Problem
gnit operates in a crowded design space — git submodules, git-subrepo, meta, and (adjacent) jujutsu all touch multi-repo / multi-workspace workflows. The README explains what gnit does, but doesn't clearly position when to reach for gnit over the alternatives — which is the first question an evaluating user asks ("why not just use submodules / subrepo?").
Proposal
Add a short prior-art / comparison section (a table works well) to the README: for each of submodules, git-subrepo, meta — and a note on jj — one line on what it does and the specific case where gnit's "one change, one pin, one push across independent repos" model is the better fit.
Why it matters
Reduces "why not X?" friction and makes the differentiator legible up front. Pure docs — no code change.
Problem
gnit operates in a crowded design space — git submodules,
git-subrepo,meta, and (adjacent) jujutsu all touch multi-repo / multi-workspace workflows. The README explains what gnit does, but doesn't clearly position when to reach for gnit over the alternatives — which is the first question an evaluating user asks ("why not just use submodules / subrepo?").Proposal
Add a short prior-art / comparison section (a table works well) to the README: for each of submodules,
git-subrepo,meta— and a note on jj — one line on what it does and the specific case where gnit's "one change, one pin, one push across independent repos" model is the better fit.Why it matters
Reduces "why not X?" friction and makes the differentiator legible up front. Pure docs — no code change.