Tiny Core Linux
Tiny Core Base => TCB News => Alpha Releases => Topic started by: Juanito on July 20, 2013, 01:28:47 AM
-
Tiny Core 5.0 Alpha4 is available for public testing:
http://repo.tinycorelinux.net/5.x/x86/release_candidates/
Changelog for 5.0 alpha1:
* kernel update to 3.8.10 with (u)efi boot enabled
* option to use vmlinuz + rootfs.gz + modules.gz or vmlinuz64 + rootfs.gz + modules64.gz (where boot loader permits)
* aterm, freetype, imlib2, jpeg and libpng factored out of Xlibs/Xprogs
* glibc updated to 2.17 and recompiled against 3.8.x kernel headers
* gcc updated to 4.7.2, recompiled against 3.8.x kernel headers and cloog, gmp, mpc, mpfr and ppl
* e2fsprogs base libs/apps updated to 1.42.7
* util-linux base libs/apps updated to 2.22.2
* scm extensions have been dropped from this release
Changelog for 5.0 alpha2:
* util-linux updated to 2.23.1 due to blkid problems with qemu
* /usr/sbin/fstype modified for new blkid output (supress "/dev/sda" type output in /etc/fstab and mount gui)
Changelog for 5.0 alpha3:
* fixed copy2fs bug with tc-load (thanks to Gerald Clark for reporting)
Changelog for 5.0 alpha4:
* localhost.local domain added to /etc/hosts, /usr/bin/sethostname for avahi compatibility/libcups printing
* 50-udev-default.rules adjusted to allow rw access to /dev/rfkill
feedback welcomed and appreciated
-
'finally managed to get several long standing problems solved:
For several kernel versions usb headsets have been broken, but now mine at least seems to work
For several versions of blueman, bluez and pulseaudio there has been a conflict in audio.conf:
bluez/blueman require Enable=Socket
pulseaudio requires Disable=Socket
..using gnome-bluetooth's bluetooth-wizard, it now seems to be possible to pair and play music to bluetooth headphones (a2dp) with "Disable=Socket"
..and use pavucontrol to switch back and forth between the local audio card, bluetooth headphones and usb headset
Screenshot attached :)
-
tinycore GUI not work in VirtualBox?
-
Sorry, but could explain in more detail what the problem is?
-
I try to boot from tinycore iso in VirtualBox,cannt enter the gui desktop.
-
When you follow identical procedure uner 4.x, do you get different results?
-
tinycore 4.7.7 ok.
-
Seems to be an issue with Hyper-V (Windows 8 ) as well. Xvesa yields a black cross on black background only, for all screen resolutions tested. Though, virtual screen seems properly sized according to chosen resolution.
Xorg works well !
-
Hi, could you tell me which programs use some /lib/ libraries in core.gz ?
I do not wish to slow down the development process. My findings are here down.
I extracted the full structure from core.gz, then choroot inside like:
$ sudo chroot . bin/busybox ash
I checked with # ls -alR / every binary simlinked or compiled executable from /bin, /sbin, /usr/bin, /usr/sbin
(to easy searching on another virgin core5.gz copy, I just deleted all busybox 1.20 simlinks from them, inserted a busybox 1.21 static linked with uclib, then run /usr/bin/busybox --install /bin; Now in /bin are only simlinks to busybox, except the blkid. In /sbin, /usr/bin, /usr/sbin are now only the core specific scripts, and some ext2fs programs. This work did not influence the results).
