Re: fsmonitor on linux is triggered due to unison changes its fingerprint cache

Hello Dale,

thanks!

On Sunday 21 September 2025, Dale R. Worley wrote:
> My understanding of all this is weak, but I notice an oddity in this bug
> report that nobody else seems to have commented on:
> 
> Ulrich Gemkow <[email protected]> writes:
> > The inotify watch tool shows this sequence about every second:
> >
> > # ~/.unison/homesync/ OPEN fpd51d53afbd71ea49fcec7a981a36e25f
> > # ~/.unison/homesync/ ACCESS fpd51d53afbd71ea49fcec7a981a36e25f
> > # ~/.unison/homesync/ CLOSE_NOWRITE,CLOSE fpd51d53afbd71ea49fcec7a981a36e25f
> > # ~/.unison/homesync/ OPEN fpd51d53afbd71ea49fcec7a981a36e25f
> > # ~/.unison/homesync/ MODIFY fpd51d53afbd71ea49fcec7a981a36e25f
> > # ~/.unison/homesync/ CLOSE_WRITE,CLOSE fpd51d53afbd71ea49fcec7a981a36e25f
> 
> That line that says "MODIFY" -- what is being modified and who is
> modifying it?  Your report says that nothing has been changed in the
> actual user files and that Unison makes no changes.

This is the output of a tool which uses the Linux inotify syscall to
watch for changes in a directory tree. It tells that the file named
at the end of the line is modified ((fpd51d53afbd71ea49fcec7a981a36e25f).

This is unisons fingerprint file.

> > Please note that the .unison dir is excluded from syncing (but this
> > is not known by the fsmonitor). So nothing is synced, unison on the
> > client side says
> >
> > # Looking for changes
> > #   Waiting for changes from server
> > # Reconciling changes
> > # Nothing to do: replicas have not changed since last sync.
> >
> > which is correct :-)
> >
> > I would expect that the fsmonitor is not triggered when there are
> > no changes in the replicas.
> 
> I don't know the details of fsmonitor, but I suspect it can be coerced
> to list *which files* have been modified.  Can we get that data?

Yes, fsmonitor is able to return the changed path(es) when
unison asks the fsmonitor to do so.

Please note that the output you commented is not the output
or a trace of fsmonitor, it is from a separate tool.

I am still investigating the problem by searching where the
"1-second" intervall comes from.

Thanks again and best regards

Ulrich

> 
> Dale
> 

-- 
Ulrich Gemkow
University of Stuttgart
Institute of Communication Networks and Computer Engineering (IKR)