The open-source world is rarely short of disagreements about when to let old code die, but a debate now unfolding in the Debian developers mailing list has unusually high stakes for a wide range of Linux users. The Debian GNOME team has announced its intention to remove the GTK2 graphical toolkit from Debian 14, codenamed "Forky", before that release ships in 2027. The decision has triggered genuine alarm among developers whose projects still depend on code that, by any measure, has had a very long run.
Debian is one of the most widely deployed Linux distributions in the world, underpinning Ubuntu and dozens of derivative systems. When it drops a library, the effect ripples outward. GTK2 was first released in March 2002, which means it has been in continuous use for nearly a quarter of a century. The GNOME project formally declared it end-of-life in December 2020, with the final release, version 2.24.33, published shortly thereafter. The case for removal is, on its surface, straightforward: the library has received no upstream maintenance for more than five years and, critically, lacks native Wayland support and does not integrate properly with modern HiDPI display scaling.
Debian would not be acting alone. As The Register reports, GTK2 has already been stripped from Red Hat Enterprise Linux 10, SUSE Linux Enterprise, openSUSE 16, Arch Linux, and others. Arch moved the GTK2 package and its dependents out of official repositories and into the Arch User Repository in October 2025, while RHEL dropped support with the release of RHEL 10 in May 2025. The direction of travel across the Linux ecosystem is not ambiguous.
The complication, however, is significant. Quite a lot of applications still use GTK2, with the Debian project linking to a list of 139 affected packages. Among the more prominent are the FreePascal compiler and its integrated development environment, Lazarus. This dependency has caused some alarm in FreePascal's community forums, where the team is discussing possible resolutions, including creating and maintaining its own packages. That is a substantial undertaking for a small project.
The more intriguing development comes from an unexpected quarter: a professional audio workstation. In early 2024, the Ardour digital audio workstation software introduced its own stripped-down version of the GTK2 toolkit, dubbed YTK, specifically because Linux distributions were removing GTK2 libraries from their repositories, and the project needed a way to keep building without that dependency. Over time, YTK picked up extra capabilities, including touch input support and improved macOS compatibility, before Ardour moved to rely on it exclusively. With Ardour 9.0 released earlier this month and version 9.2 following days later, the project is actively maintained and its toolkit fork is now battle-tested.
The argument for keeping GTK2 available, even in some form, deserves a fair hearing. Critics of the removal point out that GTK2 is stable, that it has not changed in years, and that the cost of packaging it is low relative to the disruption caused by its absence. Some community members argue that removing it will break applications that many users rely on, and that it does not hurt Debian to keep packaging software that is stable and unchanged. There is also the awkward reality that even the graphical Debian Installer itself still uses GTK2, a dependency that must be resolved before the removal can be completed.
There is a deeper philosophical point embedded in this dispute. The scenario of software being declared obsolete and unmaintained has been appearing with increasing frequency, and the pile of old software still in active use continues to grow faster than the pool of developers willing to port it to newer libraries. Forcing migrations is one way to accelerate modernisation; it is also a way to strand users who have no pressing reason to change something that works.
The strongest counterargument to outright removal is the precedent set by Ardour's YTK. There is a real possibility that multiple free and open-source projects could collaborate to make YTK into something more generally applicable, providing a maintained, community-supported fork that covers the needs of Lazarus, GKrellM, and the many smaller applications still living on the old toolkit. Whether that kind of cross-project coordination is achievable is another matter entirely.
The full list of affected Debian packages makes plain the scale of what is involved. Some projects, like the MATE and Xfce desktop environments, completed the migration to GTK3 years ago, though by all accounts it required substantial effort. Smaller, less-resourced projects have simply not had the capacity to follow. Developers have argued that porting such applications to GTK3 or GTK4 is often non-trivial, requiring major redesigns rather than mechanical updates, and in some cases is unlikely to happen upstream at all.
As a compromise, contributors have proposed moving GTK2 and its remaining reverse dependencies into side repositories rather than excising them from the official archive entirely. That approach mirrors what Arch Linux has done with its AUR, and it would at least preserve access for users who need these tools without requiring the Debian project to carry full maintenance responsibility. It seems likely that Debian developers will find some way to continue supporting applications that require GTK2, but users may have to look outside official repositories to get it.
The Debian 14 GTK2 question is, at its core, a resource-allocation problem wearing the clothes of a technical debate. Removing unmaintained code is sound engineering practice. Breaking working software for users who have no upgrade path is a real cost. The most defensible outcome sits somewhere in the middle: GTK2 shifted to a community-maintained side repository, with the Ardour team's YTK work serving as a foundation for projects like Lazarus to build on. That path requires collaboration across multiple small teams, which is genuinely difficult. But it is more honest about the trade-offs than simply declaring old code dead and walking away.