On Fri, Jul 12, 2002 at 12:31:44PM +0300, Tom Alsberg wrote:
> Hi, again, comments below, again:
>
> On Fri, Jul 12, 2002 at 02:45:58AM +0200, Dominik Vogt wrote:
> > On Thu, Jul 11, 2002 at 09:17:41PM +0300, Tom Alsberg wrote:
> > > <snip />
> > > Well, what do you mean - pulling your leg?
> >
> > Pulling somebody's leg means roughly to mislead someone for fun.
>
> Oh... You'll have to forgive me, I'm not a native english speaker,
Neither am I, but I have a new dictionary and enjoy looking up
phrases I'd use in German :-)
> so I didn't understand exactly the relation with the context until now.
>
> > > I am just saying, that now with -menter, it works the way I want it,
> > > but that isn't exactly the way it should work, as documented in the
> > > manual page.
> >
> > The man page is correct, but I did not mention that moving the
> > pointer from any window into the root window now generates an
> > ENTER_WINDOW packet too.
>
> So doesn't this mean that now every leave event is accompanied by an
> enter event?
That would be the ideal solution, but I'm not convinced this is
the case at the moment. Handling of the corresponding X events is
very complicated. There may still be situations where an enter or
leave event is sent without the other.
> What is the difference now, between using the -menter
> option and the -menterleave option?
With -menter FvwmAuto only handles enter events. The LeaveCommand
is triggered for the previous window when an enter event arrives.
With -menterleave only leave events trigger the LeaveCommand. As
long as these events arrive in pairs, -menterleave is just a waste
of resources.
> > Er, first you say that "it works the way I want it" and a few
> > lines further down "So unless ... I will not have it the way I
> > want"? So is the behaviour of -menter what you want or isn't it?
> > Automatic lowering when entering the root window can be disabled
> > by checking if the passed id is that of a real window or not (in
> > other words: in the called functions without any changes to
> > FvwmAuto).
>
> Well, it works the way I want it, now. But that seems to be not
> exactly the way it should work, according to the documentation /
> manual page.
> So when/if it will be changed, to work the way it should work, then it
> won't work the way I want it, and I will be unhappy again... :-)
Don't worry. I've already changed the documentation to reflect
the current behaviour.
> Or, to say it otherwise, what makes me happy now, probably should not
> be that way...
>
> The behaviour of -menter is what I want, because it seems to be the
> same as the behaviour of -menterleave, only without the problem with
> gv and grabbing the pointer...
Yes. But it is only the same because we specifically wrote the
function this way. It's not an inherent feature of FvwmAuto.
There may still be situations where you might want to treat enter
and leave events separately. If we did not want the raise-when-
root-window-entered behaviour, we could just change the function
to ignore this case. No code in the module needs to be touched.
> On the other hand, now that you say that the pointer leaving a window
> for the root window also generates an enter event for the root window,
> that might be the appropriate behaviour... But then what is the
> difference between the -menter option and the -menterleave option?
See explanation above.
> > > Well, I'm pretty sure... I listed the running processes after
> > > restarting FVWM, and the FvwmPager which is running is the one
> > > installed by the last make install (checked)...
> > > Also now I tried building from scratch, removing any previous
> > > installation I had locally, and installing of clean, being sure that
> > > there wasn't any FvwmPager left somewhere before, it still looks that
> > > way.
> > >
> > > Doesn't it happen for you? I checked, and even with the default
> > > configuration (running as 'fvwm -f no-rc'), and nothing done except
> > > popping up a pager, it looks different...
> >
> > Hm, default configuration. All my modules use colour sets. That
> > effectively hides any windows with BackColor. Can you give an
> > exact date for "a few days back"?
>
> "A few days back" is here since the 6th of July, to the 11th of
> July... In those times, FVWM updated from CVS head (I updated about
> twice a day) had that problem with FvwmPager.
> On the other hand, FVWM updated from CVS a half an a hour ago (12th of
> July, 2002, 12:10 AM GMT+03), seems to have fixed this...
>
> > Originally I thought this meant before the FvwmPager fixes I made,
> > but I'm not so sure anymore.
>
> Well, back when you said you updated a lot of the FvwmPager drawing
> code, some time after that it wasn't fixed yet. But apparently there
> were some FvwmPager fixes today/yesterday too, and those seemed to
> have fixed it...
I found the bug. FvwmPager was accessing the wrong information
from the packet informing it of the new window. That happened
because the size of a variable in the packet changed and the pager
accessed the data by byte offsets in the structure :P
Bye
Dominik ^_^ ^_^
--
Dominik Vogt, mail: dominik.vogt_at_schlund.de, phone: 0721/91374-382
Schlund + Partner AG, Erbprinzenstr. 4-12, D-76133 Karlsruhe
--
Visit the official FVWM web page at <URL: http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm" in the body of a
message to majordomo_at_fvwm.org.
To report problems, send mail to fvwm-owner_at_fvwm.org.
Received on Fri Jul 12 2002 - 05:43:58 BST