Tiny Core Linux
Tiny Core Base => TCB News => Release Candidate Testing => Topic started by: roberts on October 02, 2009, 03:49:25 PM
-
The Third Release Candidate of v2.4 (tinycore_2.4rc3.iso), is now posted and ready for testing.
http://distro.ibiblio.org/pub/linux/distributions/tinycorelinux/2.x/release_candidates/
Change log for v2.4 Release Candidate 3
* Updated flwm_topside with color titlebar support and cursor patch.
* Added GNU cpio - dropped busybox cpio applet.
* Updated tce-load to use boot flag and cpio for copy to filesystem (ram).
* Updated tce-setup, optimized code and calls tce-load via boot flag.
* Updated tc-config with more consistent and streamlined messages at boot.
* Updated usbinstall to operate better when launched from system menu.
-
Good job Robert. It's good to have some quick fixes.
I am testing 2.4rc3 on Qemu now. Looks like the title bar color fix works just as it should.
I did notice that the extensions loaded during boot are no longer listed. I was just wondering if this is intentional?
edit: I have Opera, emelfm2 and ace of penguins loaded.
-
Change log for v2.4 Release Candidate 3
* Updated tce-load to use boot flag and cpio for copy to filesystem (ram).
Robert could you provide a little more background on this change. I'm not sure I understand how this works.
-
In v2.3 & v2.4rc1 I was using rsync to copy to filesystem, i.e, 'install' a tcz.
This was causing upstream directories to be overwritten.
In v2.4rc2 I switched to using a tar pipe with xlist, but it was noticeable slower.
With the help of Jason, we tried several options, and decided on cpio.
The GNU version of such has a passthru mode and we can control upstream permissions better.
So rsync, was replaced by tar pipe, was replaced by cpio.
I no longer display extensions as they are being loaded, with upwards of 100 or so, it takes too much time to write to screen. Besides, I do not display each file as it is being restored, nor do I display each boot codes as it is being checked. So we now have a more consistent and cleaner boot screen.
-
Thanks for the additional information.
I also noticed that a tce directory is created in /opt even though I have a tce directory in sda. Is this as it should be?
-
That has been there for many releases. It provides a simple way to remaster.
-
Hi Robert!
The new cut seems to work pretty fine at first glance. There is one problem though that has been the same throughout the rc cycle. I boot with base, norestore, xsetup, mount beaver and emelfm and press Refresh Menu, then the screen goes black. But if I reboot with the tce directory specified and also the same with restore, everything works fine. Thanks to the TC team for continually refining the distribution.
Have fun,
meo
-
meo, I could not reproduce. Did the same steps. You are using appbrowser to load local? By using appbrowser load local the menu and icons were refreshed. No action needed. Perhaps you are using a different window manager? If you are using jwm you should be using the one in the release candidate area. Please give more information.
-
Hi again Robert!
Here comes the whole shabang. These are all the boot codes I use: quiet vga=791 noicons tz=Europe/Stockholm tcvd=hde1/harddisk base norestore xsetup. I ran it all from scratch using the appbrowser mount function. So the only extensions I installed (mounted) were beaver and emelfm. Then I pressed the Refresh Menu to get access to the extensions but the screen went black. Maybe I installed (mount option on the app browser) links too I don't remember. The strange thing is that I looked through the harddrive and the compact flash card that I have in a PCMCIA adapter but I could not find the extensions. I even manually mounted the virtual hd that I have on the flash card to see if they were there but I couldn't find them. I have a tce directory in the root of hda6 which I have formatted in ext3 and it also contains another linux distro. So I'm kind of puzzled but I hope this information can be of help.
Have fun developing tinycore further,
meo
-
Using the 'base' boot option will set the tce directory to /tmp/tce and so the mounted extensions were not saved. If you wanted them saved on the virtual drive then you would need both tcvd=hde1/harddisk tce=tvcd The first to setup the virtual drive, the second to set tce to use the virtual drive.
See: http://www.tinycorelinux.com/faq.html#qemu
-
I no longer display extensions as they are being loaded, with upwards of 100 or so, it takes too much time to write to screen. Besides, I do not display each file as it is being restored, nor do I display each boot codes as it is being checked. So we now have a more consistent and cleaner boot screen.
What is about a number of loaded extensions instead of rotating -|=\ or a simple progress bar?
-
Because rotdash is and has been the standard indicator that processing is occuring as shown above with udev and below with restore. Any other interaction will only slow down the boot process.
