[PATCH] improve xfsinvutil man page and argument processing

[Christoph Hellwig]

On Wed, Dec 23, 2009 at 10:57:13AM -0500, Christoph Hellwig wrote:
> Some notes to look into after this patch:
> 
>  - can you provide a testcase for the problems caused by the wrong
>    time_t usage?  A patch to xfstests would be perfect, but if you
>    have a raw testcase I'll vounteer to wire it up.
>  - why do we wrap both ctime and ctime_r?  Currently xfsdump isn't
>    multithreaded so we shouldn't need it.  But if the parallel
>    dump/restore ever gets merged from IRIX we need to get rid of the
>    plain ctime calls.  (Btw, are there any plans for the parallel
>    dump/restore port?)

Err, sorry - thos comments were for the time32 patch, but this patch
also looks good,


Reviewed-by: Christoph Hellwig <hch at lst.de>