Tiny Core Linux
Tiny Core Base => Raspberry Pi => piCore Final Releases => Topic started by: Paul_123 on June 23, 2023, 12:29:47 PM
-
Team Tiny Core is happy to announce the release of piCore-14.0. Both 32 and 64bit images are ready for download/testing.
rpi-kernel 6.1.25
gcc 12.2
binutils 2.39
glibc 2.37
busybox 1.36
micropython 1.19
libffi (v8 library)
piCore 14 also brings openssl to 3.1.0. For applications in the 14.x repo not yet rebuilt, there are compatibility extensions created openssl-1.1.1 and libffi7
32bit - http://www.tinycorelinux.net/14.x/armv7/releases/RPi/piCore-14.0.1.zip (http://www.tinycorelinux.net/14.x/armv7/releases/RPi/piCore-14.0.1.zip) (This should run on all rpi boards)
64bit - http://www.tinycorelinux.net/14.x/aarch64/releases/RPi/piCore64-14.0.zip (http://www.tinycorelinux.net/14.x/aarch64/releases/RPi/piCore64-14.0.zip) (rpi zero2W, pi3, pi4)
* Note, no changes from the last beta releases.
-
hi as you know I have a desktop with piCore64 currently getting a failed for waitforX on arm7.
I am trying to compare my 64 bit installed list vs 32 and so far I have eliminated
imlib2-bin as it appears 32 bit combines or does not split out the bin files.
I have not been able to eliminate libxcvt which is a dep of xorg-server on 64 bit but is missing in 32 bit.
could this be what I am missing?
-
Hi
I wanted to post on the beta page but it appears I can not so forgive me not testing 32 bit at the beta stage. I have built libxcvt for 32 bit but not submitted as it makes no difference. I am now stumped. Here is what I did, and I kept the output.
1) make my boot list on 64 MATCH as far as possible the tce.installed list here for 32 bit
cat /mnt/mmcblk0p2/tce/onboot.lst
openssh.tcz
libXcursor.tcz
nano.tcz
mc.tcz
libxcvt.tcz
liblz4.tcz
libzstd.tcz
llvm15-lib.tcz
xf86-video-fbdev.tcz
Xorg.tcz
jwm.tcz
aterm.tcz
wbar.tcz
2)cd /mnt/sda1 # its my storage usb
ls /usr/local/tc.installed > /mnt/sda1/32c.txt
diff 32c.txt 64.txt > diffc.txt
cat diffc.txt | grep +
+++ 64.txt
@@ -5,24 +5,18 @@
+imlib2-bin
@@ -40,22 +34,18 @@
@@ -72,15 +62,11 @@
so as per prev post, I am eliminating imlib2-bin as its a combined TCE AFAIK on 32 bit
and now I can eliminate libxcvt as its built on 32 bit
full reboot once libxcvt built and onboot list match
3) startx still fails....so I am stumped at this stage.
thanks for reading
##############
libxcvt.tcz
liblz4.tcz
libzstd.tcz
llvm15-lib.tcz
xf86-video-fbdev.tcz
were + diffs from diffb I had run previously
-
Only newer versions of xorg-server (not the version in the armv* repos) need libxcvt.
What happens if you start the server with "Xorg -nolisten tcp"?
Note that for xorg 3d to work in piCore /usr/local/share/X11/xorg.conf.d/fbturbo.conf needs to be deleted
-
Hi Juanito
I am loading Xorg not Xorg-3d
my ~/.xsession is the same from /etc/skel which has that stanza in the top line.
Based on your reply, I removed libxcvt and full reboot but kept the other TCEs as above. still no desktop.
Those extra TCEs were not in my original boot list and of course no desktop either.
sorry to say I am still stumped.
Maybe someone who is using 14.x can share their thoughts too?
-
When you get the “failed waiting for x” message it’s useful to troubleshoot using the first line of .xsession.
-
Hi Juanito
bad and good news
bad news (1) . That repeat command of top line of .xsession showed I needed to load prce (1)
I have rebooted with a new boot list and using openssh for this reply as no web browser currently usable for fussy me.
bad news (2) relaunching startx...for short time jwm wallpaper appears but then goes blank with invisible mouse even tho I have libXcursor loaded, but wbar appears and iconification of wbar items is kind of usable so new boot list
openssh.tcz
libXcursor.tcz
nano.tcz
mc.tcz
pcre.tcz
Xorg.tcz
flwm.tcz
aterm.tcz
wbar.tcz
full reboot....good news.....desktop appears as expected with working wbar, context menu etc etc
I plan to update jwm
-
Hi Juanito and Paul_123
TCE firefox on arm7
firefox &
tc@box:~$ XPCOMGlueLoad error for file /tmp/tcloop/firefox/usr/local/lib/firefox/libmozgtk.so:
libffi.so.6: cannot open shared object file: No such file or directory
Couldn't load XPCOM.
TCE epiphany on arm7
epiphany &
tc@box:~$ epiphany: error while loading shared libraries: libffi.so.6: cannot open shared object file: No such file or directory
[1]+ Done(127) epiphany
Looks like these need a rebuild against libffi8?
