[SOLVED] Scanner Problem after update

BrunoLafleur · 02-25-2021, 02:31 PM

Quote:

Originally Posted by business_kid

Now you know why I keep isos

. I have qt-4.8.7 from the 2020-09-28 iso. That went in, but didn't help much. I think it may have got further along, but I'm not sure

Code:

Traceback (most recent call last):
  File "/usr/bin/hp-doctor", line 297, in <module>
    num_errors, num_warns = dep.validate(DEPENDENCY_RUN_AND_COMPILE_TIME, False)
  File "/usr/share/hplip/check.py", line 367, in validate
    self.__update_deps_info(supported_distro_vrs, dep,
  File "/usr/share/hplip/check.py", line 210, in __update_deps_info
    installed_ver = self.core.version_func[deps_info[6]]()
  File "/usr/share/hplip/installer/dcheck.py", line 303, in get_pyQt4_version
    from PyQt4 import QtCore
ImportError: cannot import name 'QtCore' from 'PyQt4' (unknown location)

I don't have the machine for big compiles. I don't need to fix qt, just the scanner.

My hp-doctor from version 3.20.6 is working here with the line :
python3-pyqt4-dbus PyQt 4 DBus - DBus Support for PyQt4 OPTIONAL 4.0 4.12.3 OK -
python3-pyqt4 PyQt 4- Qt interface for Python (for Qt version 4.x) REQUIRED 4.0 4.12.3 OK -

PyQt4 should work. I depends on qt4 :
ldd /usr/lib64/python3.9/site-packages/PyQt4/QtCore.so
linux-vdso.so.1 (0x00007ffd20d56000)
libQtCore.so.4 => /usr/lib64/libQtCore.so.4 (0x00007f3ed8985000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f3ed8964000)
libstdc++.so.6 => /usr/lib64/libstdc++.so.6 (0x00007f3ed8797000)
libm.so.6 => /lib64/libm.so.6 (0x00007f3ed864d000)
libgcc_s.so.1 => /usr/lib64/libgcc_s.so.1 (0x00007f3ed8633000)
libc.so.6 => /lib64/libc.so.6 (0x00007f3ed8454000)
libz.so.1 => /lib64/libz.so.1 (0x00007f3ed8438000)
libdl.so.2 => /lib64/libdl.so.2 (0x00007f3ed8433000)
libgthread-2.0.so.0 => /usr/lib64/libgthread-2.0.so.0 (0x00007f3ed842e000)
libglib-2.0.so.0 => /usr/lib64/libglib-2.0.so.0 (0x00007f3ed8300000)
librt.so.1 => /lib64/librt.so.1 (0x00007f3ed82f6000)
/lib64/ld-linux-x86-64.so.2 (0x00007f3ed8f53000)
libpcre.so.1 => /lib64/libpcre.so.1 (0x00007f3ed827c000)

But if next version is with qt5, it will be better.

Emerson · 02-25-2021, 03:53 PM

These are two only references to Qt when I run hp-doctor.

Code:

python3-pyqt5-dbus   PyQt 5 DBus - DBus Support for PyQt5                         OPTIONAL        5.0             5.15.2          OK         -
python3-pyqt5        PyQt 5- Qt interface for Python (for Qt version 4.x)         REQUIRED        5.0             5.15.2          OK         -

business_kid · 02-26-2021, 03:39 AM

OK. Nothing called python3-pyqt<anything> is on my iso. I did a full iso install, except for kde, & testing. I took a look on bear.alienbase.nl, but there's no newer version of hplip. The git just seems to have diffs from that. I don't have time for this now as I have an appointment this morning. I may get to it later.

Getting hp-doctor singing doesn't get me a scanner, because all I expect is a repeat of the I/O error. Do you expect anything else?

BrunoLafleur · 02-26-2021, 05:43 AM

Quote:

Originally Posted by business_kid

OK. Nothing called python3-pyqt<anything> is on my iso. I did a full iso install, except for kde, & testing. I took a look on bear.alienbase.nl, but there's no newer version of hplip. The git just seems to have diffs from that. I don't have time for this now as I have an appointment this morning. I may get to it later.

