We currently ship xdvdfs-web in the desktop app via Tauri, but this introduces a number of complexities:
Tauri is still heavy to run in contrast to native tools
It requires many dependencies, and requires local environment dependencies that make building xdvdfs more difficult
It doesn't integrate with the main CI pipelines, and the desktop pipelines take significantly longer to run than the others
There are platform-specific webkit bugs with WASM that cannot be worked around or controlled for
The pros of using Tauri are UI code sharing with xdvdfs-web, and having multi-platform support for free. In spite of these, the desktop platform is likely better served with a native UI that is not bound by the limitations of the filesystem access API.
xdvdfs-web should remain backend-agnostic with respect to how it accesses the filesystem, but the Tauri wrappers should be removed as part of this work.
We currently ship xdvdfs-web in the desktop app via Tauri, but this introduces a number of complexities:
The pros of using Tauri are UI code sharing with xdvdfs-web, and having multi-platform support for free. In spite of these, the desktop platform is likely better served with a native UI that is not bound by the limitations of the filesystem access API.
xdvdfs-web should remain backend-agnostic with respect to how it accesses the filesystem, but the Tauri wrappers should be removed as part of this work.