I have not tried links so now a quick look at dillo
Navigating to this forum, multiple certificate requests....popup options include continue -> I do
Attempting to login...currently not working for me.....I will try again later. I might need to update ca-certs and add to backup will check later
-
Hi Paul_123
arm7 Apps (GUI) -> check for updates results -> I have not yet actioned.
Apps claims I need to update...checking info
atk-dev................ last updated 2020
ca-certificates...... last update 2022
gettext...................last update 2023/04/30
gtk2-dev................last update 2020............ I do plan to rebuild this so not fussed about this TCE update msg
-
ca-certificates updated to the same version in aarch64. I just missed that when building the repo.
pcre.tcz added to xorg-server, harfbuzz, fontconfig. As they all had a dependency on pcre, and could be used independently of X.
gettext is built as a part of the toolchain, PM me as to why you think it needs updated.
-
Hi Paul_123
pm not needed. in post 8 I was reporting that GUI=Apps -> check for updates thought I needed updates, not a request to update a TCE.
I have found the reason this morning. Altho I have the md5 files....they are not md5 files! as per
cd /mnt/mmcblk0p2/tce/optional/
tc@box:/mnt/mmcblk0p2/tce/optional$ file gtk2-dev.tcz.md5.txt
gtk2-dev.tcz.md5.txt: data
tc@box:/mnt/mmcblk0p2/tce/optional$ file gettext.tcz.md5.txt
gettext.tcz.md5.txt: data
tc@box:/mnt/mmcblk0p2/tce/optional$ file atk-dev.tcz.md5.txt
atk-dev.tcz.md5.txt: data
Please check what kind of files you have in the repo.
2) last night it was late....checking Apps -> update check just now the right pane claims
"You are running version 14.0
The latest release is 13.0.3"
3) I know how to generate new md5 txt files...EDIT but processed Apps updates in case there was a dep change too.
cheers :D
-
No issues with MD5's on the TC server. Are you using a mirror?
And thanks for the reminder to update the latest release files.
-
No so it must have been an internet gremlin my end. No worries.
-
piCore arm6/7/7l
Ok I was going to rebuild gtk2 against pcre2 as
root@box:/tmp/tcloop/gtk2/usr/local/bin# readelf -d gtk-query* | grep 'NEEDED' | grep libpcre
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
but gtk2 rebuild borks ....configure and config.log looked ok before I tried make -j4
SNIP
/usr/local/bin/ld: warning: libffi.so.6, needed by /usr/local/lib/libwayland-client.so.0, not found (try using -rpath or -rpath-link)
This will delay me attempting to update gtk2, pcmanfm, leafpad. I will now look at others
Thanks for reading
2) Would you like me to inject libffi.so.6 from an earlier release to continue build?
-
libffi6 added, somewhat inelegantly, for now - updating various x components/wayland is nearing the top of my todo list.
wayland dep file adjusted.
Do we still want xf86-video-fbturbo, which doesn't work in aarch64, stops xorg-3d from working in armv* and is not maintained as the default or can we move to xf86-video-fbdev?
-
In preview I am getting server error already posted before about so removing lots changing terms etc
inxi -Gxxx | grep loaded....Display: server: X.org driver: X: loaded: fbturbo gpu: bcm2708_fb
I have not fiddled with partition 1 configs.
swapping out that xf86-video-fbturbo for xf86-video-fbdev for Xorg.tcz.dep
Full reboot....yep desktop appears
inxi -Gxxx | grep loaded....Display: server: X.org driver: X: loaded: fbdev unloaded: modesetting
I do not know enough about RPi to know if I am supposed to have modesetting
-
modesetting needs graphics-KERNEL
-
Do we still want xf86-video-fbturbo, which doesn't work in aarch64, stops xorg-3d from working in armv* and is not maintained as the default or can we move to xf86-video-fbdev?
I'm not sure fbturbo did anything significant for the Raspi, and we should just move to keep 32 and 64 bit as close as we can.
-
Do we still want xf86-video-fbturbo, which doesn't work in aarch64, stops xorg-3d from working in armv* and is not maintained as the default or can we move to xf86-video-fbdev?
I'm not sure fbturbo did anything significant for the Raspi, and we should just move to keep 32 and 64 bit as close as we can.
I don't know how well it works currently in practice, but this RPi forum post (https://forums.raspberrypi.com/viewtopic.php?t=65037#p478857) describes how fbturbo was supposed to improve X performance.
But I think that was before the Linux kernel's vc4 (https://www.kernel.org/doc/html/latest/gpu/vc4.html) DRM driver was available via the Xorg "modesetting" driver. I'm not sure how well that's supported in Xorg on PiCore (EDIT: I see vc4.ko is in graphics-KERNEL.tcz, so I guess that's all you need), but in theory it would be better than either fbdev or fbturbo because it uses the GPU (I think). It's not clear how much help it is for 2D acceleration though.
-
Hi Juanito
thanks but not showing modesetting....lets leave it as I have other issues
-
Hi Paul_123
I stole a bit of your build script which reads
# Export variables needed for compilation
case $(uname -m) in
arm*)
export CFLAGS="-flto=auto -Os -pipe -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp"
export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp"
;;
aarch64)
export CFLAGS="-flto=auto -Os -pipe -march=armv8-a+crc -mtune=cortex-a72"
export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -march=armv8-a+crc -mtune=cortex-a72"
;;
esac
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:/usr/lib/pkgconfig
I am attempting to copy and paste initially from my TC64 via openssh. I am connected but inputting above...too much stuff to post but the end bits suggest it is not working as
SNIP
locale zstdmt
locate zsync
logger
root@box:/tmp# export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -march=armv8-a+crc -mtune=cortex-a72"
root@box:/tmp# ;;
sh: syntax error: unexpected ";;"
root@box:/tmp# esac
sh: syntax error: unexpected "esac"
root@box:/tmp# export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:/usr/lib/pkgconfig
root@box:/tmp# uname -m
armv7l
root@box:/tmp# echo $CFLAGS
-flto=auto -Os -pipe -march=armv8-a+crc -mtune=cortex-a72
this is my first attempt to use an arch test set flag
After sshing in....I become root via sudo su
I can do it manually but any clues?
Is not possible to do it by openssh?
###############
something is wrong with first bit...it fails and then flags are being set by the arm8 commands
-
In aarch64 and armv7l it gives this:
inxi -Gxxx | grep loaded
Display: server: X.org driver: X: loaded: modesetting dri: vc4
-
in post 20 I was using openssh copy and paste into a terminal on TC64 sshed to RPi
in this post I am on RPi with copy and paste just the flags to aterm. I am root
root@box:/tmp# export CFLAGS="-flto=auto -Os -pipe -march=armv6zk -mtune=arm1176 jzf-s -mfpu=vfp"
root@box:/tmp#
. logname
: logread
X losetup
Xnest ls
SNIP
logger zcat
login zsync
root@box:/tmp# export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -
march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp"
last line is not yet input....so me thinks its first line?
EDIT if I input the bits starting from -pipe.....I am not seeing any garbage
-
My build systems are all headless, so I'm ssh'd into them all the time. I've never seen cut and paste problems like that.
I use putty for my ssh client.
-
hmmm well the good news is that back on RPi I can use rxvt. I still get garbage when I copy and paste from
case $(uname -m) in.......(to) esac but echo $CFLAGS && echo $CXXFLAGS is returning the corrrect flags. So it looks like I will build natively.
tomorrow awaits.
thanks for your feedback
-
Congrats for the effort to release version no 14!
I have one issue: I use Roon Bridge to listen to music, and in Tiny Core v 14 sound quality seems to be lower than on picore 13. It could be from new kernel, or the new ALSA, or I don’t know. Maybe someone has experimented the same.
-
Sound quality is so subjective. I cannot say that I noticed anything.
Are you using any of the alsa plugins/conversions or the alsa hw:xxxxx devices?
-
I agree with you that is subjective, generally speaking.
But I'm an audiophile with high-end audio gears, so any small difference in SQ is immediately perceived.
I'm using exactly the same hardware components, and stable version is Roon Bridge on piCore 64 13.1.0
And I use alsa hw:xxxx device
-
on piCore64-14.x I use firefox-perf which does not need apulse, and can play youtubes without pulseaudio. Its sound sounds good to my ears.
I tend to convert to mp3 which I know causes a loss of quality and play it thru parole on bcm2835 but swapped headphones for a desktop line out. I have always kept some fav mp3s. I can not detect any significant drop in quality given its already a drop in quality. The fact that I use mp3s should confirm I am not an audiophile. ;D
-
Lol aus9. Not discounting cosminbro, but I stream a lot of internet radio. So I’m not a good candidate for evaluation either.
Since you are using the hw:xxxx card, that should mostly rule out alsa. But the newer kernel is for sure a possibility. There are many reports that old 3.x and early 4.x kernels sound better. But we have to keep up with rpi versions for new hardware support. We use basically the same kernel that RaspiOS uses.
There are many blog posts around that explain tweaks to kernel runtime parameters that are perceived to make improvements.
-
OK good news for openssh copy and paste.
the flags are defined well by changing the " to ' via openssh.
It might be possible for someone on arm 32 for 14.x series to confirm that aterm is producing garbage by inputting
case $(uname -m) in
arm*)
export CFLAGS="-flto=auto -Os -pipe -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp"
export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp"
;;
aarch64)
export CFLAGS="-flto=auto -Os -pipe -march=armv8-a+crc -mtune=cortex-a72"
export CXXFLAGS="-flto=auto -Os -pipe -fno-exceptions -fno-rtti -march=armv8-a+crc -mtune=cortex-a72"
;;
esac
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:/usr/lib/pkgconfig
if no-one sees garbage there must be something wrong my end....other than my brain ;D
-
so I had submitted but cancelled jwm on arm6/7/7l as I still had a "hidden" dep of pcre, meaning not prce2.
I am on openssh build ATM and here is what I am seeing.
P=jwm and I build at /tmp
file jwm/usr/local/bin/* | grep ELF
jwm/usr/local/bin/jwm: ELF 32-bit LSB pie executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-armhf.so.3, for GNU/Linux 5.15.1, stripped
# jwm has no shared objects
Now the misleading bit
readelf -d $P/usr/local/bin/$P | grep 'NEEDED' | grep pcre
# returns no hits as pcre nor pcre2 are hits
misleading because of some TCE needing pcre as per
ldd $P/usr/local/bin/$P | grep pcre
libpcre.so.1 => not found
libpcre.so.1 => not found
libpcre.so.1 => not found
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb6580000)
libpcre.so.1 => not found
Does anyone know how to track down....when pcre is not even loaded why ldd thinks its needed?
I am thinking I need to run the readelf to track down 4 TCEs of all currently loaded?
And I am confirming when I rebooted with jwm in my boot list without pcre....I got the failed for waitforX until I loaded pcre
Thanks for reading
-
woo hoo started with only openssh...in boot list which I should have done before sorry and we have in loading make deps for jwm
fontconfig.tcz: OK
fc-cache: error while loading shared libraries: libpcre.so.1: cannot open shared object file: No such file or directory
fontconfig-dev.tcz: OK
as fontconfig had been in my boot list .....I missed seeing these messages for pcre
/tmp/tcloop/fontconfig/usr/local/bin# readelf -d * | grep 'NEEDED' | grep pcre
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
ok I hope its only the one TCE with all of its executables ....nine of the little rascals.
ls
fc-cache fc-conflist fc-match fc-query fc-validate
fc-cat fc-list fc-pattern fc-scan
-
Nope....started to do a private update of fontconfig and now I get at the make stage
numerous mentions of
/usr/local/bin/ld: warning: libpcre.so.1, needed by /usr/local/lib/libfreetype.so
so it looks like our new list is fontconfig and freetype.
-
started to complete private update of freetype and found pcre for ldd and its harfbuzz,
tc@box:/tmp/tcloop/harfbuzz/usr/local/bin$ readelf -d * | grep 'NEEDED' | grep pcre
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
0x00000001 (NEEDED) Shared library: [libpcre.so.1]
we are getting there.
In looks like it may take me a bit longer than I hoped but I prefer to submit at least one desktop without pcre dep. Due to my pref report openssh issue I was building natively...but doing so with pcre issues. Once I resolve these I can go back to native build.
Thanks for reading
-
fontconfig, freetype and harfbuzz updated in armv* repos.
-
Hi Juanito
thanks for that. I noticed LFS thinks harfbuzz and freetype are make dependencies of each other....laughs.I was cheating off your freetype so never got to get around to harfbuzz. I admit that I over rely on readelf and I know another user who uses it too. Its not enough. I have to use ldd too but progress this morning is miles better than last night.
readelf -d $P/usr/local/bin/$P | grep 'NEEDED'
0x00000001 (NEEDED) Shared library: [libX11.so.6]
0x00000001 (NEEDED) Shared library: [libpng16.so.16]
0x00000001 (NEEDED) Shared library: [libz.so.1]
0x00000001 (NEEDED) Shared library: [libjpeg.so.62]
0x00000001 (NEEDED) Shared library: [libXft.so.2]
0x00000001 (NEEDED) Shared library: [libXrender.so.1]
0x00000001 (NEEDED) Shared library: [libXpm.so.4]
0x00000001 (NEEDED) Shared library: [libXinerama.so.1]
0x00000001 (NEEDED) Shared library: [libc.so.6]
0x00000001 (NEEDED) Shared library: [ld-linux-armhf.so.3]
root@box:/tmp# ldd $P/usr/local/bin/$P | grep pcre
libpcre.so.1 => not found
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb6490000)
I know which one it is by going into each /tmp/tcloop at my current skill level. Its libXft.But there could be more.
thanks for reading
-
Ok there is more, for arm 6/7 ...after building libXfont2 and libXft...jwm builds with no ldd hits for pcre(1)
I regret to inform you that xorg-server and libudev remain. And that is with a very small boot list. Sorry about this.
I can have a go at libudev but xorg-server is above my skill set. I can submit libXfont2 and libXft if interested?
trivia from openssh session
/tmp/tcloop$ ldd xorg-server/usr/local/bin/* | grep pcre
xorg-server/usr/local/bin/X: error while loading shared libraries: xorg-server/usr/local/bin/X: invalid ELF header
libpcre.so.1 => /usr/local/lib/libpcre.so.1 (0xb67c0000)
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb60e0000)
xorg-server/usr/local/bin/Xorg: error while loading shared libraries: xorg-server/usr/local/bin/Xorg: invalid ELF header
libpcre.so.1 => /usr/local/lib/libpcre.so.1 (0xb6940000)
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb60e0000)
/tmp/tcloop$ ldd libudev/usr/local/lib/* | grep pcre
libpcre.so.1 => /usr/local/lib/libpcre.so.1 (0xb6cf0000)
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb6a80000)
libpcre.so.1 => /usr/local/lib/libpcre.so.1 (0xb6d00000)
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb6a90000)
libpcre.so.1 => /usr/local/lib/libpcre.so.1 (0xb6c60000)
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb69f0000)
-
No need thanks, I’ve started to update the x libs.
-
fontconfig, freetype and harfbuzz updated in armv* repos.
With updated freetype I have the same problem mentioned in http://repo.tinycorelinux.net/13.x/x86/tcz/src/freetype/compile_freetype (http://repo.tinycorelinux.net/13.x/x86/tcz/src/freetype/compile_freetype) with firefox
-
Hi Juanito and Paul_123
Thanks for your patience and skills. Based on Juanito's last post I will delay any rebuilds. A number of my packages that I pulled from submission process show ldd hits for libffi6. I assume we are not concerned about lifffi6?
thanks
-
Hi mortegai
I have firefox.tcz on arm7l with 8 tabs open. How many do you see a crash please?
-
Hi mortegai
I have firefox.tcz on arm7l with 8 tabs open. How many do you see a crash please?
No tab opens. The error occurs as soon as you start Firefox
-
arm8 firefox-perf opens for me needs libffi7 loaded
arm7l firefox opens for me needs libffi6
-
With updated freetype I have the same problem mentioned ... with firefox
Firefox crashed occasionally for me before and after the update to freetype.
-
I assume we are not concerned about lifffi6?
We should aim to eliminate deps on libffi6 and libffi7.
-
Hi Juanito
I am focussing only on arm6/7 gtk this post. And glad I pulled all my submissions based on your libffi6 reply
paste is from aterm....seems to truncate the pixbuff line sorry about that...gtk2 is my baby
tce-load -i gtk2
hicolor-icon-theme.tcz: OK
libcroco.tcz: OK
g_module_open() failed for /usr/local/lib/gdk-pixbuf-2.0/2.10.0/loaders/libpixbu file or directory
librsvg.tcz: OK
adwaita-icon-theme.tcz: OK
gtk-query-immodules-2.0: error while loading shared libraries: libffi.so.6: cann
gtk2.tcz: OK
tce-load -i gtk3
libxml2.tcz: OK
wayland.tcz: OK
wayland-protocols.tcz: OK
libxkbcommon.tcz: OK
libXi.tcz: OK
libXrandr.tcz: OK
fribidi.tcz: OK
libGLES.tcz: OK
libEGL.tcz: OK
cairo.tcz: OK
pango.tcz: OK
libXcomposite.tcz: OK
gdk-pixbuf.tcz: OK
atk.tcz: OK
dbus.tcz: OK
at-spi2-core.tcz: OK
at-spi2-atk.tcz: OK
shared-mime-info.tcz: OK
gtk-query-immodules-3.0: error while loading shared libraries: libffi.so.6: cann
gtk3.tcz: OK
EDITgtk4 OOPS it has libffi6...I must wear my goggles more often. Loading of gtk4 gave no errors.
tc@box:/tmp/tcloop/gtk4/usr/local$ ldd bin/* | grep pcre
libpcre2-8.so.0 => /usr/local/lib/libpcre2-8.so.0 (0xb59f0000)
SNIP
tc@box:/tmp/tcloop/gtk4/usr/local$ ldd bin/* | grep libffi
libffi.so.8 => /usr/lib/libffi.so.8 (0xb59d0000)
libffi.so.6 => /usr/local/lib/libffi.so.6 (0xb54f0000)
SNIP
Juanito
I know you will be very busy with Xorg libX stuff. Would you like me to have a go at gtk3/4?
-
I know you will be very busy with Xorg libX stuff. Would you like me to have a go at gtk3/4?
You can if you'd like, otherwise I'll get there eventually.
-
Hi Juanito
arm6/7* at this stage only
Thanks I only have one gtk4 TCE downloaded= epiphany web browser so likely pm when built. And you are miles faster than me so if I have not sent a pm assume I am slower than snail, eh
If you don't mind me asking on a TCB post, I assume your Xorg rebuild will include libEGL as thats got a libffi6 rascal in it?
And that stops me rebuilding gtk2. gtk3 affects rebuild of ghostscipt and lxterminal. lxterminal (latest) no longer looks for gtk2 and can build against gtk4. So if I build gtk4 I will be able to test it.
-
Yes, I was speaking of armv6/7/7l.
The x libs rebuild will include mesa, wayland and xorg-server.
-
The archive with 6.1.21 kernel for 64bit is corrupt: rpi-6.1.21-piCore64.tar.xz
http://tinycorelinux.net/14.x/aarch64/releases/RPi/src/kernel/rpi-6.1.21-piCore64.tar.xz
-
just tested, its not corrupt.
The md5sum is: ea80eb251fe6d2aff56e2ee5d995c6e22533dfe8 rpi-linux-6.1.25.tar.xz
-
just tested, its not corrupt.
The md5sum is: ea80eb251fe6d2aff56e2ee5d995c6e22533dfe8 rpi-linux-6.1.25.tar.xz
it gave me error on untar
-
Hi Paul_123
cosminbro is discussing 6.1.21 and I get similar result
md5sum *
f9dce1e21b3bc9762d7501cd9e190b91 rpi-6.1.21-piCore64.tar.xz
tar -xvf *.xz
tar: corrupted data
tar: short read
tc@box:~/Downloads$ xz -d *.xz
xz: corrupted data
whether he should be downloading the 6.1.25 is something else eh?
-
That is not the released kernel. That version of the kernel had a few problems. I need to remove the 6.1.21 files.
-
Hi Paul_123 and Juanito
to save sending a third pm....I have spotted the error and manually fixed my TCE list file for the possible submission which is rejectable.
That will teach me to not read what I send...or will it? ;D
-
libudev renamed to udev-lib to closer align with x86_64
gudev files have been split out, it was not obvious if anything really depended on libgudev.so but it you find a missing dep, let me know.
armvx repo is updated, I’ll get aarch64 a bit later
-
Hi
I know this is TCB post but based on that udev-lib change, members may like to run
Apps -> Dependencies & deletions -> dependencies -> Update .dep files
With aussie net speed and lots installed that takes about 2M40 YMMV
For me, the dep changed are for bluez.tcz.dep, libdrm.tcz.dep, libinput.tcz.dep, libmbim.tcz.dep, libusb.tcz.dep, libv4l2.tcz.dep
and pulseaudio.tcz.dep
Thanks to Juanito and Paul_123 for looking into this.
-
Hi Paul_123
tc@box:/tmp/tcloop$ readelf -d libmbim/usr/local/bin/mbimcli | grep 'NEEDED' | grep udev
0x00000001 (NEEDED) Shared library: [libgudev-1.0.so.0]
0x00000001 (NEEDED) Shared library: [libudev.so.0]
The other TCEs I have named for dep changes look good to me. cheers
-
Hi Juanito or Paul_123
We appear on arm7l and maybe others missing colord?
on arm7l I try
tce-load -w gtk4-dev
gtk4-dev.tcz.dep OK
Downloading: colord-dev.tcz
Connecting to repo.tinycorelinux.net (128.127.66.77:80)
wget: server returned error: HTTP/1.1 404 Not Found
md5sum: colord-dev.tcz.md5.txt: No such file or directory
Error on colord-dev.tcz
Apps GUI total tab also says colord-dev not found. FYI
EDIT I was going to offer to see if I can build it but I have make error against sane at the moment which is a build dep of colord
that assumes I needed to update sane. Maybe keep current sane.
-
I think that's probably a mistake in the gtk4-dev dep file - adjusted
-
Hi Juanito
It might be a carry over from x86_64 repos which contain:
colord-dev.tcz
colord-gir.tcz
colord-gtk-dev.tcz
colord-gtk-gir.tcz
colord-gtk.tcz
colord.tcz
The comments from http://tinycorelinux.net/14.x/x86_64/tcz/gtk4.tcz.info state:
can optionally use the colord, libavformat4, libcups, libswscale4,
gst-plugins-bad and rest extensions
-
Hi Juanito
server error on preview so mod output does not like my xorg lines nor gtk4 stuff
I am planning on testing vkcube from vulkan-tools and have done or seen the following with no joy on modesetting
inxi -Gxxx | grep loaded
Display: server: X.org driver: X: loaded: fbturbo
cat /var/log/Xorg.0.log | grep accel
SNIP
FBTURBO(0): G2D hardware acceleration can't be enabled
FBTURBO(0): no 3D acceleration because the driver has been compiled without libUMP
bootlist
openssh.tcz
bash.tcz
nano.tcz
mc.tcz
graphics-KERNEL.tcz
Xorg-3d.tcz
jwm.tcz
leafpad.tcz
firefox.tcz
aterm.tcz
lxterminal.tcz
pcmanfm.tcz
cat /mnt/mmcblk0p1/config.txt | grep pi4
dtoverlay=vc4-kms-v3d-pi4
I know you have modesetting.....where I am going wrong please?
-
vkcube: /usr/src/Vulkan-Tools-1.2.141/cube/cube.c:3175: demo_init_vk: Assertion `!err' failed.
-
I know you have modesetting.....where I am going wrong please?
Did you delete the fbturbo code snippet in /usr/local/share/X11/xorg.conf.d?
-
No I did not. Instead as a response I removed that TCE as a dep of Xorg
initial reboot gave me a failed for waitforX
I have made some changes to config.txt for better or for worse at my own risk etc etc
configs now
hdmi_safe=1
config_hdmi_boost=4
# dtoverlay=vc4-kms-v3d-pi4
dtoverlay=vc4-fkms-v3d
Boot list relevant bits unchanged
graphics-KERNEL.tcz
Xorg-3d.tcz
then the desktop etc
finally
inxi -Gxxx | grep loaded
Display: server: X.org driver: X: loaded: modesetting alternate: fbdev
vkcube
vkcube: /usr/src/Vulkan-Tools-1.2.141/cube/cube.c:3175: demo_init_vk: Assertion `!err' failed.
Aborted
I know you set run dir but its not set with current TCE loaded
I know weston will set it
vulkaninfo
error: XDG_RUNTIME_DIR not set in the environment.
ERROR at /usr/src/Vulkan-Tools-1.2.141/vulkaninfo/vulkaninfo.h:240:vkEnumerateInstanceExtensionProperties failed with ERROR_INITIALIZATION_FAILED
tc@box:~$ export XDG_RUNTIME_DIR=/TMP
tc@box:~$ vulkaninfo
ERROR at /usr/src/Vulkan-Tools-1.2.141/vulkaninfo/vulkaninfo.h:240:vkEnumerateInstanceExtensionProperties failed with ERROR_INITIALIZATION_FAILED
-
Hi Paul_123 or Juanito
found a missing dep with change to udev
epiphany &
tc@box:~/1builds$ epiphany: error while loading shared libraries: libgudev-1.0.so.0: cannot open shared object file: No such file or directory
[1]+ Done(127) epiphany
tc@box:~/1builds$ tce-load -i gudev-lib
gudev-lib.tcz: OK
tc@box:~/1builds$ epiphany & # loads as expected
-
arm 6/7*
mtpaint which is my baby needs an update due to libffi5 EEK
mtpaint
mtpaint: error while loading shared libraries: libtiff.so.5: cannot open shared object file: No such file or directory
I was going to show a screenshot that firefox nightly is not displaying some web pages correctly while epiphany is still good.
