Syntax.Codec

Codecs for the versioned syntax artifacts.

Decoding turns the compiler’s exports into the typed Syntax vocabularies, rejecting wrong schema tags, malformed shapes, and spans that invert or reach past the embedded source with one structured error; encoding is the exact inverse, re-emitting identical bytes.

The printer here matches the compiler’s export format precisely: two-space indent, one element per line, a space after each key’s colon, [] and {} for empty collections, insertion-order object keys, and the shared escape discipline (named escapes plus lowercase \u00xx for other control characters). Field order is part of the schema: envelope keys are in declaration order, node objects are alphabetical, and optional fields are omitted rather than null.

Types

`CodecError`

type CodecError = CodecError { path: String, reason: String }

A structured decode refusal: the path of the offending value inside the document and the reason it was rejected.

`TokensDoc`

type TokensDoc = TokensDoc {
  schema: String,
  compiler: String,
  source: SourceFile,
  raw: List(Token),
  parse: List(Token),
  trivia: List(Trivia)
}

A decoded prism-syntax-tokens-v1 document: the envelope identity, the embedded source, the raw and post-layout token streams, and the trivia events, all in stream order.

`SurfaceDoc`

type SurfaceDoc = SurfaceDoc {
  schema: String,
  compiler: String,
  source: SourceFile,
  items: List(Item)
}

A decoded prism-surface-syntax-v1 document: the envelope identity, the embedded source, and the ordered item list of the parsed file.

Effects

`Decode`

effect Decode
  never fail_decode(CodecError) : a

Functions and Values

`tokens_schema`

tokens_schema : () -> String

The schema tag of the token-stream artifact.

`codec_error_message`

codec_error_message : (Syntax.Codec.CodecError) -> String

Render a codec error as one line.

`decode_tokens`

decode_tokens : (String) -> Result(Syntax.Codec.TokensDoc, Syntax.Codec.CodecError)

Decode a prism-syntax-tokens-v1 document from its exact bytes. A wrong or missing schema tag, malformed JSON, missing field, or invalid span is a structured CodecError, never a partial document.

`canonical_json`

canonical_json : (Json.Json) -> String

Render a JSON value in the export layout the encoders here use, for tools that derive a document from an artifact instead of re-encoding a decoded one.

`encode_tokens`

encode_tokens : (Syntax.Codec.TokensDoc) -> String

Encode a token document back to the exact artifact bytes the compiler emits: encode_tokens after decode_tokens is byte-identity on every well-formed export.

`surface_schema`

surface_schema : () -> String

The schema tag of the surface-syntax artifact.

`encode_surface`

encode_surface : (Syntax.Codec.SurfaceDoc) -> String

Encode a surface document back to the exact artifact bytes the compiler emits: encode_surface after decode_surface is byte-identity on every well-formed export.

`decode_surface`

decode_surface : (String) -> Result(Syntax.Codec.SurfaceDoc, Syntax.Codec.CodecError)

Decode a prism-surface-syntax-v1 document from its exact bytes. A wrong or missing schema tag, malformed JSON, missing field, or invalid span is a structured CodecError, never a partial document.

`decode_diagnostics`

decode_diagnostics : (String) -> Result(Syntax.Diagnostic.DiagnosticsDoc, Syntax.Codec.CodecError)

Decode a prism-syntax-diagnostics-v1 document. Total: every malformed input lands on a CodecError naming the JSON path that refused.

`encode_diagnostics`

encode_diagnostics : (Syntax.Diagnostic.DiagnosticsDoc) -> String

Encode a diagnostics document to the canonical artifact bytes, the exact inverse of decode_diagnostics on every Rust-produced document.

`run_decode`

run_decode : forall e0 a. (() -> a ! {Syntax.Codec.Decode, e0}) -> Result(a, Syntax.Codec.CodecError) ! {e0}

Run a decoding computation, turning a fail_decode into Err and a completed decode into Ok.

`decode_resolved`

decode_resolved : (String) -> Result(Syntax.Resolved.ResolvedDoc, Syntax.Codec.CodecError)

Decode a prism-resolved-syntax-v1 document. Total: every malformed input lands on a CodecError naming the JSON path that refused. The node tree decodes recursively; a child list is bounded by the input’s own nesting.

`encode_resolved`

encode_resolved : (Syntax.Resolved.ResolvedDoc) -> String

Keyboard shortcuts

Prism