Tiny Core Linux
Tiny Core Base => TCB News => Alpha Releases => Topic started by: roberts on May 30, 2010, 12:39:24 AM
-
Tiny Core v3.0 Alpha5 is now posted at http://distro.ibiblio.org/pub/linux/distributions/tinycorelinux/3.x/release_candidates
Change log for Alpha5
* Updated AppsAudit: added OnDemand Maintenance (add/remove) OnDemand items.
* Updated ondemand to support new feature in AppsAudit.
* Updated AppsAudit: dropped full path from "Check for Updates" display.
* Updated flwm* for potential issue when used with peristent home
* Updated tce-setdrive to exclude cdrom and dvd.
* Updated mktclocal to exclude cdrom and dvd.
* Updated usbinstall moved bzImage and tinycore.gz/microcore.gz to /boot
* Updated wbar_upd_icon.sh for missing c: in freedesktop Terminal & Icon case.
* Fixed various typos and other small tweaks and adjustments
* Fixed freedesktopcheck in startx
* Updated squashfs module with latest patch.
* Removed unused .webdata* from /etc/skel/
* Updated X libs in base to Xorg 7.5:
libICE-1.0.6
libSM-1.1.1
libX11-1.3.3
libXau-1.0.5
libXaw-1.0.7
libXcursor-1.1.10
libXdmcp-1.0.3
libXext-1.1.1
libXfont-1.4.1
libXft-2.1.14
libXi-1.3
libXinerama-1.1
libXmu-1.0.5
libXpm-3.5.8
libXrender-0.9.5
libXt-1.0.8
libXtst-1.1.0
* The user-visible part of the update is mainly the removal of the harmless libXext error message
-
Hi Tiny Core Team!
The new cut works without problems for me. Did some initial testing and the applications I use most work flawlessly. Thanks for your hard work!
Have fun working with tinycore 3.0 towards it's final state,
meo
-
This was probably there before, but when I use appsaudit to update bluez, the updated dep file contains bluetooth-KERNEL.tcz rather than bluetooth-2.6.33.3-tinycore{,64}.tcz
-
Tonight for the first time I saw the release news *without* a happy note from Meo, and thought, "This is my chance!" Unfortunately, after installing as fast as I could and quickly testing a YouTube video with getFlash10 and namaroka, I lost the race. Congratulations to Meo and ... more importantly ;) ... thanks to the team.
I like the classic FLWM desktop. The menu system based in ~/.wmx has a symbolic link named "Applications" pointing to /usr/local/tce.flwm. In the 2.X series of TC releases, this target directory was populated automatically when things like namaroka or the OSS mixer got installed. Now it seems this directory just stays empty. Can the previous functionality be restored?
-
~/.wbar apparently does not get used... on starting, wbar is located bottom centre, whereas in my ~/.wbar file it is spec'ed as being vertical and on the right.
-
tc@box:~$ wbar.sh
tc@box:~$ wbar: unrecognized option '-bpress -vbar -pos right -zoomf 1.78 -nofont'
edit: Removing the double quotes around $OPTION in wbar.sh fixes the problem.
-
Couldn' test alpha 4. With alpha 5 I have the same boot errors I had with alpha2 and alpha 3 but not with alpha1 (see http://forum.tinycorelinux.net/index.php?topic=6021.msg32363#msg3263 in alpha 3 thread for details)
-
jano: First up I don't pretend to really know much at all about the following, so I can only hope that some of the other guys are stepping in when I'm "talking nonsense" here:
- I was surprised that you reported your problems to have started with alpha2 since the kernel got build again with alpha3. The only change I could find between alpha1 (i.e. "good") and alpha2 (i.e. "bad") that I would associate with the "udev area" was the addition of KERNEL=="ramzswap*", GROUP="root", MODE="0600" in '/etc/udev/rules.d/55-tc.rules'. Maybe you could re-master (http://wiki.tinycorelinux.com/tiki-index.php?page=Remastering) the initrd of TC 3.0alpha5 with a version that removes this entry again and see whether that makes a difference (BTW I would not bother with the 'advdef' step since "squeezing out" the last bits from the file should not be critical for your testing).
- Since I assume that some sort of race condition or some other failure happens to "udev" I wonder whether you should remaster the initrd with a view of changing '/etc/init.d/tc-config' a bit: I see that /sbin/udevd --daemon 2>/dev/null >/dev/null inhibits all output, and that a '--debug' option is available. I therefore wonder if a change to /sbin/udevd --daemon --debug would show more information in the 'syslog'.
- Likewise there are several uses of /sbin/udevadm in the same file where all output is inhibited. I suppose running these commands in the background is critical and running them in the foreground might not be an option.
Well, that are just my 2c ...
-
This was probably there before, but when I use appsaudit to update bluez, the updated dep file contains bluetooth-KERNEL.tcz rather than bluetooth-2.6.33.3-tinycore{,64}.tcz
Found it.
Fixed it.
Thanks!
