Re: [gamin] another release soon

[Daniel Veillard <veillard redhat com>]

On Tue, Aug 09, 2005 at 10:22:43AM -0400, John McCutchan wrote:
> On Tue, 2005-08-09 at 10:13 -0400, Daniel Veillard wrote:
> > On Tue, Aug 09, 2005 at 09:53:17AM -0400, John McCutchan wrote:
> > > >   Listener nautilus has 16 subscriptions registered
> > > >     Subscription 3 reqno 1008 events 1008 dir 1: /media/usbdisk
> > > >     Subscription 1 reqno 1008 events 1008 dir 0: /media/usbdisk
> > > > ...
> > > > but I didn't see any of the poll callbacks every seconds which should
> > > > show in the logs. 
> > > 
> > > I'm going to add some extra debug code to the subscription layer so that
> > > we can see what method is chosen. To see the poll messages, you need to
> > > #define VERBOSE_POLL in gam_poll.c , you can just uncomment the line at
> > > the top.
> > 
> >   I'm adding a gam_poll_debug() to list polled resources too.
> 
> I just committed a debug messages that are printed out the method and
> the poll limit time when a gam_node is created. And a message when a
> subscription is added that will let us know if it was excluded, or what
> the gam_fs chose for kernel/poll/etc.

  the events are generated but don't seems to be sent through the connection

Poll: emit events 16 for /media/usbdisk
Event to nautilus : 1, 1, /media/usbdisk Changed
Event to nautilus : 1, 1, /media/usbdisk Changed
Poll: scanning directory /media/usbdisk
g_n_n: node for /media/usbdisk/test using kernel with poll timeout of 0
Poll: poll_file for /media/usbdisk/test called
Poll: file is new
Poll: emit events 32 for /media/usbdisk/test
Event to nautilus : 2, 5, test Created
Event to nautilus : 3, 5, test Created
Event to nautilus : 2, 5, test Created

  but nautilus did not refreshed the view ...

Same thing when using the standalone debug tool:

paphio:~/gamin/tests -> ./testgam -
> connect
connected
> mondir /media/usbdisk
mondir /media/usbdisk 0
>

  the events for the content are not seen. Adding a file to the directory,
no event is seen either by testgam ...

  Once again the regression tests flagged a problem which wasn't caught
before the code was commited, digging after the fact to try to find where
this comes from makes it way more complex.

Daniel

-- 
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/