The same "cleaning" was done for /lib and /usr/lib and adjusted that all library are corect by running /sbin/ldconfig to update the cache:
/sbin/ldconfig: uClibc version
/lib:
ld-linux.so.2 => ld-linux.so.2
libpthread.so.0 => libpthread.so.0
libc.so.6 => libc.so.6
* libcrypt.so.1 => libcrypt.so.1
* libutil.so.1 => libutil.so.1
* libe2p.so.2 => libe2p.so.2
* libext2fs.so.2 => libext2fs.so.2
* libcom_err.so.2 => libcom_err.so.2
* libblkid.so.1 => libblkid.so.1
* libuuid.so.1 => libuuid.so.1
libdl.so.2 => libdl.so.2
librt.so.1 => librt.so.1
libnss_files.so.2 => libnss_files.so.2
libnss_dns.so.2 => libnss_dns.so.2
libresolv.so.2 => libresolv.so.2
libnss_compat.so.2 => libnss_compat.so.2
libnsl.so.1 => libnsl.so.1
libanl.so.1 => libanl.so.1
libm.so.6 => libm.so.6 <--asked by libstdc++.so.6
/usr/lib:
? libstdc++.so.6 => libstdc++.so.6
* libsysfs.so.2 => libsysfs.so.2
* libz.so.1 => libz.so.1
* libgcc_s.so.1 => libgcc_s.so.1
libsudo_noexec.so => libsudo_noexec.so
/usr/local/lib:
After manualy tested every /sbin, /usr/bin, /usr/sbin program, I identified by ldd /path/to/prog only some used libraries:
/usr/bin/blkid needs libblkid.so.1, libuuid.so.1, libgcc_s.so.1
/usr/bin/uuidgen needs libuuid.so.1, libgcc_s.so.1
/usr/bin/bcrypt needs libz.so.1
/usr/bin/systool needs libsysfs.so.2
/sbin/pccardctl needs libsysfs.so.2
/lib/udev/pcmcia-socket-startup needs libsysfs.so.2
/usr/bin/sudo needs libutil.so.1, libcrypt.so.1, libz.so.1, libgcc_s.so.1
/sbin/mke2fs (tune2fs, e2fsk) needs libext2fs.so.2, libe2p.so.2, libblkid.so.1, libuuid.so.1
The question: who does use the rest of the core libraries? Thanks in advance.
-
Right, bcrypt is not needed in the base and can be removed.
-
Ok, for anyone interested or looking for headers needed to compile apps in TC-5 alpha, I put up a script which prep's the full sources in lieu of headers. Maybe it'll be useful to someone
http://forum.tinycorelinux.net/index.php/topic,15680.msg91785.html#msg91785 (http://forum.tinycorelinux.net/index.php/topic,15680.msg91785.html#msg91785)
have fun
-
Hi, could you tell me which programs use some /lib/ libraries in core.gz ?
I do not wish to slow down the development process. My findings are here down.
It is not clear what you are about.
Do you imply you suspect or verified some particular difference between 4.x stable vs. 5 alpha?
-
Right, bcrypt is not needed in the base and can be removed.
I think bcrypt is needed in base at least with boot code "protect".
-
Right, bcrypt is not needed in the base and can be removed.
I think bcrypt is needed in base at least with boot code "protect".
Hmmm... Maybe you are right. I have never used 'protect'. BTW it is missing in piCore.
-
@nick65go
All the NSS libraries are dynamically loaded, so you can't determine them as unused via ldd. I think sudo also loads the sudo lib dynamically.
librt, libdl, libnsl, and libresolv may not be used in the base, as busybox has its own implementations, but they are very often used by extensions, such that close to all systems would have them loaded.
libanl, the glibc async DNS library, is the only rarely used one. Whether it makes sense to move into an extension at its small size, I don't know.
-
Hi guys!
Please leave bcrypt in TC!!! That was my only contribution to DSL and subsequently TC and I still have use for it. Would like to have it in TC 5.0 final. Please! :'(
From a grateful user of TC now posting from ophcrack linux,
meo
PS By the way would not mind an update of john.tcz either DS
-
@nick65go
All the NSS libraries are dynamically loaded, so you can't determine them as unused via ldd. I think sudo also loads the sudo lib dynamically.
librt, libdl, libnsl, and libresolv may not be used in the base, as busybox has its own implementations, but they are very often used by extensions, such that close to all systems would have them loaded.
libanl, the glibc async DNS library, is the only rarely used one. Whether it makes sense to move into an extension at its small size, I don't know.
OK, thank you very much! I am just learning during "remastering" my optimised core (kernel 3.10 and busybox 1.21.1, hunting at uclib or musl instead of libc).
