Ok... who thought it was a good thing to break printer accounting in the new lpr? 0.47 in RawHide had the same weirdism, but I figured hey, that's RawHide, and I didn't say anything. But 0.48 has the same flaw. With 0.46 --> Jan 7 21:06:01 rei ifpap[21657]: starting for mwilson Jan 7 21:06:01 rei ifpap[21657]: accounting with psa[21658] Jan 7 21:06:01 rei ifpap[21657]: sending to pap[21659] Jan 7 21:06:01 rei ifpap[21657]: PostScript Jan 7 21:06:50 rei ifpap[21657]: 21659 done Jan 7 21:06:50 rei ifpap[21657]: 21658 done Jan 7 21:06:50 rei ifpap[21657]: done Notice it works, and I get toner on paper... and with 0.48 --> Jan 7 21:23:26 rei lpd: lpd shutdown succeeded Jan 7 21:23:26 rei lpd: lpd startup succeeded Jan 7 21:23:38 rei ifpap[21856]: Too many arguments Jan 7 21:23:38 rei lpd[21855]: laser: job could not be printed (cfA546rei.moonkingdom.net) And now it doesn't. RH6.0 with patches and netatalk-1.4b2+asun2.1.3-6. If accounting only works now with rh-printfilters, you could at least tell someone. Or maybe the calling syntax for the filter has been changed, in which case changing the documentation might be a good idea so someone might have a chance of altering the filter. My /etc/printcap looks like: lp|laser:\ :sd=/var/spool/lpd/lp:\ :af=/var/account/lpacct:\ :mx#0:\ :sh:\ :lp=/dev/null:\ :if=/usr/lib/atalk/filters/ifpap: The printer is a LaserWriter IIg on ethernet... there is no Mac involved.
Apparently ifpap is not recognizing the -j option now being passed to it. I set up a fake printer to dump the arguments to if: -w132 -l66 -i0 -n john -h linus.localdomain -j stdin /var/tmp/acct Comparing this with netatalk/etc/psf/psf.c reveals that -j is not recognized.
Ok, what's the -j option supposed to do? Tell the filter where to get input from?
I assume -j is the jobname as seen in lpq. The filter always reads the standard input.
Try 0.50-2 or 0.50-3