http://foo2zjs.rkkda.com/ foo2zjs is an open source printer driver for printers that use the Zenographics ZjStream wire protocol for their print data, such as the Minolta/QMS magicolor 2300 DL, HP Laserjet 1000/1005 etc. Latest Version: n/a (20041109), see http://foo2zjs.rkkda.com/ChangeLog
Created attachment 40264 [details] app-misc/foo2zjs-20042309 seems to dislike 64 Bit compilers, it compiles but it doesn't work. I hope it's a good decision do append -m32 to CFLAGS for amd64 here... Use the local use flag 'getweb' to instruct the shipped getweb script to get additional files (firmware, icm color profiles, some ppd files) from the web.
Created attachment 40265 [details, diff] 20042309-Makefile.patch
Created attachment 40266 [details] Script to refresh foomatic db
Created attachment 40316 [details] net-print/foo2zjs/foo2zjs-20040923.ebuild Hi, I think I made your ebuild better, added some use, made portage download the getweb files. Basically I merged my old and your ebuild to get the best of both. I left out the foomatic reconfigure because the directory does not exist in my 2 gentoo installations, and the printer apperently works without it, do you need it? Can you please test the ebuild so that we can assure that it works on 64bit?
Add myself as cc
Created attachment 40317 [details] files/foo2zjs-20040923-Makefile.patch
Seems alright to me, works fine on amd64 here. You're most certainly right about this foomatic directory thing, works here without that directory. My experiences so far: (Laserjet 1000) Automatic firmware transferation with hotplug doesn't work here Paper Size is must be set to A4 instead of letter, but does still cut of parts of the things to print - most certainly a configuration issue, how could I fix this? Still prints the 'Media Tray empty' warning, but prints though.
laserjet 1005 here - hotplug works. Is your printer on usb, what does the logfile tell us? Did you install with USE=getweb to get the firmware? Hotplug should work .. its a needed feature.
It works on my 2.4 kernel server - on the 2.6 client I cant do hotplug .. :( any clue why it does not work?
Now I know what's wrong: Firstly, usb_printerid.c doesn't like being compiled with -m32, so I figured out a (dirty?!) workaround to fix this for amd64. Secondly, hotplug seems to run /etc/hotplug/usb/hplj1000 _before_ udev has finished creating the device nodes. I inserted a sleep 20 into the script, which is executed if it detects udev. This dirty hack is sufficient for me, but shouldn't hotplug call this script _after_ a successful udev run?
Created attachment 40604 [details] foo2zjs-20040923-r2
Created attachment 40605 [details, diff] foo2zjs-20040923-amd64.patch makes udev_printerid.c work as it should on amd64
Created attachment 40606 [details, diff] hp-printer-udev.patch inserts sleep into hplj1000 hotplug script
Created attachment 40883 [details] files/hp-printer-udev.patch An updated hotplug patch that loads the firmware when the device is there and does not waste ~7 sec before.
We still need to get the hotplug script working. But maybe 64bit works now: 2004-10-23 Rick Richardson <rickr@mn.rr.com> * Make code 64-bit clean for x86_64 (but not well tested). Can you please check and report?
Created attachment 42700 [details] foo2zjs-20041023.ebuild version bump, removed amd64 specific stuff, works without this now. hotplug script works too. Shouldn't the ebuild install the hotplug stuff only if the use flasg 'usb' is set?
Created attachment 44146 [details, diff] foo2zjs-20041117-Makefile.patch net-print: Can I take that bug and add foo2zjs to portage?
Created attachment 44147 [details, diff] foo2zjs-20041117.ebuild New ebuild with cleaned up patch, moved fixes from the ebuild to the patch.
yes, go ahead
Fixed in cvs. Thanks.