-
On a clean system the first loop device is /dev/loop0. Installing extensions in RAM is consuming a loop device for each extensions. For example after installing Midnight Commander and its 3 dependencies first loop device is /dev/loop4 according to 'losetup -f'
It means, that after loading 320 extensions to RAM in my case no any free loop device available even when no any mounted loop.
Just for checking mounting and unmounting an extension manually as a loop device next free loop device remains the same after.
EDIT: Extensions at the end of the list (in alphabetic order) are not installed as no free loop device available.
-
As excited as I am to get started using 2.4... I'm having problems with 2.4rc3. I had a bunch of .tcz's in my hdc4/tce directory. I have been boothing 2.3.1 with the options:
tce=hdc4 noswap laptop
In my menu.lst, I copied the 2.3.1 section for 2.4rc3, pointed it to the new bzImage and tincycore.gz files, and used the same boot options listed above. 2.4rc3 boots right up and displays the flwm desktop, but no extensions are loaded, and if I go to the status window in the control panel, the extensions list is empty. I can can use AppBrowser to install extensions, but they get stored in the home folder (/home/tc), and if I reboot, they will not be automatically reloaded.
I tried to boot with the option tce=hdc4/tce_2_4, to create a new 2.4-specific tce directory, but booting like this does not create the hdc4/tce_2_4 directory, nor does installing apps from appbrowser. ??? What am I doing wrong?
My restored files are ok that come from mydata.gz. I tried to reset .xsession and .profile from /etc/skel to make sure those files where not causing a problem, but I did not see any change in the tce/ situation. Is there something else I can do to troubleshoot this? Or do differently to set up a tce/ that works? Thanks for any help.
--
Mike L.
-
How to identify version of running system? I do not find a solution. A simple text file like /etc/tcversion with 2.4rc3 in would be sufficient.
-
Use version command.
-
/etc/issue would be a good place.
-
Use version command.
Thanks, I didn't know this command.
-
See http://forum.tinycorelinux.net/index.php?topic=724.0:
-
See http://forum.tinycorelinux.net/index.php?topic=724.0:
I have never used TC 1.x started with 2.0rc..
-
There doesn't seem to be much if any difference in the boot time with this version, and I'll certainly miss watching what is going on while booting. I like to see which programs are the pigs, conflicts, things that aren't installing correctly, permission problems, etc, etc, while it's booting. The rotdash is great for those who don't want to know what's happening. (psychologically, it makes the boot seem longer).
-
Let's clean up the DHCP message as well. Patch, also uploaded:
--- dhcp.sh 2009-10-03 11:12:48.000000000 +0300
+++ dhcp.sh.new 2009-10-03 11:13:27.000000000 +0300
@@ -10,7 +10,7 @@
for DEVICE in $NETDEVICES; do
ifconfig $DEVICE | grep -q "inet addr"
if [ "$?" != 0 ]; then
- echo "${GREEN}Network device ${MAGENTA}$DEVICE${GREEN} detected, DHCP broadcasting for IP.${NORMAL}"
+ echo -e "\n${GREEN}Network device ${MAGENTA}$DEVICE${GREEN} detected, DHCP broadcasting for IP.${NORMAL}"
trap 2 3 11
/sbin/udhcpc -b -i $DEVICE -h $(/bin/hostname) -p /var/run/udhcpc.$DEVICE.pid >/dev/null 2>&1 &
trap "" 2 3 11
-
Hi again Robert!
Thanks for your response but I still think there is something that isn't well with the Refresh Menu option, The screen went black when pressing it. All was made from scratch with no backup or restore. I usually test this out when there is a new cut. After that (if nothing dramatically has changed) I just use tce=hda6 and restore=hda1. Those contain the apps I need and the files that have to be restored to make tinycore 2.4.x work on my 8,5 years old laptop. So I really find tinycore very swift and comfortable to use in my daily use of a computer. Thanks a lot to all of you who are making tinycore progress.
Have fun with tinycore,
meo
-
Hello, Robert
How about including the tftp/nfs patches from http://forum.tinycorelinux.net/index.php?topic=2705.0 ?
They allow an nfs share to be used just like a drive for PPR/TCZ.
Other people are using the patch, and have said they would like it added.
-
I have a NIC that requires additional firmware. In some previous versions, I had to install the separate firmware extension. However, in 2.4rc1, it looked like the firmware was being installed automatically. Now in 2.4rc3, I'm getting a message at the end of boot-up "etho: failed to load firmware tigon/tg3_ts05.bin". On the other hand, the NIC appears to be working properly.
-
I just sorted tcz2ram.lst file and learned that each extensions are listed 5 times. In fact in /tce there are 83 extensions. Deleting tcz2ram.lst and rebooting 80 extensions are mounted on /dev/loop0 .... /dev/loop79 while others are lost without error message.