-
As I tried to understand some of the changes between TC 3.0alpha4 and 3.0alpha5 I made the following observations:
- In '/usr/lib' a fair amount has changed (as mentioned in the release notes), but I've stumbled upon a few files that I'm not sure to which degree they are still required:
- libSM.so.6.0.0 has no symbolic link pointing to it (those are now pointing to libSM.so.6.0.1)
- libX11.so.6.2.0 has no symbolic link pointing to it (those are now pointing to libX11.so.6.3.0)
- libXi.so.6.0.0 has no symbolic link pointing to it (those are now pointing to libXi.so.6.1.0)
- libXrandr.so.2.1.0 has no symbolic link pointing to it (those are now pointing to libXrandr.so.2.2.0)
- There appear to be only three libraries left which have not been updated: libXfontcache.so.1.0.0, libfontenc.so.1.0.0, and libjpeg.so.62.0.0.
- The latter one I could explain myself. Maybe the latest stable release (http://www.ijg.org/files/jpegsrc.v8b.tar.gz) is seen as still too "fresh" given the fact that it's merely two weeks old.
- But the two former ones I thought might be part of Xorg and would therefore be updated.
I'd appreciate if someone wiser might be kind enough to explain this to me.
-
I think the leftover files are already taken care of ;)
Seems like libXfontcache can be removed (deprecated) and that libfontenc might already be the latest.
Not using the 'highest' version was done to keep a compatible interface.
-
Seems like libXfontcache can be removed (deprecated) and that libfontenc might already be the latest.
I confirm this, Xfontcache is not a part of Xorg-7.5, and libfontenc has not received an update since the 7.4 days.
-
cpanel/Update Apps displays "Reboot to use new updated extensions." on fresh system instead of "Your system is up to date".
-
Are we lacking a linux-kernel-headers extension for the alpha versions of TC3?
-
Technically, yes, for the convenience of building modules without the full kernel source.
Until one gets made, the full kernel source can be used instead.
-
I held off posting one until things settled down. I'll see about it sometime.
-
cpanel/Update Apps displays "Reboot to use new updated extensions." on fresh system instead of "Your system is up to date".
Updated to acknowledge when system is up-to-date.
-
Technically, yes, for the convenience of building modules without the full kernel source.
Until one gets made, the full kernel source can be used instead.
Ah, ok. Makes sense.
-
I hope this is an appropriate place to post this bug:
Load galculator. It runs. Now load and run gnumeric. Now, galculator segfaults.
-
Just tried the same test booting off usb stick and it didn't happen. On the usb stick I run xvesa while on the harddisk I use Xorg.
-
i am not sure whether it still applies to 3.0 but there is one issue that could use a fix (i've seen it on 2.11 at least).
when you have a external device/filesystem or anything mounted somewhere in /home/tc, and decide to shutdown TC, the backup script will try to pack home dir along with mounts into mydata.tgz.
-
/home/tc/mnt is in the default xfiletool.lst, so if you mount under that, things will work.
-
I can confirm the finding from jur that galculator "dies" with a segfault if gnumeric is installed as well. I did my tests on a "plain" TC 3.0a5 system and considering my findings I can only assume that his second report resulted due to a different set of extensions (but most likely unrelated to the X server).
In principle my test consisted of the following steps:- I installed 'galculator.tcz' and confirmed that 'galculator' is working. I also took note of the ldconfig -p and ldd $( which galculator ) output.
- I installed 'gnumeric.tcz' and confirmed that 'galculator' is crashing with a segfault. Comparing the respective output showed that several libraries (e.g. 'libxml2.so.2', and 'libglade-2.so.0') had been replaced by different versions installed to '/usr/lib' which were included in 'gnumeric.tcz'.
- I then removed one of the "offenders" (with sudo rm -f /usr/lib/libxml2.so.2* ; sudo ldconfig) and this allowed for 'galculator' to work again.
There are quite a few libraries contained in the '/usr/lib' directory of 'gnumeric.tcz' that are available via distinct library extensions. I spotted 'pcre.tcz', 'ORBit2.tcz', and 'libgpg-error.tcz' as other candidates (but there are probably more) by just eyeballing 'gnumeric.tcz.list'. And there is a fairly good chance that some other incompatibilities will "break" also other extensions. I'd therefore suggest to remove 'gnumeric.tcz' from the 3.x repository until it has been rebuild taking into account the possibly much larger set of available libraries via extensions in the current repository.
Due to this finding I then had a quick look across the .list files of the 3.x repository and saw a few other cases where libraries had been put into '/usr/lib': - ayttm.tcz contains '/usr/lib/libltdl.so*' which could possibly be removed and replaced via the 'libltdl.tcz' extension.
- inkscapelite.tcz contains '/usr/lib/libpopt.so*' which could possibly be removed and replaced via the 'popt.tcz' extension.
- stardict.tcz contains '/usr/lib/libsigc-2.0.so.0*' which could possibly be removed and replaced via the 'libsigc++.tcz' extension.
- udev-lib.tcz contains files for '/usr/lib' (and '/lib') and I wonder if '/usr/local/lib' would not be the better "place".
-
maro: Thanks for the confirmation. Your findings for libxml2 also explains why I had "random" cases where right-clicking on rox filer sometimes exited as reported before.