Re: [coreutils] [PATCH 1/2] stat: support printing birthtime

[Pádraig Brady]

On 01/10/10 00:32, Eric Blake wrote:
> * src/stat.c (print_stat): New %w and %W formats.
> (do_stat): Include %w in verbose format.
> (usage): Document them.
> * doc/coreutils.texi (stat invocation): Likewise.
> * NEWS: Likewise.
> Suggested by Andre "Osku" Schmidt.
> ---
> 
> I've tested that this works on cygwin.  On Fedora 13 with an
> ext4 partition, the birthtime appears to not be returned in
> stat().  If the kernel guys will ever commit to a stable
> xstat() interface, which we can then write gnulib wrappers
> for, we can use that instead.  I'm assuming this will also
> work on BSD systems, although I have not yet tested that.
> 
> I wasn't sure how to write a test for this - how to tell if
> a filesystem has birthtime support exposed by the kernel?
> Ideas on that front are welcome.
> 
> For an example of what it looks like on a file system with
> all four times:
> 
> $ src/stat ChangeLog
>   File: `ChangeLog'
>   Size: 496660    488        IO Block: 65536  regular file
> Device: 10897b43h/277445443dInode: 562949954522488  Links: 1
> Access: (0644/-rw-r--r--)  Uid: ( 1007/  eblake)   Gid: (  513/    None)
> Access: 2010-09-30 16:58:48.859000000 -0600
> Modify: 2010-09-30 16:52:50.000000000 -0600
> Change: 2010-09-30 16:58:06.176250000 -0600
>  Birth: 2010-09-30 16:58:06.098125000 -0600

So by default on Linux we'll get:

  Birth: -

Might it be better to suppress the line if not present
given it's lack of support by file systems?

> +  %x   Time of file birth, or - if unknown\n\
> +  %X   Time of file birth as seconds since Epoch, or - if unknown\n\

s/x/w/

cheers,
Pádraig.