Steve Dickson wrote:
> Chuck Lever wrote:
>> After I enabled NFS_MOUNT_DEBUG the compiler started spitting out 
>> warnings.
>> Fix up commas, output formatting, and double-wide character support.
> Aren't these debugging statements basically useless because one has to
> recompile to turn them on... Wouldn't it make more sense to add
> '-o mntdebug' flag so these could be turned on at will?

Useless for folks who don't want to or can't compile nfs-utils 
themselves, yes.  But I think the usefulness of these is for anyone who 
is developing mount.  They aren't for debugging problems mounting 
shares, but for debugging changes to mount itself.

I'm not opposed to making these debugging messages available dynamically 
(if we really believe they will be useful for everyday system 
administrators).... but....

Miklos has illuminated an important distinction between mount options 
that affect the behavior of the mounted file system, and mount options 
that control who can mount and how the mount takes place.  NFS is one of 
the worst offenders in mixing the two.

A "mntdebug" option would be of the latter type, and I'm beginning to 
think twice about the value of adding more of these.  I'm thinking about 
a way of specifying those types of options in some other way; for 
example, by using a separate NFS mount policy file like nsswitch.conf to 
set these up.  Just a random thought.

Or, rather, instead of "-o mntdebug" we might consider --mntdebug 
instead; that is, use a command line option instead of a mount option.

It's a nit, I know.