SliTaz Forum
You are not logged in.
Hi gibor,
At first, thanks for your test and your compliments.
Nice to see that current work as well for this usage, hope continue to improve day to day usage 
For firefox, infact i don't build it directly, but firefox-official is available.
Strange you have black screen, maybe miss firmwares 
.
I have a same issue when i try first time SliTaz Current on my laptop, miss firmware-amdgpu to have screen.
Offline
Hi, although it is of little consequence I am getting an error that I cannot trace back to the cause:
“CreateProfile failed: org.freedesktop.DBus.Error.ServiceUnknown:The name org.freedesktop.ColorManager was not provided by any .service files”
The printer that fails is an old grayscale laserjet2100, to which I associate the ppd I find on the net. The ones in hplip don't go, or if it starts they glitch in printing.
Maybe it could be that some packages are compiled without the necessary dependencies?
--------------------------------------------------------------------------
Ciao, anche se ha poca importanza mi ritrovo un errore che non riesco a risalire alla causa:
“CreateProfile failed: org.freedesktop.DBus.Error.ServiceUnknown:The name org.freedesktop.ColorManager was not provided by any .service files”
La stampante che non va è una vecchia laserjet2100 in scala di grigi, a cui associo il ppd che trovo in rete. Quelli presenti in hplip non vanno, o se parte fanno difetti nella stampa.
Magari potrebbe essere che qualche pacchetto è compilato senza le dipendenze necessarie?
Offline
Hi gibor,
If understand i found it's related to colord, don't packaging, not sure if cause issue for your printer. I found this link on gentoo forum that report this issue
https://forums.gentoo.org/viewtopic-t-1138212-start-0.html
Maybe need to build it, you said with laserjet2100 ppd file on hplip not work ? What version do you find on net ? hplip on current it's 3.20.5
Offline
Hi Shann, I understand it's strange, but cups long ago that print warns that it will no longer be supported. Also added complication is parallel port seen with a usb adapter.
To sum up: I only print pdf wirelessly, clients are always with the same setting.
On the server hplip drivers with cups-2.3.3 on your test build does not print; or at least prints a small portion at the top/ or writes something in text with the blank sheet then calls in an infinite loop new page.
On my build with cups-2.3.3 the printer works with the configuration
HP LaserJet 2100M - CUPS+Gutenprint v5.3.4
The old working server based on slitaz-3.0 and cups-1.4.0 had configuration
HP LaserJet 2100M - CUPS+Gutenprint v5.2.7
Colord I tried to compile it on your version but it doesn't come out. I don't remember where, but it doesn't have everything you need to do it.
Sorry again for going on and on... cups for me is strange, each printing process invokes often different sub processes. So I have written many words but few on the substance of the codes.
Translated with DeepL.com (free version)
Offline
Hmm if understand
On current
cups 2.4.7 / gutenprint 5.3.4 / hplip 3.20.5 failed and same if you build cups 2.3.3
On your build
cups 2.3.3 / gutenprint 5.3.4 / hplip (?) ok
On 3.0
cups 1.40 / gutenprint 5.2.7 / hplip 3.9.10 ok
Maybe hplip version is too new to work with your printer.
On cooking it's work cups 2.4.1 / gutenprint 5.3.4 / hplip 3.22.2 ?
Offline
With rolling it automatically comes out a blank page, then it asks for confirmation on the printer and the cut page comes out as in the image.
If it had printer problems, with the old server it would not print the correct sheet....
----------------------------------------------------------------------
Con rolling esce in automatico una pagina bianca, poi richede conferma sulla stampante ed esce la pagina tagliata come nell'immagine.
Se avesse problemi la stampante, con il vecchio server non stamperebbe il foglio corretto...
Rolling
[attachment=53229,3717]
slitaz-3.0
[attachment=53229,3718]
Offline
Ok thanks
For your own build, which hplip version you use or it not important ?
Infact i understand rolling not work in case cups 2.4.x and hplip 3.22.
In case we need cups 2.3.3 with gutenprint 5.3.4 and with hplip same version you used on your build.
Offline
By enabling cups debugging and observing the printing process, it can be seen that hplip is never called, while cups-filter and gutenprint routines participate directly.
I ran two tests with the two servers and the same PDF from the same client location. I collected the two logs and posted them on pastebin:
https://pastebin.com/GEetrn7F << mybuild(slitaz-10.64bit)
https://pastebin.com/QeyQmkRS << oldserver(slitaz-3.0)
I will try to recompile the dependencies of cups-lib, poppler, cups, gutenprint, without installing anything else related to the slitaz packages on printing.
Translated with DeepL.com (free version)
Offline
Hi gibor,
If understand for your build you use cups 2.3.3 / gutenprint 5.3.4.
I don't found cups-filters version, not exist on SliTaz 3.0/4.0 version
slitaz-10.64bit mean you use LFS 10 as base ?
