Summary #
A user-facing way to build a multibuffer from a chosen set of open buffers, not just :grep results.
Motivation #
:grep command is the only producer of multibuffers right now, but buf/new-multibuffer API is already general (arbitrary {:id :start :end} excerpts). A manual command lets users merge a few files into one linear editing surface.
Current behavior #
buf/new-multibuffer(buffer.janet) builds a multibuffer from excerpts; edit routing,:save-all,:open-excerpt, context, and syntax highlighting all work.- The prompt engine (prompt.janet) is single-select only.
Proposed approach #
- Add a multi-select mechanism: either a
prompt.janetmulti-select view (mark/unmark candidates, accept the marked set, generic + reusable) or a two-step mark/stitch flow (:bmarkthen:multibuffer). - The command collects the chosen buffers and calls
buf/new-multibufferwith an excerpt per buffer (whole file, first N lines, or a picked range).
Scope / notes #
Smaller than per-cursor selections; mostly prompt UI. Reconsider whether window splits (once done) are the better answer for "see/edit several files at once".