Anyone not using headless setup can you confirm please, otherwise I might be running with too few fonts?
https://github.com/lanoxx/tilda
at some later date.....I am tempted to look at a decent web browser for my needs too
EDIT
I forgot to mention that I solved one font issue before by unticking "allow pages to choose their own fonts"
in advanced settings for firefox.....this does not work for me at above site
and quite a few other github gitlab type sites
YMMV
-
at some later date.....I am tempted to look at a decent web browser for my needs too
On x86_64 there is brave-browser.tcz. It is FOSS, based on chromium, and comes with many privacy-friendly features (e.g., anti-fingerprinting) enabled by default. I've been happy with it.
I was going to show a screenshot that firefox nightly is not displaying some web pages correctly while epiphany is still good.
Is your issue that some webpages show text with huge spaces between letters and digits, like this?
(https://forum.tinycorelinux.net/index.php?action=dlattach;topic=26307.0;attach=6606)
When using Brave in TCL, solution was "Settings" -> "Appearance" -> "Customize fonts", then to change every listed font type from "Custom" to one of the fonts in the pull-down menu (which are the fonts actually available on your system). End result:
(https://forum.tinycorelinux.net/index.php?action=dlattach;topic=26307.0;attach=6608)
-
Hi GNUser
the image is supposed to show text to the right of data, docs etc
https://imgur.com/aqr3OUp
I looked at what fonts I have loaded, and available in piCore-14.x. I will need to add some.
However I forced a terminal font onto FF....reload shows a thinner font....so the font was accepted but still no joy on the missing text
2) I am pretty sure I will need to compile a web browser. This can wait until we have some updated packages. TC64 can use FF and I think brave from unpack without packaging.....that does not work on arm.
It might be possible to convert arm debian package to TC....which is cheating....something I am highly skilled at ;D
cheers
-
WOO HOO
on arm32 I have converted a debian chromium and it displays my tricky test web page
I started a new post on this not to pollute TCB post any further
http://forum.tinycorelinux.net/index.php/topic,26322.0.html
-
the image is supposed to show text to the right of data, docs etc
https://imgur.com/aqr3OUp
The Javascript on the page didn't load, or is still loading. That should be fixable, but it's too hard to understand exactly what you're doing/running so as to give further advice.
-
Hi CNK
my reply is here...to reduce my waffle on a TCB post
http://forum.tinycorelinux.net/index.php/topic,26322.0.html
-
Hi Paul_123
arm 7 Maybe alsa-utils needs alsa as a dep?
tce-load -i alsa-utils
alsa-utils.tcz: OK
tc@box:~$ alsamixer
ALSA lib conf.c:4556:(snd_config_update_r) Cannot access file /usr/local/share/alsa/alsa.conf
ALSA lib control.c:1529:(snd_ctl_open_noupdate) Invalid CTL default
cannot open mixer: No such file or directory
tce-load -i alsa
alsamixer # works as expected
-
hi Paul_123
There is a difference in your kernel lines to enable/disable headphones and hdmi.
For reasons I can not explain....the kernel line that works for arm8 piCore-14.x reads in part
snd_bcm2835.enable_headphones=0 snd_bcm2835.enable_headphones=1
snd_bcm2835.enable_hdmi=1 snd_bcm2835.enable_hdmi=0
I am now confused by above anyhow the main thing is you have enable headphones disabled hdmi and
aplay -l
**** List of PLAYBACK Hardware Devices ****
card 0: Headphones [bcm2835 Headphones], device 0: bcm2835 Headphones [bcm2835 Headphones]
Subdevices: 8/8
Subdevice #0: subdevice #0
Subdevice #1: subdevice #1
Subdevice #2: subdevice #2
Subdevice #3: subdevice #3
Subdevice #4: subdevice #4
Subdevice #5: subdevice #5
Subdevice #6: subdevice #6
Subdevice #7: subdevice #7
# no card1
arm7 kernel line similar for headphones....disabled then enabled but
hdmi is enabled and there is no string to disable it.
For those who want a temporary fix and use line out/headphone,
here is my suggestion add to bootlocal.sh after confirming with root powers aplay results in headphones only output
rmmod snd_bcm2835
modprobe snd_bcm2835 options enable_headphones=1 enable_hdmi=0
thanks for reading
-
to anyone on arm6/7*
If you are using headphones/line out you may find no /var/lib/alsa/asound.state to interact with if you run a command
sudo alsactl store
the solution for me, which I hope works for you if interested is
alsamixer # set your volumes
sudo alsactl store -f /home/tc/asound.state
Then I append to my bootlocal as discussed above...one more line
aslactl restore -f /home/tc/asound.state
-
audio options on the kernel command line are set from the pi firmware. You should not be using anything on the command line to manipulate.
Audio is configured in config.txt I do not believe you can enable/disable the headphone jack and HDMI separately. Just pick the appropriate alsa device.
If you are using the VC4 driver, then there are audio options with that too, as the VC4 driver will take over.
