I tried GlazeWM and it does what it says “technically” but there’s a lot of edge cases where it doesn’t and things just glitch out or don’t tile. Applications that don’t work well where you have to edit the configs per app etc. I settled on using Komorebi [1] instead and it’s a lot more of “just works”, just run it and go.[1] https://github.com/LGUG2Z/komorebi
bsnnkv|1 year ago
Applications behaving badly is a big problem for any window management project on Windows because application developers, especially since the rise of Electron, are increasingly throwing established Win32 application development guidelines to the wind with reckless abandon.
Komorebi deals with this by exposing functionality for handling edge cases to the end user with a powerful set of application matchers, and pooling together all of the weird application-specific fixes into a single upstream which is loaded for all new users on first install[3] and can be easily updated by the end user[4].
This upstream is now also consumed by at least one other window management project, Whim[5].
I think a lot of existing tiling window manager users from Linux and macOS look for straight "equivalents" on Windows if they end up here, but from my point of view both as an end user of tiling window managers and a developer of a tiling window manager, this is a huge missed opportunity.
Komorebi initially took a lot of inspiration from bspwm and yabai, but as I spent more time working with different monitors (square, wide, ultrawide) in different orientations, I became more and more convinced that a layout algorithm-driven design provides for a better overall experience than painfully manually manipulating a BSP tree to do something as simple as getting 3 columns on an ultrawide. More recently I also integrated some interesting ideas from more niche WMs such as PaperWM[6].
Tiling window manager development on Windows is really at the bleeding edge of the entire field these days, where the boundaries (especially in UX) continue to be pushed unencumbered by the inertia of established heavyweights and tradition.
Within the next decade I wouldn't be surprised to see this come full circle and find people implementing "Window Manager ___ From Windows, but for Linux/macOS".
[1]: https://www.youtube.com/watch?v=JNZw0qUf_PE
[2]: https://github.com/LGUG2Z/komorebi/commit/45894be4ff5162e1c2...
[3]: https://github.com/LGUG2Z/komorebi-application-specific-conf...
[4]: https://lgug2z.github.io/komorebi/example-configurations.htm...
[5]: https://dalyisaac.github.io/Whim/docs/customize/filtering.ht...
[6]: https://github.com/LGUG2Z/komorebi/commit/b799fd30774b5ae928...
lucianbr|1 year ago
Or is there some reason this behavior is not useful? I see some hints that it would be. More and more apps have tabs - windows explorer now, all IDEs of course, all text editors, browsers, terminals. And often they let you drag the tabs to some other convenient place, like Firefox lets you drag tabs to another Firefox window. Would it not make sense to generalise this to the entire screen?
toothbrush6|1 year ago
What apps did you find were bugging out? The default handling has been tweaked and improved a lot over the years but there's still some edge cases that are tough to handle
emptysongglass|1 year ago