[Python-Dev] Bizarre mtime behaviour

[Robert Collins]

On Sun, 2009-11-01 at 12:03 -0500, Eric Smith wrote:
> Antoine Pitrou wrote:
> > Adam Olsen <rhamph <at> gmail.com> writes:
> >> Looks like an OS bug to me.  Linux I'm guessing?
> > 
> > Yes, but only on certain boxes. I could never reproduce on my home box.
> > RDM (David)'s buildbot is a Gentoo vserver with a reiserfs filesystem.
> 
> You'll occasionally see something similar on Windows boxes running FAT, 
> because the timestamps stored in the filesystem have a 2 second 
> granularity there. Not sure if this might be similar, or just a reiserfs 
> issue.

Some years ago there was a bug where the dentry cache held timestamps
more granular than those that the filesystem could represent. We saw
this cause build failures which were troubling to reproduce :) I haven't
seen Linux arbitrarily invent time travel so far:).

The FAT rounding issue is a possibility, but I didn't think reiserfs was
short that much precision.

I'd check that the work area you had really was reiser, not a mounted AT
partition, and if its not look up the ReiserFS precision for mtime.

-Rob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/python-dev/attachments/20091102/c7b81020/attachment.pgp>