-
correction to my above restore info....on reboot I lost the setting so added a sleep and reboot is good again.
so my opt now reads
rmmod snd_bcm2835
modprobe snd_bcm2835 options enable_headphones=1 enable_hdmi=0
sleep 3
alsactl restore -f /home/tc/asound.state
-
Hi Paul_123
I may be misreading your response but I can assure that I have disabled hdmi audio out as per
inxi -Axxx
Audio:
Device-1: bcm2711-hdmi0 driver: N/A bus-ID: N/A chip-ID: brcm:soc
class-ID: hdmi
Device-2: bcm2711-hdmi1 driver: N/A bus-ID: N/A chip-ID: brcm:soc
class-ID: hdmi
API: ALSA v: k6.1.25-piCore-v7l status: kernel-api
tc@box:~$ aplay -l
**** List of PLAYBACK Hardware Devices ****
card 0: Headphones [bcm2835 Headphones], device 0: bcm2835 Headphones [bcm2835 Headphones]
Subdevices: 8/8
Subdevice #0: subdevice #0
Subdevice #1: subdevice #1
Subdevice #2: subdevice #2
Subdevice #3: subdevice #3
Subdevice #4: subdevice #4
Subdevice #5: subdevice #5
Subdevice #6: subdevice #6
Subdevice #7: subdevice #7
inxi is misleading but aplay is not!
We both agree that audio is enabled by the config.txt
at the moment I have bootlist lines
graphics-KERNEL.tcz
Xorg-3d.tcz
#### config.txt lines ...some deviate from original
[PI4]
initramfs rootfs-piCore-14.0.gz,modules-6.1.25-piCore-v7l.gz followkernel
kernel kernel6125v7l.img
hdmi_safe=1
config_hdmi_boost=4
dtparam=audio=on
[PI4]
# Enable DRM VC4 V3D driver on top of the dispmanx display stack
#dtoverlay=vc4-kms-v3d-pi4
max_framebuffers=2
[all]
dtoverlay=vc4-fkms-v3d
thanks for reading
## trivia
if card0 was hdmi I would agree but its not as hdmi is disabled
-
Hi.
Looks like I'm late to the party
Just trying it out on my Pi3 model B starting fresh
Here's what I've run into, I couldn't download KERNEL related files
tc@piCore:~$ . /etc/init.d/tc-functions
tc@piCore:~$ getMirror
tc@piCore:~$ echo $MIRROR
http://repo.tinycorelinux.net/14.x/armv7l/tcz
tc@piCore:~$ echo $BUILD
armv7l
tc@piCore:~$ uname -r
6.1.25-piCore-v7
tc@piCore:~$ uname -m
armv7l
tc@piCore:~$ uname -a
Linux piCore 6.1.25-piCore-v7 #1 SMP Sun Apr 30 13:34:55 EDT 2023 armv7l GNU/Linux
By default, without touching config.txt on Pi3B
You'll load up the armv7 kernel, but 'uname -m' says you have armv7l
This results in tce-load trying to download v7 modules from v7l repo
I think the easy fix for this is to adjust the config file
[PI3]
initramfs rootfs-piCore-14.0.gz,modules-6.1.25-piCore-v7l.gz followkernel
kernel kernel6125v7l.img
#initramfs rootfs-piCore-14.0.gz,modules-6.1.25-piCore-v7.gz followkernel
#kernel kernel6125v7.imgCould anyone kindly share some insight to which kernel I should pick for each type of board ?
P.S. the openbox.tcz.dep present on both armv6 and armv7 need a little fix
libpcre.tcz --> pcre.tcz
BTW, there's no openbox on armv7l repo, it's working though.
-
A pi3 is armv7.
Pi4 in 32bit uses armv7l
I’ll have to take a look at logic around the Mirror setting.
-
Hi
Here's the difference between the two
--- tc-functions-13
+++ /etc/init.d/tc-functions
@@ -273,18 +273,23 @@
} ' "$1"
}
+getFullVer() {
+verid=$(grep '^VERSION_ID=' /etc/os-release)
+echo ${verid#*=}
+}
+
getMajorVer() {
-awk '{printf "%d", $1}' /usr/share/doc/tc/release.txt
+fullver=$(getFullVer)
+echo ${fullver%.*}
}
-
getBuild() {
-BUILD=`getrpicpu`
+BUILD=`uname -m`
case ${BUILD} in
- BCM2835) echo "armv6" ;;
- BCM2836) echo "armv7" ;;
- BCM2837) echo "armv7" ;;
- BCM2711) echo "armv7l" ;;
+ armv6l) echo "armv6" ;;
+ armv7) echo "armv7" ;;
+ armv7l) echo "armv7l" ;;
+ aarch64) echo "aarch64" ;;
*) echo "armv6" ;;
esac
}
-
Yes, I already fixed it locally. That’s what happens when git doesn’t get updated. I found lots of cases where Core scripts on piCore did not match, I just missed that one.
I’ll push up a dot release image soon.
-
New 32bit image uploaded. http://www.tinycorelinux.net/14.x/armv7/releases/RPi/piCore-14.0.1.zip (http://www.tinycorelinux.net/14.x/armv7/releases/RPi/piCore-14.0.1.zip)
-fixes Pi2/3 using wrong repo.
No other changes.