-
Please leave bcrypt in TC!!! That was my only contribution to DSL and subsequently TC and I still have use for it. Would like to have it in TC 5.0 final. Please! :'(
hi meo,
i want to support your wish!
bcrypt is essential and tc without protected mode...
is it a distribution for nudes? ;-)
never mind,
i've heard about people that have nothing to hide...
-
tinycore GUI not work in VirtualBox?
The issue seems to be that 'waitforX' fails due to missing libX11.so.6 (formerly part of 'Xlibs.tcz'). Surprisingly enough that library was already "AWOL" in 'TinyCore-5.0.alpha3.iso' without being noticed.
Of course that issue would not just be limited to showing up in VirtualBox.
-
It's not missing on my system.
-
It's not missing on my system.
I can see that it is hard to believe, as it is something rather obvious when all "/usr/lib/*.so" files are missing from such an important TCZ file. Nevertheless I've redone my tests with QEMU (e.g. v0.11.1), VirtualBox (e.g. v4.2.16) and used the "7-Zip File Manager" (all on a Win7 x64 host) to browse the ISO and the contained 'Xlib.tcz' file. But as my earlier observation still stands, I can only conclude that I've got some faulty downloads.
For comparison here are my MD5 checksums for the ISO files in question:bd218f13d8dfd7ff87272a5073b9189c TinyCore-5.0.alpha2.iso
5f0926ba84620d1f3f739b99a128c034 TinyCore-5.0.alpha3.iso
b8644951f99c7fc87604f36b586afac7 TinyCore-5.0.alpha4.iso
with the first one containing a 'Xlib.tcz' (i.e. 'f1d1507e66a1c5ad6f7ac17c4776f779') that still contains the shared libraries in question, and the two latter ones containing a 'Xlib.tcz' (i.e. 'c66b89774b6ba0071fc3f4b9551f075c') where those file have "vanished".
-
I did not use the TinyCore ISO, but used tce-load from a core install. Try a tce-update.
-
My virtual box systems (4 at last count with TC-5) have no issues. I follow the technique advised earlier, apps browser to remove all X extensions, then reboot to command line and use tce-ab to install Xorg which loads all required dep's. on reboot the system boots normally into X
This procedure works when updating from the old Xvesa to latest Xorg-7.7 on virtual box systems and with real hardware
-
I'm not interested in Xorg. One (if not my main) reason for using TC is that I can quickly boot into a small (!!) live system (typically using a VM). IMHO it is crucial that 'Xvesa' continues as the default (or at least the initial) X11 server.
-
Interesting and i can sympathize with your objective, meanwhile core + xvesa + additional ext's still functions right? Or at least did for me until updating to Xorg, I think the choice is yours
Am interested to see the official response for this scenario. Maybe I should try booting a clean system without Xorg ??
-
Things should still work with Xvesa (my current hardware will not work with Xvesa so I cannot test).
Did you try marking Xlibs for removal, rebooting and re-downloading Xlibs to force it to recognise the new deps for Xlibs in tc-5.x?
-
... can confirm that with Hyper-V I also find that after booting from iso, libX11.so.6 is not found. Doing a find cannot locate it. Booting the iso into command line and tce-load -wi the required extensions gets Xvesa working.
-
Confirmed, tinycore.iso needs the additional extensions included.
-
OK, thank you very much! I am just learning during "remastering" my optimised core (kernel 3.10 and busybox 1.21.1, hunting at uclib or musl instead of libc).
Sounds more like a port or a fork than like a remaster, and I fail to understand the reason why such would be posted in the "Tiny Core 5.0 Alpha 4 Testing" thread.
-
OK, thank you very much! I am just learning during "remastering" my optimised core (kernel 3.10 and busybox 1.21.1, hunting at uclib or musl instead of libc).
Sounds more like a port or a fork than like a remaster, and I fail to understand the reason why such would be posted in the "Tiny Core 5.0 Alpha 4 Testing" thread.st
Without much verbal noise, my post in this thread was because my query about the need for all the /lib and /usr/lib files. My "fault" is that I provided unsolicited info about how I play with core.gz. Please ignore the last part of my previous post. Thanks.
