Re: Separation into dock widgets

[John Swensen]

On Apr 29, 2011, at 5:23 AM, Jacob Dawid wrote:

> 
> 
> 2011/4/29 Michael Goffioul <address@hidden>
> On Fri, Apr 29, 2011 at 10:50 AM, Jacob Dawid
> <address@hidden> wrote:
> > I wrote about that here:
> >
> >> [...] but DockWidgets are not intended by design what they should do in
> >> this case, they rather act as toolbars around the central widget (it is
> >> possible to "hack" it by assigning an empty central widget and not allowing
> >> them to overlap, but that's is not working well for all situations).
> 
> Could you be more specific about what's not working well?
> Can't the problems be worked around?
> 
> 
> The root problem is that you trying to use a solution that it is not designed 
> for the problem. From these there are several other issues, for example you 
> only get that separate and flexible window solution by trading off some other 
> things, like deactivating overlapping for all subwindows. Further, you can't 
> place windows anywhere - there is no central window anymore. Subwindows will 
> always divide the space and fill it out.
>  
Jacob, this isn't true.  Try setting the dockOptions for the QMainWindow to 
(QMainWindow::AllowNestedDocks|QMainWindow::AllowTabbedDocks).
The problem I was running into was that I couldn't get the QDockWidgets to 
allow docking to two different QMainWindow's.  Mailing lists said this should 
be possible, but I couldn't get it to work.


> > Please do not judge that by merely saying shortcuts are a bad thing before
> > trying out how comfortable this is, we're not average Windows users :). In
> > fact, there are also arguments against having a terminal window and the
> > editor side-by-side, like wasting screen space for example. Let's recall,
> > what are the benefits?
> 
> Debugging. When debugging code, having the terminal window and the
> source code is important. This is just my opinion, but I think I'm not the
> only.
> 
> Michael.
> 
> Excuse me that question, but I don't know much about Debugging in Octave. How 
> do you debug code? Do you run it command by command from the editor? I mean, 
> usually you mark some breakpoints, run it, so the focus switches to the 
> terminal until it reaches some breakpoint, you stop and switch back to the 
> editor. Or is it different?

I see that you cut out a bunch of stuff that I had in octave_server for the 
OctaveLink class.  Some of that was for setting, clearing and querying debug 
information from Octave.  When debugging I usually have the editor showing the 
code I am debugging, along with the breakpoints that have been set, where the 
current execution is stopped, etc. along with the terminal so I can quickly 
inspect variable dimensions, values, etc.

I am still adamant that the right solution is to let the user configure their 
windows how they wish.  As an undergrad, almost everyone used the Matlab GUI in 
almost the same configuration.  As a PhD student almost every student has the 
Matlab GUI in a different configuration.  However, one of the constants is that 
the command window and editor window are visible simultaneously.

John Swensen