(I hope you don't mind if I bring that back to the list)
On Sat, 2011-01-22 at 15:50 -0500, Hans-Christoph Steiner wrote:
----- Original message -----
On Sat, 2011-01-22 at 15:31 -0500, Hans-Christoph Steiner wrote:
----- Original message -----
On Fri, 2011-01-21 at 18:02 -0500, Hans-Christoph Steiner wrote:
It was changed to work on GNOME/metacity, which is the default on Fedora, Debian, Ubuntu, Mint, etc. So it is by far and away the most common. Since it is going to be different with each window manager, I think it makes sense to have the defaults work for the most people.
Didn't it work for everything before? Or am I even wrong in thinking that up to 0.42 Pd used to write the window geometry into the patch file and not the canvas geometry? Sorry to repeat myself: Why was it changed to storing the canvas geometry?
Roman
from my experience, it was not working, that's why I changed it. I didn't realize at the time that it was different on each window manager in X11, I only use GNOME/metacity.
Not that I care _that_ much about it, but I still wonder: In what way is it working better now than it was before? I wouldn't call it 'working' when it actually only works with one window manager (metacity).
OTOH, I must say that it is quite easy to make it working in non-metacity window managers. Only two numbers need to be adjusted in <pd-src>/tcl/pdtk_canvas.tcl.
Is there a way for tcl/tk to know in which window manager it is running in? If so, a real solution wouldn't be too hard, hopefully.
Roman
there has got to be a way, at worst you could always "ps auxww | grep metacity" for each supported WM. I only use the default Debian and Ubuntu setup, so I not the one to take this project on.
True.
However, I think it's not even tied to the window manager, but to the theme it uses. At least I tested it also on Gnome/Metacity (default theme on Lucid) and it was still off (which might be an indication that your Gnome is different from mine). Also I believe it is dangerous to assume that the default theme uses the same measures on different Ubuntu versions or the measures are the same across different distros (most likely they are _not_). I don't see how it could be solved in a consistent way. I consider the current implementation rather 'broken' than 'working'.
This is not a rant, but I think we should be aware of where things might fail. Personally, I am very fine with adjusting two numbers in a file whenever I install Pd.
Roman