Tiny Core Linux
Tiny Core Base => TCB News => Release Candidate Testing => Topic started by: roberts on February 07, 2010, 11:28:54 PM
-
The Second Release Candidate of v2.9 (tinycore_2.9rc2.iso), is now posted and ready for testing
http://distro.ibiblio.org/pub/linux/distributions/tinycorelinux/2.x/release_candidates
tinycore_2.9rc2.iso
tinycore_2.9rc2.iso.md5.txt
Change log for Tiny Core v2.9
Release Candidate 1
* Updated appbrowser - added "Provides" to search for "What provides", also speed improvements.
* Updated appsaudit - added "Display All Not Depends On".
* Updated tce-load - added support for recursive dependencies.
* Updated tc-config - persistent home and/or opt, changed ext only support to not Microsoft file system support.
* Updated ab to use tmp work files.
* Improved language support by auto loading locale support extension(s) when both "lang= " and the "-locale.tcz" extension(s) are available in tce/optional.
* Renamed *.core.gz to simply *.gz for ISO9660 compliance.
* Clean up of reported relics and empty directories.
* Updated f2 & f3 boot message screens.
* Fixed bug in tce-update easy mode.
* Dropped the embedded help file.
Notes:
Attention extension makers, please test the new recursive dependency process, but hold in abeyance until v2.9 is final.
Since the embedded help file has been dropped, your system menu, likely in your backup will need to be adjusted. Edit/remove such entry.
Release Candidate 2
* Upgraded busybox to stable 1.15.3
* Upgraded udev to 151.
* New "Set TCE Drive" in control panel to setup TCE directory, optional and mydata.tgz from cloud mode.
* Upgraded tce-load and appsaudit to support copy2fs flag and list in the base tce directory.
* Updated OnDemand to also support re-launching applicatons. Now one can always use the OnDemand menu section.
* Updated support for easy mode remastering /opt/tce now uses onboot.lst
* Updated tc-config to also make, if necessary, optional directory when tce= boot code is used.
* Added cache, .logo.xpm, and stats.txt to default /opt/.xfiletool.lst
* Updated /opt/.xfiletool.lst to persist OnDemand menu item across reboots via normal backup.
Notes: Files likely in your backup that need updating:
The following have been added to the default /opt/.xfiletool.lst:
cache
home/tc/.wmx/Applications
home/tc/.logo.xpm
home/tc/stats.txt
-
No sound (alsa)... I suppose this may have something to do with the recursive dependencies?
Also, Chromium-browser complained about libIDC.tcz (sp?) which it never had a problem with before.
Is this the right place to report these?
-
tce-load in 2.9rc2 should work with standard and recursive deps - alsa with standard deps works for me with 2.9rc1 (I didn't try with 2.9rc2).
Chromium-browser is probably complaining about libIDL, a dep of ORBit2, which is not listed in the dep file - I'll add it
Edit: ..and the missing dbus/dbus-glib deps for the new GConf
-
There should be no recursive dep files in the repository. I have only asked for extension makers to be testing such. I have also asked that any and all such recursive dep files be held, i.e., not submitted to the repository, until 2.9 is final.
-
Hi there!
I just saw that someone else had problems with the sound with tinycore 2.9rc2. The same happened to me this morning (local time, like 9 hours ago or so) when I tried it out. The sound was muted. I tried to use Xmms to listen to a part of an audio book. So i just switched back to tinycore 2.9rc1 (just the core, not making any other changes) and now the sound is up just as usual. Can't explain this but it seemed very odd. Otherwise it really seems that the 2.9 is getting along well.
Have fun making tinycore progress,
meo
-
OSS or ALSA?
Would you check the sound modules loaded (lsmod).
Compare the working one 2.9rc1 and that which 2.9rc2 display?
