread if kernel requested so because of going to suspend
+ *
+ * \return Nonzero if freeze has been performed, zero otherwise
+ *
+ */
+unsigned int kas_try_to_freeze(void)
+{
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,10)
+ return 0;
+#else
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,12)
+ return try_to_freeze(PF_FREEZE);
+#else
+ return try_to_freeze();
+#endif
+#endif
+}
+
/** \brief Wait for the event
*
* If event is already signalled, return right away.
Well not quite SuSE
I want to stress that this is not only a openSuSE problem and *I think* it is another change in mainstream the kernel, just as 2.6.12 upset a few drivers.
I say this as I have read reports of Fedora Core 7 Test{2,3} having trouble with the same config, 64bit, 2.6.21-rc* and fglrx(8.35.5).
I personally think NO DISTRO should hack the kernel. I do not object to building say AppControl as a module, this seems ok to me but why does the kernel have to be altered from Vanilla? Most changes are simply just config (or should be).
If they want to backport fixes or experimental then thats their opinion, I either disagree or back em.
I always remember the Mandrake 9.1 physically damaging some optical drives. This is an example of what imho happens when ditro's go it alone on performance issues.
What has seriously annoyed me was a SuSE developer fixed an issue with the acpi error with my laptop, this was back with 2.6.18 and it has taken for 2.6.21 with this fix becoming vanilla for SuSE to build it into their kernel.
I have been wanting to go Gentoo, but as of now if I do I would either get a melted laptop or a 32bit system on my 64bit laptop. I could disable acpi (like most sane people) but I lost my Bios password so I would have to dismantle a laptop & remove battery or pay for ti to be done for me. I wouldn't mind but I have to remove everything but my processor to get to my battery.
And to be even more honest as soon as CLFS is more stable I will be building my own as close to vanilla/minimalist system as possible.