-
Confirmed, tinycore.iso needs the additional extensions included.
Tinycore iso removed for the time being
-
On Hyper-V (Windows 8 ) the attached virtual disks are not visible to Core client when booting Core-5.0.alpha4.iso .
fdisk -l yields no output. Used to work with Core-4.7.7.iso .
edit:
problem solved. hv_storvsc driver needs to be loaded separately from scsi-3.8.10-tinycore.tcz . Can we please have that driver included in core.gz ?
-
Does anyone have the egilbc build script from the 4.x series? RPC support has been removed from glibc as of 2.14. The headers are still there, which causes problems, and TI-RPC isn't a drop in replacement.
Andyj
-
I should ask, going forward has TC switched from eglibc to glibc?
Andy
-
PS By the way would not mind an update of john.tcz either DS
1.8.0 is in the repo
-
For tc-5.x, yes
-
TinyCore-5.0.alpha4.iso reposted - please report any problems found
-
Does anyone have the egilbc build script from the 4.x series?
A little off-topic for tc-5.x, but see:
http://www.tinycorelinux.net/4.x/x86/release/src/toolchain/
Note that eglibc was rebuilt after the original toolchain was built as there were some difficulties with making it i486 compatible
-
May I know how to find tc-install or somewhat similar; or does the alpha version not come with that package?
-
I use the tc-install from tc-4.7.7
-
Hi coreplayer2!
Do you just substitute the actual files to test TC-5.0-Alfa4 after installing TC 4.7.7?
Kind regards,
meo
-
You would need to substitute core.gz and vmlinuz
..or rename them to core5.gz/vmlinuz5 and create a new boot loader menu entry
-
Thanks Juanito!
Trying TC 5.0 alfa 4 as I am typing, will just pass on some error messages given during this tryout,
Trying to start chromium,
usr/local/chromium-browser/chromium-browser: error while loading shared libraries: libXss.so.1: cannot open shared object file: No such file or directory :(
Trying to start JTR,
fopen: john.ini: No such file or directory :(
Another obvious thing is that there is no minimize button ( - ) in neither of the extensions and the only web browser with a gui I could make work (using now) is Opera.
Hope this can be of some help and thanks for all the hard work you guys are putting in,
meo
PS Forgot to mention that I was making my tryout using a live cd DS
-
I suggest you read through the alpha testing threads
Xlibs and flwm have changed/been recompiled and Xorg-7.6{,-lib} has become Xorg-7.7{,-lib} - you basically need to be sure you start with a clean set of extensions from the tce-5.x repo.
-
Well as I mentioned post scriptum (PS) I was using a cd with only the iso-file of the reposted TC 5.0 Alfa 4 (about 12 MB in size) burned on it. Diskless in other words, so the environment must have been as pristine as possible when I made this trial of the latest Alfa cut. Nevertheless I got the error messages mentioned above. Hope there are no misunderstandings left.
Have fun improving TC,
meo
-
Sorry - I misunderstood...
Could you confirm the *iso file name you downloaded and the contents of the cde folder on the cd you burnt from it please?
-
Another obvious thing is that there is no minimize button ( - ) in neither of the extensions and the only web browser with a gui I could make work (using now) is Opera.
I find the minimize button has been relocated to the top left of each window (to the left of the title)
And have grown fond of chromium-browser with "Lightning-newtab" google extension in tc-5.0
-
Hóla Juanito!
No tengo el iso que usé pero el cde en el disco contiene:
Folder optional latest change 2013-07-24 16.45 (4.45 PM)
File copy2fs.lst latest change 2013-07-24 16.45 (4.45 PM)
File onboot.lst latest change 2013-07-24 16.45 (4.45 PM)
File xbase.lst latest change 2013-07-24 16.45 (4.45 PM)
Ojalá que esta información sea de ajuda, ;)
meo
PS Tenía el iso en mi instalación de Kali Linux: TinyCore-5.0.alfa4.iso DS :) :) :)
-
It looks to me like chromium-browser is missing a dep on Xorg-7.7-lib
Could you try with this loaded and see if things work?