Getting hp-doctor singing doesn't get me a scanner, because all I expect is a repeat of the I/O error. Do you expect anything else?

I think it was just to try to have a diagnostic on your scanner. But if hp-doctor don't work, it is probably not necessary to make it work if you don't need it.

Do you have some .new files or new configs files not adapted in /etc/sane.d directory or in HOME ?

I don't usually touch those files but maybe with your devices there is some problem in particular for network devices.

See for example : https://help.ubuntu.com/community/sa...r_Installation

business_kid · 02-26-2021, 10:32 AM

We appear to have narrowed the difference between your Qt stuff and mine.

I thought of reinstalling the old system on a usb disk, but there's no point because it works. PyQt4 & PyQt5 are represented in my /usr/lib64, and each have libs called QtCore.so /usr/share/sip/ has the bulk of QtCore. /usr/bin/hp-doctor is a link to a python script. I ran it with python2, and it puked and told me to reinstall hplip. It needs python3 to run. It apparently has needed python3 for some time, as my ~Current from last September needed python3 on the doctor.py script (just tried it).

The saned.conf has my scanner IP configured in it and always worked. I fixed it in the router to get a steady IP. But I wasn't running saned (never used it, IIRC). So I started it with debug & exit, then ran xsane

Code:

/usr/sbin/saned -d128 -o

The output was

Code:

Feb 26 16:15:01 RoseViolet xsane: io/hpmud/model.c 532: no laserjet_mfp_m129-m134 attributes found in /usr/share/hplip/data/models/models.dat 
Feb 26 16:15:01 RoseViolet xsane: io/hpmud/model.c 543: no laserjet_mfp_m129-m134 attributes found in /usr/share/hplip/data/models/unreleased/unreleased.dat 
Feb 26 16:15:08 RoseViolet xsane: common/utils.c 245: unable to load library libm.so: /usr/lib64/libm.so: invalid ELF header 
Feb 26 16:15:08 RoseViolet xsane: common/utils.c 112: unable to open /var/lib/hp/hplip.state: No such file or directory 
Feb 26 16:15:08 RoseViolet xsane: common/utils.c 162: validate_plugin_version() Failed to get Plugin version from [/var/lib/hp/hplip.state] 
Feb 26 16:15:08 RoseViolet xsane: common/utils.c 206: Plugin version is not matching  
Feb 26 16:15:08 RoseViolet xsane: common/utils.c 277: Invalid Library hanlder pLibHandler = NULL.

That libm.so is a static lib from one of the combination glibc multilib packages, but that puts me on 'a definite line of enquiry' as the police here say (when they know who did it, they just need proof). I'll go exploding packages and try and kill this today. Having a static lib labelled as a dynamic one smells like a package error, but judgement would be premature. There's just been an update here.

business_kid · 02-26-2021, 11:03 AM

Ok. I have compared the /usr/lib64/libm.so from glibc-2.32 package from 2021-02-11, and the glibc-2.33_multilib package. Both are the same 110 byte ascii text file, calling (as needed) /lib64/libm.so.6 --> /lib64/libm-2.33.so, & /libmvec.so.1 -->/lib64/libmvec-2.33.so. These final destinations are elf64 executable libs, but I'm left with the error in my last post

EDIT: readelf -h throws up nothing suspicious to my eye.

Emerson · 02-26-2021, 12:47 PM

If your scanner needs binary plugin then you have to install it, it is not part of hplip package.

Emerson · 02-26-2021, 01:06 PM

Code:

hp-align               hp-diagnose_queues     hp-linefeedcal         hp-print               hp-systray
hp-check               hp-doctor              hp-logcapture          hp-printsettings       hp-testpage
hp-clean               hp-fab                 hp-makecopies          hp-probe               hp-timedate
hp-colorcal            hp-faxsetup            hp-makeuri             hp-query               hp-toolbox
hp-config_usb_printer  hp-firmware            hp-pkservice           hp-scan                hp-uiscan
hp-devicesettings      hp-info                hp-plugin              hp-sendfax             hp-unload
hp-diagnose_plugin     hp-levels              hp-pqdiag              hp-setup               hp-wificonfig

