Unison 2.10.2 fast update check broken?

[Marcus Picasso]

Rolf Campbell wrote:
> Marcus Picasso wrote:
> 
> > Seems that Cygwin port of the unison file synchronizer does not do the
> > -fastcheck very well. Transcript follows:
> > 
> > ...
> > 
> > Can somebody confirm / explain this behaviour? I have a large tree that
> > I'm synchronizing across two hard-disks, and got suspicious when
> > re-running synchronization takes longer than expected. The above
> > transcript functions as expected using linux or native Win32 unison
> > builds.
> > 
> > Regards,
> > -Marcus.
> 
> I have noticed a change in how -fastcheck which seems to be caused by my
> upgrade from cygwin 1.5.14 -> 1.5.16. I tried doing a unison sync between
> a maching running 1.5.14 and a machine running 1.5.16 when I noticed the
> 1.5.16 machine spent a lot of time grinding the disk. So, I upgraded the
> 1.5.14 machine to 1.5.16 and it too went from a 10 second scan time to a
> half hour of heavy disk access.
>
> ...

That's exactly my issue also.

I think it's the recent ctime changes in Cygwin that has broken unison. The
ctime stamp that gets recorded in the unison archive database is slightly
off, compared to the actual ctime stamp of the file that got modified by
unison. 

Could it be that unison reads the ctime stamp, then closes the
file, which results in an update of the stamp, causing the mismatch of the
stamps? Andrew, any ideas how to fix this? :)

-Marcus.


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/