Bug switching views

[Dirk Hohndel]

Chris Lewis <chrislewis915 at gmail.com> writes:

> OK i'm now struggling with this.
>
> I'm running using x forwarding over SSH to an x server running on a windows
> laptop.  After a few switches between views using Ctrl (1,2,3,4) i
> eventually lose the ability to get all three panes displayed again.  Ctrl+4
> does nothing, selecting it from the menu also doesnt work
>
> In fact Its now stuck like this and i cant get the three pane view back.

you should always be able to drag the edges (the panes aren't gone, they
are just pushed all the way to the edge - the handles are still
there...)

Can you add a printf in save_pane_position that would help us figure out
what happens?

/D


> On 4 March 2013 08:36, Dirk Hohndel <dirk at hohndel.org> wrote:
>
>> Dirk Hohndel <dirk at hohndel.org> writes:
>>
>> > "Robert C. Helling" <helling at lmu.de> writes:
>> >
>> >> On Sun, 3 Mar 2013, Dirk Hohndel wrote:
>> >>
>> >>> "Robert C. Helling" <helling at atdotde.de> writes:
>> >>
>> >> Dirk,
>> >>
>> >>> I decided to tackle this one myself.
>> >>
>> >> thanks. After yesterday's patch, I had not been able to reproduce
>> >> Sergey's problem anymore but if this mornings message would have been
>> that
>> >> it is persistent I would have implemented a similar thing with a global
>> >> variable to indicate THREE_VIEW.
>> >>
>> >> That said, I can confirm that this seems to work for me (but so did my
>> >> last version ;-)). There is just one comment: Why don't you do the
>> testing
>> >>
>> >> if (pane_conf == PANE_THREE)
>> >>
>> >> in list_view() and the other two and not in save_pane_position()? There
>> it
>> >> would replace my if()'s testing against window geometry.
>> >
>> > I didn't know if those might come in handy in other situations...
>> >
>> >> This would eliminate some code duplication and, more importantly,
>> >> save_pane_position() is also called from save_window_geometry() whose
>> >> calling pattern is not obvious to me but I guess is it is called on
>> >> quit at least. So, theoretically, we could still have the situation
>> >> where the pane position is saved upon exit if we are not in
>> >> PANE_THREE.
>> >
>> > Ok, that makes sense.
>> >
>> >> Here is a patch to show what I mean.
>> >
>> > I'll take that and recreate the binaries / sources - the save on exit
>> > thing makes a difference :-(
>>
>> Updated binaries / sources are on subsurface.hohndel.org
>>
>> /D