Podcasts 3000

Minimal, offline-first podcast player built with Go, pure-Go SQLite, and vanilla HTML5.

Run the desktop application

go run -tags 'production webkit2_41' ./cmd/podcasts3000

This launches the Wails desktop window. Data is stored under the platform user configuration directory. Set PODCASTS3000_DATA_DIR to choose another location.

Linux desktop builds require the WebKitGTK development/runtime libraries required by Wails. Windows builds use WebView2.

On Linux, the production and webkit2_41 build tags are required. The repository includes WebKitGTK 4.1 support; verify it is installed with:

pkg-config --modversion webkit2gtk-4.1

If Go is not available in a new shell, enable the installed toolchain first:

export PATH=/usr/local/go/bin:$PATH
go version

The package path starts with ./ (a relative path), not /.:

./cmd/podcasts3000

go run uses a temporary executable and does not leave a binary in the repository. Use make build when a persistent podcasts3000 binary is required. make run-dev is available for Wails development builds.

Use make release to create release/podcasts3000-linux-amd64 and release/podcasts3000-windows-amd64.exe. The Windows release is linked with the GUI subsystem, so launching it does not create a console window.

GitHub Actions runs the test, Wails-tagged runtime test, and vet targets on pushes and pull requests. Pushing a v* tag builds both release executables and publishes them to a GitHub release.

The Wails frontend source lives exclusively under cmd/podcasts3000/frontend and is embedded into the binary. The root web/ directory is not used and may remain absent.

Verify

go test ./...
go vet ./...

For a Wails-tagged integration compile and embedded-frontend smoke test:

make runtime-test

This validates the production Wails package, embedded assets, and the bound podcasts.App namespace. A full GUI interaction test still requires a graphical desktop session; manually verify subscribing with a reachable RSS feed after launching make run.

The backend is organized around internal/podcasts.App and is bound directly to the embedded Wails frontend. No Node toolchain is required.

Disclaimer

This application was produced using AI (Gemini, Copilot).