Software 44861 Published by

MangoWM 0.17.0 has officially landed as a milestone release, marking the completion of the Wayland compositor's original feature roadmap. Built on dwl and wlroots, the update brings polish to existing tools like scratchpad animations, dwindle layout improvements, and a toggleable floating workspace while consolidating legacy input directives under a unified devicerule system. With the core feature set now locked down, maintainers confirm that future development will shift entirely toward bug fixes, performance refinements, and long-term stability. 



MangoWM 0.17.0 Hits the Milestone, Locking Down the Feature Set

The Wayland compositor built on dwl and wlroots says the hard part is over, and now it's just polish and bug fixes.

If you've been tracking the Wayland compositor scene outside of the KDE and GNOME heavyweights, you've probably seen MangoWM pop up with increasing frequency. It's a project that leans heavily into a simple premise: take dwl, slap wlroots underneath, and make it actually usable for your daily tasks. The result is a tiling compositor that's fast, lightweight, and stubbornly opinionated about what makes a window manager good.

With the release of version 0.17.0, the maintainers are calling it a milestone. Not because they're adding a dozen new features, but because they're finally adding none. The original feature list is complete, the codebase has been reorganized for better IDE support, and the project is officially shifting from feature development to refinement.

That's a bold move for a project that's only three years old, but it aligns with their stated philosophy of practicality over novelty. You're not going to find experimental features here. You're going to find a compositor that's solid enough to actually run your desktop.

Mangowc

What Changed in 0.17.0

The headline feature is the feature freeze itself. According to maintainer @DreamMaoMao, the core development cycle is done. Future updates will focus on bug fixes, incremental improvements, and stabilizing the existing set of tools. It's a clear signal to potential users: this isn't going to change dramatically in the next six months.

There are eight new features in this release, though most are refinements rather than groundbreakers. The most noticeable additions include a special floating workspace that you can toggle on demand, which is handy if you want a persistent utility window without cluttering your regular tags. The dwindle layout now remembers focus state and supports client movement via IPC, which should make workflow adjustments less painful.

Scratchpad windows now animate when you show or hide them. It's a small touch, but it softens the otherwise jarring instant pop-in behavior that plagues many tiling compositors. Mouse and touch follow gestures also got smoother, with interpolation replacing the previous jump-to-target behavior. Gamers might appreciate the new confine_pointer window rule, which locks the cursor inside a game window instead of letting it wander across screens.

There are also a couple of breaking changes to keep in mind. The tablet_map_to_mon and touch_map_to_mon directives have been dropped in favor of the more flexible devicerule system introduced in 0.16.1. If you're using custom tablet or touch mapping, you'll need to migrate your configuration. It's a minor headache, but it consolidates input handling under a single, more capable framework.

Why This Matters for Wayland Users

MangoWM sits in a niche that's been underserved for years. It's not trying to compete with i3 or Sway in terms of raw features, and it's certainly not going to match the polish of a major desktop environment. What it does is take the dwl philosophy and stretch it into something that handles modern workflows. Tags with independent layouts, scratchpads, swallow, animations, and IPC are all there. It's also built on wlroots, which means you're not locked into a proprietary protocol.

The project has a steady release cadence throughout 2025, with versions rolling out roughly every week or two. That pace suggests an active community, and the source material backs it up. The release notes list contributors like @Greensky-gs and @Wateir, and the repository has over 2,000 commits on main. It's also packaged in Arch Linux, Fedora, Gentoo, Guix, NixOS, openSUSE, and a couple of other distributions, so you don't have to build it from scratch if you'd rather not.

The codebase reorganization in 0.17.0 is worth a mention. The team has restructured the C project to be LSP-friendly, which should make contributions easier for new developers. That's not just a nice-to-have; it's a practical step toward long-term sustainability. A project that's hard to work on will eventually stall out, no matter how popular it becomes.

The Catch

It's a rather niche tool for a rather niche audience. If you're used to a full desktop environment with system trays, network managers, and automatic window spawning, you're not going to find any of that here. MangoWM is a compositor, not a desktop. You'll need to bring your own panels, status bars, and launchers.

That's not a flaw, but it's a commitment. The project's tagline is "dwm but Wayland," and that's exactly what you're getting: a tiling window manager that expects you to know what you're doing. It's fast, it's minimal, and it's opinionated. If that sounds like your kind of setup, you'll find it's one of the more mature options in the dwl-derived space.

The maintainers have been clear about stability being the priority. Breaking changes will be minimal going forward, and the feature set is locked down. That's a good sign for anyone looking to adopt it as a daily driver.

Where to Find It

The source code is available on GitHub under the GPL-3.0 license. If you want to follow along with development or report issues, the repository is active, and the maintainers are responsive.

Head here to download the release.