hello, please see the comments below
--
Martin Gracik
----- "Chris Lumens" <clumens@xxxxxxxxxx> wrote:
> > @@ -43,6 +45,11 @@ class childWindow:
> > moduleName = (_("Network Setup"))
> > moduleClass = "reconfig"
> >
> > + needsparent = 1
>
> Please use "True" instead of 1.
>
> > @@ -166,6 +173,30 @@ class childWindow:
> > #Run rhn_register so they can register with RHN
> > pid = start_process("/usr/bin/system-config-network")
> >
> > + # get the x window id
> > + xid = None
> > + start = time.time()
> > +
> > + # keep trying for 3 seconds
> > + while xid is None and (time.time() - start < 3):
> > + rc, out = commands.getstatusoutput('xwininfo -name' \
> > + ' "Network Configuration" -int')
> > + pattern =
> re.compile("xwininfo:\sWindow\sid:\s(?P<xid>\d+)\s.*")
> > + for line in out.splitlines():
> > + m = pattern.match(line)
> > + if m:
> > + xid = long(m.group("xid"))
>
> I assume you're waiting to make sure that system-config-network has
> started. Did you encounter problems in testing that required
> waiting?
yes, it seems it takes some miliseconds for the system-config-network to create the window,
if i run xwininfo right after system-config-network, i get "no window" error,
so we have to wait some time, just 1 second worked for me, but it may not work on slower systems.
i hope 3 seconds must be enough, otherwise there's some problem, and we ignore the new code.
>
> I think a more natural way to format this loop is:
>
> xid = None
> i = 0
>
> while not xid and i < 3:
> rc, out = blah
>
> pattern = blah
> for line in out.splitlines():
> m = blah
> if m:
> xid = blah
>
> i += 1
> time.sleep(1)
>
> This avoids having to calculate time as well as avoids hammering the
> system by calling xwininfo as fast as possible. Note you're not
> doing
> any sleeping in your loop so as soon as xwininfo exits
> unsuccessfully,
> you're going to try it again.
yeah, i can do it like this
>
> Note also that you're not checking the return value of
> commands.getstatusoutput. I hope that out is "" in the case where
> the
> commands fails.
i think i don't need to check the return value, because later on,
i check if i have a window number in xid, so if the command fails,
i have no xid number, and we will get no traceback, but the dialog window
will work as before, that means, it will be hidden when you click somewhere else,
that's a fallback...
>
> > + # if we have the xid, embed the window
> > + if xid is not None:
>
> You can more simply say:
>
> if not xid:
i think you meant, "if xid:",
well, i don't know if it's possible, probably not,
but what if xwininfo returns a window id = 000000 ?
then this check will fail, and that's not what should happen.
i check explicitly for None, because i set it to None in the beginning
of the loop, so if there is something wrong in the xwininfo output,
it will stay as None.
also, maybe it was a stupid article, but i read somewhere, that it's a good
habit to check for None.
maybe you have some experiences where it's not a good idea, tell me why,
i'm happy to learn from someone more experienced
>
> > + network_dlg = gtk.gdk.window_foreign_new(xid)
> > + self.parent.win.realize()
> > + network_dlg.set_transient_for(self.parent.win.window)
> > + network_dlg.set_modal_hint(True)
> > + # XXX if we keep the main window fullscreen,
> > + # and use alt+tab, the network window is gone forever
> > + self.parent.win.unfullscreen()
> > +
> > flag = None
> > while not flag:
> > while gtk.events_pending():
> > @@ -178,5 +209,8 @@ class childWindow:
> > else:
> > time.sleep(0.1)
> >
> > + # make the main window fullscreen again
> > + self.parent.win.fullscreen()
> > +
> > i.grab_remove ()
> > self.updateLabels()
>
> I can't really speak to the window manipulations here, so I'll assume
> if
> it works and avoids the problem of having the gnome panel appear over
> top, it's probably okay.
well the gnome panel still appears, but only when the system-config-network window is opened.
during the rest of the process, the panel is not visible.
i can't figure out what's wrong with it, and how to make it go away,
i tried to talk to guys in #pygtk, but still couldn't figure it out.
>
> - Chris
>
> _______________________________________________
> Anaconda-devel-list mailing list
> Anaconda-devel-list@xxxxxxxxxx
> https://www.redhat.com/mailman/listinfo/anaconda-devel-list
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list
