YAML

Batch YAML parsing for Telo manifests. Multi-document-aware, returns plain JS values.

Why use this

• Multi-doc native — yaml.parseAllDocuments under the hood; single-doc input still lands in docs[0].
• Plain JS values — output is regular objects, arrays, and scalars — no proxy or wrapper types.
• Typed errors — invalid YAML surfaces as ERR_PARSE_FAILED with the original parser error (line/column when available) preserved on data.
• Batch, not stream — built as a Telo.Invocable because parsing needs the whole document; no stream-codec overhead.

Kinds

Kind	Purpose
Yaml.Parse	UTF-8 string in, { docs: unknown[] } out. Each entry mirrors the source document's top-level value.

Example

kind: Telo.Application
metadata: { name: parse-demo }
imports:
  Yaml: std/yaml@0.4.1
  Run: std/run@0.10.0
targets: [ !ref ParseExample ]
---
kind: Run.Sequence
metadata: { name: ParseExample }
steps:
  - name: parse
    invoke: { kind: Yaml.Parse }
    inputs:
      text: |
        kind: Telo.Library
        metadata: { name: example }
outputs:
  firstKind: ${{ steps.parse.result.docs[0].kind }}

Reference

• Yaml.Parse — parse a UTF-8 YAML string into one or more documents.

Why a dedicated module (and not a codec)?

The format-codec abstracts (Codec.Encoder / Codec.Decoder) are stream-oriented — they carry Stream<Uint8Array> end-to-end so manifests can pipe bytes without buffering. YAML parsing is fundamentally batch: the parser needs the whole document before it can build the syntax tree, so a stream-typed input/output adds nothing. Yaml.Parse is therefore a plain Telo.Invocable, not a Codec.Decoder.

Roadmap

• Yaml.Stringify (object → string) lands when the first consumer needs it.