-
Hi again Juanito!
I seems you were right. Chromium works now. Just posting this from it. ;)
Thanks for your good work with TC,
meo
-
Thanks - dep file updated :)
-
Why the switch from eglibc to glibc?
Andy
-
rhythmbox posted as a way to test various extensions from the tc-4.x repo using copied over and renamed
d gstreamer extensions in tc-5.x repo
-
Maybe these two suggestions about unicode are weird, but I still want to put forward. First is the modification of /etc/profile. Maybe the variable G_FILENAME_ENCODING=UTF-8 will be better. It doesn't affect western file names and can resolve the problems of displaying east Asian file names. The other suggestion is updating fltk to 1.3.x to support unicode, and re-build flwm_topside and Xprogs. At present, the display of ControlPanel will be a mess after I loaded base-locale and set LANG=zh_CN.utf8.
-
When I use the xf86-video-vmware driver in Xorg 7.7 I don't see a mouse cursor. There is still a mouse, because I can roll over the wbar and select windows, but no cursor. I didn't see this problem with 7.6 in TC4, or with the xvesa driver in 7.7. It doesn't seem to make a difference if I use the vmmouse input driver or not. Any ideas?
Andy
-
And it does this pretty regularly too:
[ 1212.151] (EE)
[ 1212.151] (EE) Backtrace:
[ 1212.151] (EE) 0: /usr/local/bin/Xorg (xorg_backtrace+0x49) [0x817f98d]
[ 1212.152] (EE) 1: /usr/local/bin/Xorg (0x8048000+0x13a855) [0x8182855]
[ 1212.152] (EE) 2: linux-gate.so.1 (__kernel_rt_sigreturn+0x0) [0xb77ce410]
[ 1212.152] (EE) 3: /usr/local/lib/xorg/modules/drivers/vmware_drv.so (0xb72e5000+0x78b5) [0xb72ec8b5]
[ 1212.152] (EE) 4: /usr/local/bin/Xorg (0x8048000+0x97068) [0x80df068]
[ 1212.152] (EE) 5: /usr/local/bin/Xorg (0x8048000+0x95e0a) [0x80dde0a]
[ 1212.152] (EE) 6: /usr/local/bin/Xorg (miPointerUpdateSprite+0x17e) [0x817056e]
[ 1212.153] (EE) 7: /usr/local/bin/Xorg (0x8048000+0x128706) [0x8170706]
[ 1212.153] (EE) 8: /usr/local/bin/Xorg (0x8048000+0xa0a52) [0x80e8a52]
[ 1212.153] (EE) 9: /usr/local/bin/Xorg (0x8048000+0xd7858) [0x811f858]
[ 1212.153] (EE) 10: /usr/local/bin/Xorg (0x8048000+0x33ebe) [0x807bebe]
[ 1212.153] (EE) 11: /usr/local/bin/Xorg (WindowHasNewCursor+0x2e) [0x807cf7e]
[ 1212.154] (EE) 12: /usr/local/bin/Xorg (ChangeWindowAttributes+0x998) [0x809b153]
[ 1212.154] (EE) 13: /usr/local/bin/Xorg (0x8048000+0x28a49) [0x8070a49]
[ 1212.154] (EE) 14: /usr/local/bin/Xorg (0x8048000+0x2d5b9) [0x80755b9]
[ 1212.154] (EE) 15: /usr/local/bin/Xorg (0x8048000+0x1f83f) [0x806783f]
[ 1212.154] (EE) 16: /lib/libc.so.6 (__libc_start_main+0x13e) [0xb73c1522]
[ 1212.154] (EE) 17: /usr/local/bin/Xorg (0x8048000+0x1fa35) [0x8067a35]
[ 1212.154] (EE)
[ 1212.155] (EE) Segmentation fault at address 0xb60f2000
[ 1212.155] (EE)
Fatal server error:
[ 1212.155] (EE) Caught signal 11 (Segmentation fault). Server aborting
Andy
-
It doesn't seem to make a difference if I use the vmmouse input driver or not. Any ideas?
Sorry, but I've never used vmware anything so I can't help there...
