pam is so wrong, and autconf is so wrong, that people using both even
manage to fuck up configure tests.
so it compiles on linux, it must be right ?
pathetic...
more syslogging of login failures so we have have more info on why.
patch-etcafpd-volumec
Update to partially resolve the current problem with afpd not behaving
correctly with permisions on /etc/netatalk.
Log correctly to /var/log/daemon when we can't access configuration files.
Put in place proper error checking when reading them. Would be nice if the
origional programmers actually checked for failure for anything.
People can now see just why things are not working as they like.
Mar 29 18:08:38 kashmir afpd[6210]: session from 39148.187:250 on
39148.169:129
Mar 29 18:08:38 kashmir afpd[6210]: login dingo (uid 1002, gid 10)
Mar 29 18:08:38 kashmir afpd[6210]: unable to access
/etc/netatalk/AppleVolumes.system: Permission denied
Mar 29 18:08:40 kashmir afpd[6210]: done
Mar 29 18:08:40 kashmir afpd[21593]: asp_chld 6210 done
Pass the CORRECT arguments to creatvol. from "Benninghoff, John"
<JABenninghoff@dainrauscher.com>
Return AFPERR_PARAM when we can't access configuration files to
the appleshare client requesting access. This stops possible DOS under
MacOS. In it's current form the Appleshare client has to be killed on the
MAC side by killing the "CHOOSER" Application. By returning proper errors
the appleshare client exits gracefully with error: "An Appleshare system
error occured."
patch-etcpapd-lpc
cosmetic change: remove an unused variable.
patch-version
changes made bump version.
--
From: maintainer
- add HOMEPAGE, another distsite, NEED_VERSION and some more docs
- respect CC and CFLAGS
- move RESDIR from share to share/netatalk
- move examples from lib/netatalk to share/examples/netatalk
- remove the need for having the source tree installed, just use the
headers in /usr/include/netatalk
This update implements changes to afpd (appletalk filing protocol daemon)
so that it handles users groups correctly for acess and handles the
effictive user and group of the parent process correctly under OpenBSD
2.5-beta. Previous fix didn't work quite as expected.