In case i see cups 2.3.3 with cups-filters 1.27.5 and gutenprint 5.3.3 (maybe you have bump it to 5.3.4)
Offline
Hi Shann, sorry but I have been compiling freewheeling sources for a long time, and when I say build10 it is LFS-10 with base files slitaz-3.0.
Since I have no real repository, I use an external hard drive where I compile the sources. Because of this without rules I often make a soup of packages.
Going back to cups is treacherous, what you read in the driver is often duplicate, or not matching what you use. You need to look at the version of the PPD inside the file in /etc/cups/ppd with the name print, only there is certainty about the match indicated in the test page.
Translated with DeepL.com (free version)
*FormatVersion: "4.3"
*FileVersion: "5.3.4"
*LanguageVersion: English
*LanguageEncoding: ISOLatin1
*PCFileName: "STP02296.PPD"
*Manufacturer: "HP"
*Product: "(HP LaserJet 2100M)"
*ModelName: "HP LaserJet 2100M"
*ShortNickName: "HP LaserJet 2100M"
*NickName: "HP LaserJet 2100M - CUPS+Gutenprint v5.3.4"
*PSVersion: "(3010.000) 0"
Offline
Epilogue on the LaserJet 2100 printer that only worked with the old server:
I don't know how or why, but the malfunction was caused by a simple 4-port USB hub.
By pure chance, I realized that this was the cause, because I connected the external hard drives to another hub with a separate power supply. And in that case, the printer worked correctly.
This also explains why the driver changed on its own when I connected the old hub...
I don't know the reasons, but I'll frame this one nicely. :-D
Translated with DeepL.com (free version)
------------------------------------------------------------------------
Epilogo sulla stampante laserjet-2100 che funzionava solo con il vecchio server:
Non so in che modo, e per quale motivo, ma a causare il malfunzionamento era un semplice hub a 4 porte usb
Per puro caso mi sono accorto che la causa era lui, perché gli hard disk esterni li collegavo con un altro hub con alimentazione separata. Ed in quel caso la stampante funzionava correttamente.
Da qui pure la causa del driver che cambiava da solo quando mettevo il vecchio hub,,,
I motivi non li so, ma questa me la metterò per bene in cornice. :-D
Offline
Hi gibor,
i search stp02296.ppd file and didn't found on cups and related package on current 
.
This could be explain issue happen.
If undertstand your last message, related to usb hub ?
It's work on both rolling / current now ?
Offline
Hi Shann, I'll try to explain myself better, even though I don't need it anymore: the hub is only connected to the tablet, which has only one USB port to which I have to connect a mouse, keyboard, and 4 printers and scanners.
The one I normally used had that problem. Using a newer one, it doesn't.
The difference was also noticeable when sending a document to print, which was much slower with the old one, and at the end, the printer required me to press the print button (manual loading).
Unfortunately, while playing around with the settings in the tablet's BIOS, I disabled xhci_hcd in the hope of improving the situation. Instead, I find myself without USB ports, with no way to reactivate them.
https://xdaforums.com/t/xhci-disabled-on-windows10-tablet-usb-ports-doesnt-work.3619759/
For the stp02296.ppd file, it is cupsd in combination with the other utilities (gutenprint, ghostscript hplip, etc.) that generates it independently. If you look in the “configure” compilation, you will see the option to generate them all directly, or leave the option to do so only directly when the specific driver is requested.
Translated with DeepL.com (free version)
Offline
Hi gibor,
if understand it's issue with your hub, work with new hub.
Now you don't have usb port due of change in BIOS you can't revert .
Don't know configure have option to generate ppd on the fly when required, interesting.
Offline
Yes, that's right, it works with a different hub... or rather, it used to work, because now the tablet is lost forever
For Cups, try installing a printer and you will find a file called “printer.ppd” in /etc/cups/ppd/ that didn't exist anywhere before.
Offline
Hi,
Sorry for delays, i finish to rebuild all packages on my local chroot for current.
I build core and base iso, test in progress to ensure same as current64 all work as expected before push isos to my space.
For people want to upgrade current to new stack (glibc 2.34 / gcc-lib-base 10.2.0), can be change /var/lib/tazpkg/mirror to point on [c]https://people.slitaz.org/~shann/slitaz-current-stuff/packages/[/c] instead [c]https://mirror.slitaz.org/packages/current/[/c]
Due of tank stack, i can't build new stack in current chroot, need to use my local env that i use for both current/current64 chroot.
Offline
Hi,
Maybe useless, but i begin to rework tazpkg install feature, goal it's to avoid copy package twice (from cache to tmp folder), and same for copy fs twice (from tmp to rootfs).
At first test, i remove copy package to tmp, and rework fs.cpio.lzma to not uncompress it and extract directly on /, but using [c]|[/c] is not good, it's use memory as cache, test for rust package on my dell x300 (1Go of memory) , it's use ~700Mo memory during extract process, it's not great for less memory machine.
To achieve it, i do this :
- Use same process as slackware package, tarball compressed with xz instead of cpio.lzma, compress ratio is best or same as lzma (implied tar/xz package on cook chroot but for install, busybox tar is enought, also i use [c]-9[/c] for xz compression)
Test on dell x300 (pentium M 1.2Ghz), same time ~40s for extract fs.tar.xz / fs.cpio.lzma
- With first point, this implied to rework "CONFIG_FILES" feature, infact we extract on rootfs directly, need to deal with config files on /etc, i use same feature used on slackware package with dotnew config.
=> In case we copy config file with suffix .new, if file not exist or not modified we move it as original name else we keep it with .new (need to create feature to manage it).
Build in progress on my laptop to test it, after test on x86_64, do same for x86 version.
Edit: mingw32-gcc failed to be build, after investigation, interesting think, seem directly related to gnu tar version installed on chroot.
On receipt we create symlink for gcc-$VERSION before extract gcc-ada, gcc-g++, gcc-$VERSION symlink replaced by folder when tar extract.
I need to use "busybox tar" to avoid this.
Offline
Hi,
I finish to rebuild dev stuff and base/base64.
I doing quick test to create base chroot, minus rework config_dotnew, wrong place [c]for loop[/c] when check if config_files is folder ^^, seem work as expected.
I install package from mirror [c]nano[/c], work well, infact it's backward compat (it's necessary to rebuild all packages without break due of all deps not rebuild yet)
I compare packages builded with tar.xz vs lzma, seem we gain few Mo 
:
root@arion:~# du -csh /home/slitaz/packages
568.4M /home/slitaz/packages
568.4M total
root@arion:~# du -csh packages_lzma
625.0M packages_lzma
625.0M total
root@arion:~# du -csh packages_lzma/gcc*
34.2M packages_lzma/gcc-10.2.0-x86_64.tazpkg
364.0K packages_lzma/gcc-lib-base-10.2.0-x86_64.tazpkg
124.0K packages_lzma/gcc-lib-math-10.2.0-x86_64.tazpkg
15.4M packages_lzma/gcc49-4.9.2-x86_64.tazpkg
224.0K packages_lzma/gcc49-lib-base-4.9.2-x86_64.tazpkg
50.4M total
root@arion:~# du -csh /home/slitaz/packages/gcc*
16.4M /home/slitaz/packages/gcc-10.2.0-x86_64.tazpkg
364.0K /home/slitaz/packages/gcc-lib-base-10.2.0-x86_64.tazpkg
124.0K /home/slitaz/packages/gcc-lib-math-10.2.0-x86_64.tazpkg
9.9M /home/slitaz/packages/gcc49-4.9.2-x86_64.tazpkg
224.0K /home/slitaz/packages/gcc49-lib-base-4.9.2-x86_64.tazpkg
26.9M total
Offline
Hi,
I finish to rebuild all x86_64 packages, doing test with base64 and core64 iso.
I doing tests with old iso for migration from fs.cpio.lzma > fs.tar.xz, found big side effect.
Infact when we use [c]tazpkg up -r[/c], if tazpkg updated, we stop it and ask to reinstall tazpkg first, it's good think but due of change tazpkg can't be correctly upgraded, i add step in pre_install function.
Also another point it's on remove files function, infact before transition to dotnew, files that now in configs_files removed and replace by new (/etc/default/grub, /etc/slitaz/tazpkg.conf for example), i tweak it to add condition to exclude file in "$CONFIG_FILES"
Continue to doing tests to ensure no other side effects
Offline
On current64, please add the zstd library to the dependencies of slitaz-toolchain. Otherwise, gcc will not work.
Offline
Hi gibor,
nice catch, thanks for report, sorry for issue .
I check on my chroot and see that zstd already installed, don't see this issue before and think why.
[c]cook setup[/c] install rsync that have zstd on depends.
I update receipt, rebuild of slitaz-toolchain in progress.
Offline
Hi,
After tests seem good and don't see regression, i switch /packages-ng (that use fs.tar.xz) to /packages for 6.0 x64. If you do upgrade you switch to new package archive.
In same time i adjust slitaz-release, now current will be know as "6.0".
I'm doing rebuild for 6.0 32bits version with new package archive.
Offline
Hi Shann,
I'm having trouble: how do I update tazpkg on a standard installation of current64 so that it installs the new xz format?
Offline
Hi Gibor,
Sorry for inconvenience, you need to do [c]tazpkg -gi tazpkg --forced[/c] for use new xz format, if you do [c]tazpkg up -r[/c], it's purpose automatically to upgrade tazpkg first.
Offline
Thanks Shann,
The installation works permanently offline, automatic updates do not work.
So I find myself with installed packages that have failed as regularly installed. Which they are not. So tazpkg automatically cannot understand which ones are corrupted and which ones are not.
I pull up a backup and try again, first updating tazpkg.
Offline
[ Generated in 0.034 seconds, 7 queries executed - Memory usage: 1.6 MiB (Peak: 1.77 MiB) ]