-
Here's mine...
tc@box:~$ lsmod
Module Size Used by Not tainted
i915 101560 2
drm 84616 3 i915
i2c_algo_bit 2956 1 i915
i2c_core 9980 3 i915,drm,i2c_algo_bit
cpufreq_ondemand 2928 2
acpi_cpufreq 3672 0
freq_table 1208 2 cpufreq_ondemand,acpi_cpufreq
snd_hda_codec_realtek 141844 1
snd_hda_intel 12532 0
snd_hda_codec 29368 2 snd_hda_codec_realtek,snd_hda_intel
snd_pcm 36852 2 snd_hda_intel,snd_hda_codec
snd_timer 10340 1 snd_pcm
snd 25936 5 snd_hda_codec_realtek,snd_hda_intel,snd_hda_code c,snd_pcm,snd_timer
ath9k 192004 0
soundcore 2192 1 snd
mac80211 82100 1 ath9k
snd_page_alloc 4028 2 snd_hda_intel,snd_pcm
intel_agp 14984 1
cfg80211 21604 1 mac80211
agpgart 15724 3 drm,intel_agp
squashfs 11732 66
scsi_wait_scan 260 0
atl1e 19596 0
battery 5976 0
ac 1732 0
eeepc_laptop 5264 0
hwmon 640 1 eeepc_laptop
backlight 1404 1 eeepc_laptop
rfkill 4012 5 ath9k,eeepc_laptop
I don't have a copy of 2.9rc1 so can't show that.
-
Just did a test in Qemu; OSS worked straight, ALSA had one issue in that it didn't tell udev to reload rules after placing its new rules there.
Please retry with the latest alsa.tcz.
-
OK, after updating alsa I don't have sound in 2.8.1 either, so let me fix that first.
-
losetup -f still does not work
losetup only shows first 10 devices.
Everything else I checked looks good.
OSS and ALSA both working.
PXE boot to NFS store working.
-
It looks like the new BusyBox version has the "VI editing mode" enabled. Thanks a lot Robert for picking up this user request!!
Mind you it's pretty limited, e.g. no shell command history search, which is the feature I'm very much used to. This is not a TC problem, but rather something to suggest to the BusyBox folks. I don't want to come across as ungrateful, I just thought I'd share my little observation.
-
Hi again!
Been busy. I think that the boot code I get might be useful. I'm using alsa by the way. These are the two final lines I get while booting:
Starting sound driver: snd-nm256
amixer: Mixer attach default error: No such file or directory
I hope that this can pinpoint the problem with the sound in tinycore 2.9rc2. When I start Xmms the sound is put to zero and even if I move the sound slider there is still no sound. I had alsamixergui installed too so I tried to start it but got an error message of a missing file if I remember correctly. Hope this is of any help.
Have fun advancing this amazing distro,
meo
-
OK, after updating alsa I don't have sound in 2.8.1 either, so let me fix that first.
Got my sound working by manually adding 2 commented-out lines in the start-up script to .xsession, so now all's well with rc2.
-
alsa is working for me in 2.9rc2 using:
alsa-modules-2.6.29.1-tinycore
libasound
libpci
pci-utils
alsa-config
and then: $ sudo /usr/local/etc/init.d/alsasound start
-
Looks like alsa sound is being taken care of at the extension level supporting both the udev update as well as the prior version.
One area tha I have gotten no public feedback, yet had many past requests, is recursion. These requests came as a result of the tedious process to collect and specify all dependencies to the top level application. And then again, if any changes occurred. It is probably more difficult to undo these top level monolithic dep files than if we had started with recursion.
In that regard, I want to point out a script that Helander created that might be helpful in (re)creating the dep files. http://forum.tinycorelinux.net/index.php?topic=4599.msg24115#msg24115
I am hoping to hear some positive feedback from extension makers now that we can support recursion.
-
I am hoping to hear some positive feedback from extension makers now that we can support recursion.
Robert,
what kind of feedback are you expecting?
-
Well, usually I only hear when things don't work!
In this case, I would like to hear assessments such as:
faster or slower showing some time comparisions.
easier or more difficult.
success stories, for example, "I now have LXDE working with recursive dependencies!" ;D
-
Well, the problem is the testing method. There are some issues.
I know which are the direct and indirect dependencies of extension created by myself and can adjust .dep files of my own extensions, but not necesarily others used in the test.
One test is to look for changes booting the system with the changed dep files. This can be done with the limitation see above for LXDE and Xfce as a good example having long and longer dep lists :)
The other test would be to see installation from the repo with appbrowser on a stock system using the new dep files. It requires not only new dep files but an online test repository. Having a test repo would be really useful for in vitro testing.
