==> Regarding Re: [autofs] csh/cd not getting the right cwd at the mount point; [EMAIL PROTECTED] adds:
raven> On Fri, 9 Jul 2004 [EMAIL PROTECTED] wrote: I'm having raven> difficulty understanding the problem you describe. >> We have a problem. Here are the conditions: >> >> 1. Shell is (t)csh (does not happen under (ba)sh) 2. cd to an autofs >> fa�ade (e.g. /tools in our case) 3. cd relatively to the subdirectory >> that is the mount point. NOTE> The mount must not yet be made. (e.g. perl5.8.3) >> 4. cd relatively to a subdirectory in the mounted files system (e.g. bin >> or lib) [snip] raven> I recommend you try autofs4 release 20040508. Unfortuneately this raven> release requires a kernel rebuild as there is a small mandatory raven> change outside of the autofs4 module needed for correct raven> operation. This release is the one that should be used as it raven> addresses all known problems and works best with the latest autofs raven> release. raven> You should be using autofs-4.1.3. There is an additional patch that raven> I have recommened a couple of times now, without adverse feedback, raven> that my fix your problem. I've uploaded it to kernel.org and it can raven> be found at: [snip] NOTE: I was not able to reproduce the problem in my environment, which included our next update kernel that is fully patched with Ian's latest kernel patch kit. I'm using autofs-4.1.3-10, which will appear on my people page soon. raven> As a RHEL user visiting Jeff Moyers autofs page is a must as there raven> are patches included their that have not yet been included in an raven> autofs release raven> http://people.redhat.com/~jmoyer raven> Jeff, would it be possible to include the two patches mentioned raven> above in your rpms please. Sure thing. I'm doing builds for this now. I'm currently tracking the following patches that I _think_ are queued for the next release. Please correct me if I'm wrong, or if you want me to resend anything. Patch12: autofs-4.1.2-option-parsing.patch - Make regular expressiosn for options parsing a bit more strict. Fixes bugs I have filed on the package for false matches of either mount points or other fs options. Patch14: autofs-4.1.3-underlinei18n.patch - This is a patch which makes translation prettier (AFAIUI). Note, I didn't write this, and didn't send it to you. I've attached it to this mail. Patch15: autofs-4.1.3-rpc-ping.patch - Ian, you know very well what this is. ;) And, the two you just had me put in: Patch16: autofs-4.1.3-bad_chdir.patch Patch17: autofs-4.1.3-mtab_lock.patch -Jeff --- autofs-4.1.3/samples/rc.autofs.in.i18n 2004-06-21 13:22:12.000000000 -0400 +++ autofs-4.1.3/samples/rc.autofs.in 2004-06-21 13:31:29.000000000 -0400 @@ -244,12 +244,10 @@ # function status() { - echo $"Configured Mount Points:" - echo "------------------------" + echo -e $"Configured Mount Points:\n------------------------" getmounts echo "" - echo $"Active Mount Points:" - echo "--------------------" + echo -e $"Active Mount Points:\n--------------------" ps axwww|grep "[0-9]:[0-9][0-9] $DAEMON " | ( while read pid tt stat time command; do echo $command; done ) @@ -356,7 +354,7 @@ fi [ $RETVAL = 0 ] && touch /var/lock/subsys/autofs else - echo -n "" $"No Mountpoints Defined" + echo -n $"No Mountpoints Defined" success "$prog startup" fi rm -f $TMP _______________________________________________ autofs mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/autofs