All this is in hplip package, have you tried to run hp-setup? It should detect if binary blob is needed, download and install it for you.

business_kid · 02-26-2021, 02:00 PM

AFAIK, it just works with no plugin.

hp-setup just installs the printer bit

hp-scan gave me

Code:

-----------------
| SELECT DEVICE |
-----------------

  Num       Scan device URI                                        
  --------  -------------------------------------------------------
  0         hpaio:/net/HP_LaserJet_MFP_M129-M134?ip=192.168.178.103
  1         v4l:/dev/video0                                        

Enter number 0...1 for device (q=quit) ?0
warning: No destinations specified. Adding 'file' destination by default.
Using device hpaio:/net/HP_LaserJet_MFP_M129-M134?ip=192.168.178.103
Opening connection to device...
error: SANE: Error during device I/O (code=9)

There's a gnu type library error in post #21, and my efforts to chase it in post#22. That's the one firm fault I can see. The I/O error appears to be linked to that (post #21). There's an error if I can get teeth into it. BUT (and this may be relevant) the system came with glibc-2.32, but the Compat32 stuff (which reinstalls glibc in 64+32 form) was built on glibc-2.33. Am I being punished?

gauchao · 02-26-2021, 02:03 PM

I have the HP MFP 135a multifunctional connected via USB to my Slack64-current up-to-date PC. Which driver version have you installed? I run this machine with a HP basic driver, V1.00.39:00.12_00.15, from HP site:

https://support.hp.com/us-en/drivers...model/24494379

Everything works fine, but every kernel upgrade I have to reinstall the scan drivers with the provided script. Not an issue at all for me.

Is your MFP130 fully supported by HPLIP or do you need a basic driver?

Emerson · 02-26-2021, 02:09 PM

If 2.33 is the system glibc then there should be no problem as glibc is backwards compatible, OTOH anything built against newer glibc will break if system glibc is older ...

BrunoLafleur · 02-26-2021, 02:59 PM

Quote:

Originally Posted by business_kid

There's a gnu type library error in post #21, and my efforts to chase it in post#22. That's the one firm fault I can see. The I/O error appears to be linked to that (post #21). There's an error if I can get teeth into it. BUT (and this may be relevant) the system came with glibc-2.32, but the Compat32 stuff (which reinstalls glibc in 64+32 form) was built on glibc-2.33. Am I being punished?

The last Slackware current is with glibc 2.33 since february the 15th and so multilib from Eric also (was compiled some hours after).

Maybe your update is a little broken :

So may be you have some 2.32 part of glibc before 15 (11 as your first post) and 2.33 multilib taken after 15 february for other part and so you have some strange behavior like bad libm.so (lib math from glibc).

BrunoLafleur · 02-26-2021, 03:04 PM

I have also multilib and I download it immediatly after updating my current Slackware.

business_kid · 02-27-2021, 05:50 AM

The libm.so from glibc-2.32 is identical with glibc-2.33. Unless there's been some upheaval in glibc, I would expect it to be backward compatible as Emerson says.

libm.so references 2 libs. ldd /usr/bin/xsane only mentions one, /lib64/libm.so.6 --~> /lib64/libm-2.33.so. I tried a direct link, but that didn't cut it either.

/Wishes to retire to a home for the bewildered, but can't afford it.

BrunoLafleur · 02-27-2021, 07:03 AM

Quote:

Originally Posted by business_kid

The libm.so from glibc-2.32 is identical with glibc-2.33. Unless there's been some upheaval in glibc, I would expect it to be backward compatible as Emerson says.

libm.so references 2 libs. ldd /usr/bin/xsane only mentions one, /lib64/libm.so.6 --~> /lib64/libm-2.33.so. I tried a direct link, but that didn't cut it either.

/Wishes to retire to a home for the bewildered, but can't afford it.

But there seems to be a mismatch between aaa_glibc-solibs, glibc-solibs and glibc multilib or not installs.

glibc 2.33 can work with older binaries but not the contrary and some parts of the libc packages seems to be from 2.32 in your case.