I'm ready to correct my dep files. However such test (and also the transition) requires work of all extension makers and a longer, few weeks test period.
-
One easy test is trying to install galculator.tcz. In fact this is the first time I noticed that it needed ttf-bitstream-vera.tcz by depending on libglade.tcz
Recursive dependency worked for 2.9rc1 and rc2 as it didn't in 2.8.1.
Moving forward, one thing that puzzled me is that when pressing 'Set TCE' button on the Control Panel, it merely complains that it is alread set, which is true but I was expecting the button to allow me to reset it. I wonder if someone expected the same. I am not talking about a feature request, but merely stating that in my particular case that button served no obvious purpose.
-
"Set TCE Drive" is for new users!
Here is the typically senario...
"I booted Tiny Core, installed some apps and when rebooted they are gone!"
Typical answer: "You need to make nested directories of tce/optional on some drive and you need to get the permissions just right!. You also need to touch a file called mydata.tgz in the tce folder"
The official answer has been: "Just use the boot code tce=hda1. It will setup everything on drive hda1. You only need to do this boot code one time"
So, beacuse this senario plays out so often, I thought I would add an option to setup the tce directories and default empty backup file when booted in cloud mode.
It is not for the creation of multiples as that can become quite confusing.
Re: Testing recursion can be done locally with tce-load -i extname.tcz
-
Re: Testing recursion can be done locally with tce-load -i extname.tcz
How -i (and -w) switches work? Tried to find a description with no success :(
-
# Example usage:
# tce-load -i /mnt/hda1/tce/optional/nano.tcz Loads local extension
# tce-load -w -i nano.tcz Download into tce/optional directory, updates OnBoot and installs
# tce-load -w nano.tcz Download only into tce/optional directory
-
As for the conversion of the existing dep files, I am testing a script that will convert automatically.
All repo conversions are scripted, it is always too much to do by hand even with many hands involved. I just need to work in wgetting existing dep files for the script to work against. In making extensions, you can list as many redundant entries that you want, and running the script will purge the dep file of deps that are called upon by other deps in the list. Just need to test a little more. There are others working on a similar thing I believe, so soon testing will be much more simple.
-
rebuildfstab[2685]: segfault at 8 ip 0806f590 sp bffd8510 error 4 in busybox[8048000+7c000]
-
"Set TCE Drive" is for new users!
[...]
It is not for the creation of multiples as that can become quite confusing.
Certainly. I was not taking into account that irremisible consequence. However, I see an opportunity here for education, changing the message from a simple "Already done" or "Don't do that" to something along the line: "TCE location is already set. It is strongly discouraged to allow multiple disparate repositories to reside in the same machine".
-
Well, usually I only hear when things don't work!
In this case, I would like to hear assessments such as:
faster or slower showing some time comparisions.
easier or more difficult.
success stories, for example, "I now have LXDE working with recursive dependencies!" ;D
I set up a private repo. Tested, it works as expected. Next I replaced dep files with files created by Jason's script. Tried Appbrowser and tce-load with the same result. Do not see any recursive resolution, only those extensions loadced which are listed in the .dep file of extension installed, but nothing next level.