-
Mike...
try booting without the ...noswap....option
noswap ..caused similar problems on earlier version for me...think it's a timing issue
-
I just sorted tcz2ram.lst file and learned that each extensions are listed 5 times. In fact in /tce there are 83 extensions. Deleting tcz2ram.lst and rebooting 80 extensions are mounted on /dev/loop0 .... /dev/loop79 while others are lost without error message.
It is a well known fact our kernel was compiled for 80 loops 0 - 79.
http://forum.tinycorelinux.net/index.php?topic=3008.msg15673#msg15673
Even though, in v2.4, there are now /dev/loop items to 254, you must inform the kernel that such are available. You need to add the boot code max_loops=255
http://www.tinycorelinux.com/faq.html#loops
This will provide loop 0 -254
The other item concerning doubling up of tc2ram.lst is a bug, still part of the full pathing error.
It is an easy fix, and is in fact done for the next cut.
-
As excited as I am to get started using 2.4... I'm having problems with 2.4rc3. I had a bunch of .tcz's in my hdc4/tce directory. I have been boothing 2.3.1 with the options:
tce=hdc4 noswap laptop
In my menu.lst, I copied the 2.3.1 section for 2.4rc3, pointed it to the new bzImage and tincycore.gz files, and used the same boot options listed above. 2.4rc3 boots right up and displays the flwm desktop, but no extensions are loaded, and if I go to the status window in the control panel, the extensions list is empty. I can can use AppBrowser to install extensions, but they get stored in the home folder (/home/tc), and if I reboot, they will not be automatically reloaded.
I tried to boot with the option tce=hdc4/tce_2_4, to create a new 2.4-specific tce directory, but booting like this does not create the hdc4/tce_2_4 directory, nor does installing apps from appbrowser. ??? What am I doing wrong?
My restored files are ok that come from mydata.gz. I tried to reset .xsession and .profile from /etc/skel to make sure those files where not causing a problem, but I did not see any change in the tce/ situation. Is there something else I can do to troubleshoot this? Or do differently to set up a tce/ that works? Thanks for any help.
--
Mike L.
Make sure hdc4 is still hdc4 as you think it is, this is especially true for removal devices that may now be attached to your system.
Upon first boot, go to shell and use fdisk to make sure hdc4 is the proper physical disk.
After you first boot is hdc4 mounted?
Another item to check, could be a needed waitusb=5 boot option. Typically tce directory not getting mounted may need more time to be detected.
Where is your backup data? Also on hdc4 or ?
-
Hello, Robert
How about including the tftp/nfs patches from http://forum.tinycorelinux.net/index.php?topic=2705.0 ?
They allow an nfs share to be used just like a drive for PPR/TCZ.
Other people are using the patch, and have said they would like it added.
With it being so late in this release cycle, and with major infrastructure changes as in 2.4, lets hold off until v2.5 rc1. When 2.4 goes final make me a patch and email to me. I will put it in early in the next release candidate cycle.
-
Robert, is there a non-hastle way of including a visible boot option? If not, I can live with it. Thanks
-
Hi again Robert!
Thanks for your response but I still think there is something that isn't well with the Refresh Menu option, The screen went black when pressing it. All was made from scratch with no backup or restore. I usually test this out when there is a new cut. After that (if nothing dramatically has changed) I just use tce=hda6 and restore=hda1. Those contain the apps I need and the files that have to be restored to make tinycore 2.4.x work on my 8,5 years old laptop. So I really find tinycore very swift and comfortable to use in my daily use of a computer. Thanks a lot to all of you who are making tinycore progress.
Have fun with tinycore,
meo
Booting tinycore with base norestore and selecting Refresh Menu still works fine for me.
I wonder why there is a need to be selecting Refresh Menu, the extension loader handles menu, icons and desktop update. You are using flwm and tinycore?
A black screen and with a large X typically means no window manager was found. If you boot your system and then first select a shell and then Refresh Menu, you should be able to use the Shell window to see if you can type flwm& to restart the window manager.
But at this point I still cannot reproduce your error.
-
Robert, is there a non-hastle way of including a visible boot option? If not, I can live with it. Thanks
I am still tyring a few concepts. I want the average use to see a cleaner consistent boot screen. I do not like hundreds of files being printed to screen. I do not do that with the files being restored, which is very similar. Some my argue that seeing files being restored is more important than extension names being displayed during boot. Some want only a GU curtain like Slim. I do however, wish to provide a way for developer types to see more information as extensions load. Still in progress.