-
Apparently for now DRI for VMware on TC5 isn't really optional, which means that I have to load even more extensions that I normally wouldn't bother with. I did finally manage to get open-vm-tools working again. But I have to ask, why are so many of the X libraries unbundled? Can they really be used for something other than X, and don't they all get loaded by default anyway when the Xorg extension gets loaded? The only application that I have loaded is open-vm-tools and that's mounted on loop 88. Doesn't that seem like a lot, especially when so many of the extensions are under 40kb?
Andy
-
Yes, it'd be more efficient to have them together, but the other end is maintenance: it's much easier to update one lib if it's in its own extension.
-
Now that Linux 3.10 has LTS status, any chance of upgrading while we're still in the alpha stage?
-
fluxbox.tcz (v1.3.5) + docs submitted for the tc-5 (3.8.10) repo
tested on Xorg
(http://db.tt/tdg49wNY)
(http://db.tt/blTi7BLP)
and Xvesa
(http://db.tt/LWYVbau3)
fluxbox-locale.tcz maybe copied from tc-4 repo
-
Sorry, but I think it's a little late for that..
-
Yes, I think we're to RC soon.
-
Aterm under VMware DRI doesn't always refresh like it should, resulting in garbage in the window. Xterm from 4.x doesn't have this problem. Has anyone noticed if aterm has issues in any other DRI environment (Intel, ATI, NV)? Why are we using aterm anyway? According to the website, it's only one step away from abandonware.
-
Why are we using aterm anyway? According to the website, it's only one step away from abandonware.
There's no particular reason to use aterm, if you'd like to submit a different extension, please feel free :)
-
It was the smallest workable terminal the last time things were measured.
-
It was set up as the default which is why I was using it, but xterm from 4.x seems to work fine. Almost anyway, it isn't being added to tce.installed when it's installed the first time from onboot.lst. Running tce-load -i fixes this, but then there are two terminal apps in wbar. I also learned that urxvt from the 4.x series has the same refresh problem that aterm has. Not a huge surprise, since there is probably some inherited code.
-
So how many actual alpha testers are there anyway? I have plymouth working in a VM with support for Intel and Radeon but no way to test these. Nouveau is broken at the moment. Any takers?
-
Several window managers have been posted - fluxbox, icewm, openbox, hackedbox, jwm - testing and comments on possible solutions to any problems found would be appreciated :)
-
Hi guys!
Isn't it time for the first RC now? I think that would make it easier for many to test out different combinations of window managers etc. Just a suggestion since it's a lot easier to install TC from a CorePlus.iso than put together a testing environment in another way.
Kind greetings,
meo
-
Isn't it time for the first RC now?
For me the question is more like, "What milestones do we have to reach before we can move from the alpha to the beta (RC) stage?"
Is there an official project plan listing such milestones? If not, then how do you decide?
-
I've been trying to get Xorg running on a tablet with the dreaded GMA500 video. On TC5, simply loading gma500_gfx instantly causes my tablet to crash hard, corrupting the screen and locking up, while the Xorg vesa and fbdev drivers also fail for various reasons, albeit more gracefully. Anybody masochistic enough to want to help me get this cruddy GMA500 stuff working on TC5? Ubuntu 13.04 loads up gma500_gfx and Xorg with the modesetting driver right out of the box, so I know that the hardware is working and the drivers are capable. Thanks!
-
Actually CONFIG_DRM_GMA3600=y option is needed, but it's not set because it's seen as 'experimental', eventhough on other distros (I've tested meanwhile), it is set and it is working beautifully.
I've tested the same kernel on this type of computers (that needs this option in order to use xorg in modesetting mode), and on other older ones (10 years old that work without).
I didn't notice any problem on any machine because of this 'experimental' option.
I feel I'm stuck, unless:
I go my way in my corner to solve my problem, just to keep using TC on ALL my computers,
or
make a wish that a TC kernel could be provided, with this option set, and be an 'exceptionnally' exception for the team policy, and start testing TC5 as soon as possible.
thank you for considering this as a request from me, and I think from many others on this forum they didn't show up yet.
Regards