hook-dsh-normalize-ellipsis

Family position (the @-sentence Hook @ DSH @ Normalize @ Ellipsis): a hook child of the plugin-dsh-factory service and the hook-dsh-core machinery; the third of the six stream normalizer siblings:

A non-manifest factory consumer: it injects ["pluginFactory"] and uses only State (cell unwrap + shared Ledger/Enabled mappings + its own fields) and Append; the config is composed by the factory's standalone Schema helper with shared: false - the minimal block, no fs/observed dead fields. It touches no files, so fs/write-intent and fs/observed never see it; it wraps the downstream result and always calls next(), so it composes with other llm/stream listeners regardless of registration order.

Text processors collapse the typographic three-dot run into a single code point - U+2026, the horizontal ellipsis. In code blocks, commands and diffs the collapsed form is simply wrong: three literal dots are what every parser, shell and search expects. This flavor makes the ASCII spelling the one that reaches the transcript.

The transform - exactly the core's Ellipsis class (@playform/hook-dsh-core's Normalize/Ellipsis), applied per text segment through the core's generic class-to-string Replace:

U+2026 (horizontal ellipsis - the typographic three-dot run collapsed into one code point by text processors) becomes the config replacement (default ..., the ASCII three-dot sequence - the ellipsis's own ASCII spelling). No context rules - hermes-style whole-chunk replacement, no lookahead; chunk-boundary-safe. The replacement is applied with a function replacer, so a custom replacement containing $ patterns is inserted literally. Replaced characters are counted per stream for the ledger line - each replaced ellipsis code point counts once, even though the default replacement emits three characters.

Volatile cells commit without remounting the plugin; the factory's State builder unwraps them defensively. Example cordis.patch.yml row:

One stream, one transformation. The model emits the collapsed typographic ellipsis; the live UI and the transcript receive three ASCII dots:

Before → after - the model stream, then what reaches the transcript

The single U+2026 code point becomes the three-dot ASCII sequence ...; dots that were already ASCII are untouched. Each replaced code point counts once in the ledger even though it expands to three characters - which is why one ellipsis can show up as normalized 3 ellipsis char(s) for three collapsed code points, not nine output characters (the count line shown below). The replacement is hot-editable: change it in the config (say to . . .) and the very next stream picks it up with no remount. The same pass runs over reasoning deltas when normalizeReasoning is on, and over tool-call arguments when the example patch enables normalizeToolArguments - with the edit name exempt and raw-marker calls passing through unnormalized.

Two lines, both written through the factory's Append (the hook-dsh-normalize-ellipsis: prefix is the logger's <State.Module>:; the durable file line is [<ISO>] <message>):

The activation line is written by apply(); the count line only follows a normal stream completion and only when N > 0 (a thrown-away stream writes no ledger line).

License: MIT.