Why GitHub Flow Isn't Ideal for RDK

[Ulrond]

While GitHub Flow offers a simple branching model, it doesn't fully align with the requirements of RDK development for scale:

• Limited Targeted Fixes: GitHub Flow's linear model lacks a built-in mechanism to address issues in specific released versions. This becomes critical for RDK, as platforms in the field (like Sky Q) often require targeted fixes after release.
• Frozen Vendor and Middleware Layers: In RDK projects, these layers often freeze after a certain point, with updates limited to security fixes. GitHub Flow's focus on continuous deployment doesn't accommodate this reality.

Key Considerations for the future

• Customer/In-Field Builds: RDK development needs to support targeted fixes for older versions of the codebase deployed in the field. This often requires branching from specific release points to apply isolated fixes.
• Release Cadence and Freezing: The release cadence for a layer often slows down after the initial launch, with layers freezing entirely at some point. A branching model should accommodate this to avoid disruptions.