Re: [Monotone-devel] Monotone fails to detect certain changes in files

[Nuno Lucas]

On 8/25/07, Richard Levitte <address@hidden> wrote:
> In message <address@hidden> on Sat, 25 Aug 2007 16:49:37 +0100, "Nuno Lucas" 
> <address@hidden> said:
>
> ntlucas> Off course that if I do a "refresh_inodeprints" or remove the
> ntlucas> inodeprints file it will work, but that is not the point.
> ntlucas>
> ntlucas> My point is I'm trying to understand if this is a bug or a
> ntlucas> "work as designed" kind of thing.
>
> I would say it's work by design.  The inodeprint for each file is a
> hash of the concatenation of ctime, mtime, mode, inode number, device
> number, uid, gid and size.  If none of those change and only content,
> then inodeprints isn't for you, as your way of changing the files
> REQUIRES checking the contents, making inodeprints a moot feature.

Ok, it makes sense and I can live with that.
At the same time, as I noticed on my previous answer to Pavel, it
seems if I do a msync at the end Linux should update the file mtime,
even if atime is not. So that shows a POSIX workaround that can't be
used on Linux.

Well, that's life, and only bothered the list with this because I feel
it's a behavior that other people can notice later and this way
becomes documented.

Best regards,
~Nuno Lucas

> Cheers,
> Richard
>
> --
> Richard Levitte                         address@hidden
>                                         http://richard.levitte.org/
>
> "When I became a man I put away childish things, including
>  the fear of childishness and the desire to be very grown up."
>                                                 -- C.S. Lewis
>