Arch Linux User Repository

Canon also supports this driver on i686 and MIPS64 .

IF

you run archlinux (or something close to it) on such a processor

AND have a supported printer

AND are willing to help test necessary adjustments

THEN please send me an email.

The printers supported by this package are often networked.

problems can be in the cnrdrvcups-lb driver, but also with authentication over smb, vpn settings etc .

Troubleshooting those is often very hard.

Archlinux and derivatives are not supported by canon. Use whatever works for you, even if that means using canon drivers in a VM that runs a supported distro .

I figured out the solution to the 100% CPU bug. There's an intermediate process that attempts to dynamically load the libjbig.so.0 shared library - the only problem is, that library doesn't exist in the jbigkit package.

In fact, that library is completely made up by the Debian maintainers (see https://github.com/mvanderkolff/jbigkit-packaging/tree/master/debian/patches for a list of the patches they add).

I manually built the shared library by applying the allNewMainMakefile.diff, pbmtoolsMakefile.diff, and shared-lib.diff patches to the jbigkit source tree and copied the resulting libjbig.so.0 library to /usr/lib. Then I restarted CUPS and was able to print a test page.

Those lines appear to come from the package() function, but I'll need more info to investigate.

You do have base-devel group installed ?

run LC_ALL=C makepkg --log to get output in english and post the logs somewhere publicly accessible

(if you don't know how to do that, check https://wiki.archlinux.org/index.php/List_of_applications#Pastebin_clients )

I Makepkg install, it print some error,(sry, my chinese environment): target=install; for dir in cngplp buftool backend rasterfilter cnjbig; do (cd $dir; make $target)|| exit 1; done make[1]: 警告: jobserver 不可用: 正使用 -j1。添加 “+” 到父 make 的规则。 make[1]: 进入目录“/home/snail/Aur/cnrdrvcups-lb/src/extracted-cnrdrvcups-lb-5.20/cnrdrvcups-common-5.20/cngplp” make[1]: 没有规则可制作目标“install”。 停止。 make[1]: 离开目录“/home/snail/Aur/cnrdrvcups-lb/src/extracted-cnrdrvcups-lb-5.20/cnrdrvcups-common-5.20/cngplp” make: [Makefile:18：install] 错误 1

It seems like the first service(cngplp) can't install, then the for script exit, this problem makes cups to miss rasterfilter directory, the system log print:

File \"/usr/lib/cups/filter/rastertoufr2\" not available: No such file or directory

I have virtually the same problem as redeyed. I use a CANON imageCLASS MF229dw via USB. I add the printer using the CUPS web administration page. When printing it's clear the laptop communicates with the printer and there's blinking lights but printer just hangs.

I'm migrating from another laptop that uses cndrvcups-lb which prints using the same printer. In this laptop I can't even downgrade to cndrvcups-lb since I have a build src error when installing the dependency cndrvcups-common-lb... I also tried cnrdrvcups-lb-bin. I think it gave me the same problem.

Let me know if you have suggestions to debug.

edit: my printer is supported UFR II v5.10 on Ubuntu.

I have not tried that; that's a good idea. I don't have a box with such a distro handy; I'll set one up and try this weekend. If that doesn't work maybe I'll try to get in touch with Canon about it.

Canon does mention the MF632C as being supported by this version on the README-ufr2-5.1xUK.html .

Have you tried this version on canon-supported linux distro like fedora or ubuntu to verify if this version does work with your printer ?

I have a Canon imageCLASS MF632Cdw. Unfortunately, I was not able to get it to print using this package; I had to fall back to the deprecated cndrvcups-lb. Even after installation of libjpeg6-turbo (which solved a problem wherein cnpkmoduleufr2r would consume 100% CPU), the printer never actually printed anything, and instead waited forever for the computer to send the file to be printed.

I did not find cndrvcups-lb to be in a building state--some symbols were multiply defined in cgnplp. Specifically, load.h defines some function pointers; that file is then included in mainwnd.c and load.c. Objects from both translation units are then linked into the cgnplp executable, which causes the function pointers to be multiply defined. I fixed it by declaring them external in load.h and defining them in load.c.

hnws, sounds like you managed to make the changes needed to build with gcc 10?

If so, please email me your PKGBUILD (address is in the PKGBUILD).

As now gcc upgraded to 10.1, we need -fcommon to compile this package. And for this to happen, we need -e 's:${_cflags}:"${_cflags}":g' \ to quote the cflags when generating the make.Arch.

With all these being done, I am still struggling with the 100% cpu usage issue. The code is very simple, it's pending a read from a file descriptor and it's reading non-stop. But I am not sure who should pass data to it.

May not be related to all issues we have with Canon drivers, but I fixed my printing problem (and 100% one core CPU usage) by creating the file /etc/samba/smb.conf. Of course, my printer is shared via Samba.