For example hal.tcz.dep is
dbus-glib.tcz
pci-utils.tcz
usb-utils.tcz
udevadm.tcz
Installing hal.tcz all above are loaded, but for example libusb.tcz isn't which is in the usb-utils.tcz.dep
P.S.: Maybe I misunderstand recursive dependency resolution :(
-
It seems to be working here with tce-load from 2.9rc2 when using tce-load from the command line and the extensions are in /optional. Here is the adjusted rhythmbox deplist:
totem-pl-parser.tcz
gnome-media.tcz
libsoup-gnome.tcz
gst-plugins-good.tcz
gst-plugins-bad.tcz
gst-plugins-ugly.tcz
gst-python.tcz
gmime.tcz
pygtk.tcz
pyorbit.tcz
gnome-python.tcz
libnotify.tcz
avahi.tcz
libgpod.tcz
gnome-icon-theme.tcz
shared-mime-info.tcz
And here is what ends up being installed, according to /usr/local/tce.installed:
GConf
ORBit2
SDL
Xlibs_support
atk
avahi
bzip2-lib
cairo
dbus
dbus-glib
eggdbus
expat2
faad
flac
fontconfig
gcc_libs
glib2
gmime
gnome-icon-theme
gnome-media
gnome-python
gnome-vfs
gnutls
graphics-libs-1
gst-plugins-bad
gst-plugins-base
gst-plugins-good
gst-plugins-ugly
gst-python
gstreamer
gtk2
libIDL
liba52
libasound
libbonobo
libcdaudio
libcdio
libcroco
libdaemon
libdca
libdvdnav
libdvdread
libffi
libgcrypt
libglade
libgnome
libgnome-keyring
libgpg-error
libgpod
libgsf
libid3tag
libmad
libmpeg2
libnotify
liboil
libproxy
librsvg
libsoup
libsoup-gnome
libtheora
libxml2
nss-mdns
ogg-vorbis
openssl-0.9.8h
pango
pixman
popt
pycairo
pygobject
pygtk
pyorbit
python
sqlite3
taglib
tcl_tk
totem-pl-parser
ttf-bitstream-vera
-
Jason, can you share how are you using tce-load ? What is about Appbrowser?
-
using "tce-load -i rhythmbox.tcz" when the extensions were downloaded and waiting in ondemand mode in an empty /optional.
-
Both Ondemand with Appbrowser or tce-load -w hal.tcz install only what is listed in hal.tcz.dep
-
Straight from the repo:
http://distro.ibiblio.org/pub/linux/distributions/tinycorelinux/2.x/tcz/galculator.tcz.dep
atk.tcz
cairo.tcz
pango.tcz
glib2.tcz
fontconfig.tcz
expat2.tcz
libxml2.tcz
graphics-libs-1.tcz
pixman.tcz
Xlibs_support.tcz
gtk2.tcz
libglade.tcz
http://distro.ibiblio.org/pub/linux/distributions/tinycorelinux/2.x/tcz/libglade.tcz.dep
expat2.tcz
glib2.tcz
graphics-libs-1.tcz
libxml2.tcz
fontconfig.tcz
ttf-bitstream-vera.tcz
gtk2.tcz
atk.tcz
cairo.tcz
pango.tcz
pixman.tcz
Xlibs_support.tcz
Notice ttf-bitstream-vera.tcz within libglade.tcz.dep. TC 2.8.1 didn't load it as it hadn't recursive dependency resolution, while 2.9rcX does.
-
ttf-bitstream-vera was at one point used as the default font for gtk2 apps. It is no longer a requirement so it is not listed in many newer dep files. Neither of those dep files in the repo were intended to be recursive. And since extension makers have some liberty to include dependencies that may not be absolutely necessary, there are likely other ones that contain non-essential entries.
-
Maybe, but it just gives us an opportunity to test the recursive dependency feature.
-
Oh, ok. I misunderstood.
-
The firmware isn't loading. I have the firmware-2.6.29.1-tinycore.tcz extension mounted on boot but am getting this message:
e100: eth0: e100_request_firmware: Failed to load firmware "e100/d102e_ucode.bin": -2
This firmware is found in the firmware extension at /lib/firmware/e100/d102e_ucode.bin
I get the same results with the cpio and upxed kernels. depmod -ae does not fix this. Neither does a rmmod e100 then modprobe e100.
What else can I try to get this to work?
-
The firmware bug has already been posted: http://forum.tinycorelinux.net/index.php?topic=4966.0
It is fixed in rc3. I will be it posting soon.
-
# Example usage:
# tce-load -i /mnt/hda1/tce/optional/nano.tcz Loads local extension
# tce-load -w -i nano.tcz Download into tce/optional directory, updates OnBoot and installs
# tce-load -w nano.tcz Download only into tce/optional directory
Could this post be updated with how tce-load works relative to the copy2fs list and flag?
Also, is there a way to perform a "tce-load -i .../mylocation" such that the extension is copied to the file system?
-
Add -c for a one time copy into filesystem to be used in combination with -i.
-
Add -c for a one time copy into filesystem to be used in combination with -i.
Thanks for the tip. I didn't know there was a -c option.