search

skx / kpie

Simple devilspie-like program for window manipulation, with Lua.
GNU General Public License v2.0
80 stars 13 forks source link

Add Option to Prevent Window Focus #28

Closed Myrddin-Wyllt closed 3 years ago

Myrddin-Wyllt commented 5 years ago

I've sought a method to apply per application window focus rules for too long now. While easy to accomplish in compiz, openbox, and KDE, finding something for XFCE has been exceptionally hard. devilspie2 lacked this feature & devilspie doesn't seem to have an obvious way to prevent input from being given to a window.

The reason is mainly for Steam Keyboard compatibility. Without this as a rule, usage of Steam's on screen keyboard becomes cumbersome, having to select each input field while holding the Steam button before being able to type because the Steam Keyboard chokes on input strokes. Steam will freeze & you will be having a bad time. Preventing input focus should solve this issue.

This patch is by no means complete. I've never had experience in this syntax, so I know in its current state, this is unusable. With a little polish, this would be ready. Here's some references you might find useful: https://tronche.com/gui/x/xlib/ICC/client-to-window-manager/wm-hints.html https://tronche.com/gui/x/xlib/ICC/client-to-window-manager/XSetWMHints.html https://tronche.com/gui/x/icccm/sec-4.html#s-4.1.7

As a bonus, I threw in the PKGBUILD some arch member decided to put up here not long ago. It seems to have had a few variable errors which I believe have been sorted out.

Lastly, the readme should probably be updated as autoreconf --install doesn't work, nor does ./configure. This program only builds with make alone (at least on my setup).

skx commented 5 years ago

It seems like this pull-request says three things:

The autoconf support was merged as a result of #14, but you'll see there that I never got confirmation it was working - My solution will be to remove it from the repository and the README file.

The PKGBUILD update seems fine.

So then we get to the real meat of your pull-request. The code isn't perfect but it's not horrid. If you could confirm you'd been able to successfully test it I'd accept it, even if I cleaned it little afterward. But without knowing that it works I'd be crazy to merge as-is :)