-
On a clean system the first loop device is /dev/loop0. Installing extensions in RAM is consuming a loop device for each extensions. For example after installing Midnight Commander and its 3 dependencies first loop device is /dev/loop4 according to 'losetup -f'
It means, that after loading 320 extensions to RAM in my case no any free loop device available even when no any mounted loop.
Just for checking mounting and unmounting an extension manually as a loop device next free loop device remains the same after.
EDIT: Extensions at the end of the list (in alphabetic order) are not installed as no free loop device available.
This has already been addressed. The umount command did not free loop device when tcz was temporarily mounted to copy to ram. Using the busybox umount -d will and does free the loop device.
So, it was already brought to my attention and is solved in the next cut.
-
I noticed that "zcat $BOOTDIR/tinycore.gz | sudo cpio -i -H newc -d" doesn't seem to work; are we using something different now? (sorry if this is already posted somewhere...)
I just saw that busybox cpio was dropped ; I'll check out GNU
I'm not sure what docs to be looking at regarding remastering. gnutils?
-
Hello Robert!
It beats me to. The thing is that, as I mentioned, boot with the noicons boot code. So there were no icons to start beaver or emelfm and it didnt work to use Run command in the tools menu either. The only option left was to refresh the menu. When I did just that the screen went black with a - or _ sign twinkling in the upper left corner. It really isn't a problem since I have a tce directory and a backupfile I can specify in the boot code. I just thought it was something to mention.
Have fun,
meo
-
gwalther: That did it! :D Thanks greatly. Everyone else: 'noswap' is to be avoided, at least for now I guess. ??? Should this be considered a bug?
roberts: I'm usually running frugally with both tce/ and mydata.gz on a linux ext2 partition on a real hard disk. The main thing I do with that 'other' Linux is download a new tc .iso and adjust my menu.lst. ;)
--
Mike
-
Will do. Thanks!
-
Robert, is there a non-hastle way of including a visible boot option? If not, I can live with it. Thanks
I am still tyring a few concepts. I want the average use to see a cleaner consistent boot screen. I do not like hundreds of files being printed to screen. I do not do that with the files being restored, which is very similar. Some my argue that seeing files being restored is more important than extension names being displayed during boot. Some want only a GU curtain like Slim. I do however, wish to provide a way for developer types to see more information as extensions load. Still in progress.
How about a -verbose boot option.
[ $VERBOSE ] && echo " $FILE"
-
Hello Robert!
It beats me to. The thing is that, as I mentioned, boot with the noicons boot code. So there were no icons to start beaver or emelfm and it didnt work to use Run command in the tools menu either. The only option left was to refresh the menu. When I did just that the screen went black with a - or _ sign twinkling in the upper left corner. It really isn't a problem since I have a tce directory and a backupfile I can specify in the boot code. I just thought it was something to mention.
Have fun,
meo
meo, Eureka! It only took me to realize that you also were doing noicons! I don't know how I missed it Anyway, I can reproduce the fact that menu update does not occur. So now it will be fixed.
Thanks for mentioning it! Everyone's feedback improves the release.
-
gwalther: That did it! :D Thanks greatly. Everyone else: 'noswap' is to be avoided, at least for now I guess. ??? Should this be considered a bug?
roberts: I'm usually running frugally with both tce/ and mydata.gz on a linux ext2 partition on a real hard disk. The main thing I do with that 'other' Linux is download a new tc .iso and adjust my menu.lst. ;)
--
Mike
Likely a timing issue. Would you try again, add noswap waitusb=5
That will help determine if indeed a timing issue.
-
I would like key bindings for enlarge horizontally and vertically at the maximum size, like the 2 buttons near the hide button do.
legalize
-
jls_legalize:
OK, I'll add Ctrl+Alt+H to toggle between max horizontal size and previous non-max size. Same for Ctrl+Alt+V and the vertical resizing. Also try to make the taller/pageup and shorter/pagedown resizing look more intuitive.
--
Mike L.
-
roberts: I tried rebooting with:
tce=hdc4 laptop noswap waitusb=5
I have the same problem. hdc4 does not mount, and I don't have any extentions or my personal files.
-
I've implemented the changes I mentioned to jls_legalize, plus I also added a Ctrl+Alt+'m' to maximize in both dimentions. The updates have been forwarded to the Core Team. I hope people find these new hotkeys and the modified resizing behavior useful and more intuitive. I know I like these changes.
Note that these maximize hotkeys will un-maximize the window if it is already at the maximum, meaning the window will return to the prior non-maximal size.
--
ML