On Thu, 14 Apr 2005, Jeff Moyer wrote:
> ==> Regarding Re: [autofs] Re: [ANNOUNCE] Important update - autofs 4.1.4
> release; Ian Kent <[EMAIL PROTECTED]> adds:
>
> raven> On Thu, 14 Apr 2005, Ian Kent wrote:
> >> On Wed, 13 Apr 2005, Jeff Moyer wrote:
> >>
> >> >
> >> > This new version has regressed since beta2 for /net mounts in my
> >> environment. A simple ls of /net/<filer> will now show the contents of
> >> > /net/<filer>/exporteddir. So, for example, we have devfiler which
> >> exports /vol/vol1. When I do an ls of /net/devfiler, I get the
> >> contents of devfiler:/vol/vol1.
> >> >
> >> > Okay, after further debugging, it seems we're not identifying this as
> >> a multimount entry. It's getting late, so I'm going to turn this over
> >> to you. I think that the checking of first_chunk is bogus. For
> >> instance, with the below change, everything works as expected for me.
> >>
> >>
> >> Cool, I'll check this out and get back.
>
> raven> Could I have the example map entry that breaks the parse please.
>
> Sheesh, you'd think that I could generate a bug report by now... I'll blame
> it on the late night thing. =)
>
> This is an auto.net entry. Here's what auto.net spits out for this
> particular host:
>
> -fstype=nfs,hard,intr,nodev,nosuid \
> /vol/vol1 devfiler:/vol/vol1
>
> I'm guessing you didn't test auto.net against a host that only exported one
> dir? Just a guess. Below are the debug logs.
>
I think this should do the trick.
--- autofs-4.1.4/modules/parse_sun.c.first-multi Fri Apr 15 14:30:18 2005
+++ autofs-4.1.4/modules/parse_sun.c Fri Apr 15 14:42:04 2005
@@ -768,6 +768,10 @@ static int check_is_multi(const char *ma
int multi = 0;
int first_chunk = 0;
+ /* If first character is "/" it's a multi-mount */
+ if (*p == '/')
+ return 1;
+
while (*p) {
p = skipspace(p);
_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs
