Crate ox_mf2_parser

Expand description

High performance MessageFormat 2 parser core.

This crate provides the Rust core for parsing Unicode MessageFormat 2.0 messages. It builds a recovering, lossless concrete syntax tree (CST), materialises diagnostics, optionally lowers semantic records, and can encode parse results into the crate’s Binary AST snapshot format.

§Entry points

Use parse_message for a one-shot parse of an in-memory message. Use parse_source when the caller owns a SourceStore, and parse_source_session with ParseWorkspace when repeated parsing should reuse allocations.

§API layers

The crate root re-exports the supported parser, CST, diagnostic, semantic, source, and workspace types. The snapshot module is the public namespace for Binary AST snapshot encoding, decoding, and zero-copy views.

§Stability

This crate is pre-1.0. Public APIs are intended to be useful and documented, but minor releases may still refine names, shape, and snapshot details while MessageFormat 2 integration work continues. Error code values are exposed so language bindings can map failures without parsing display strings.

§Example

use ox_mf2_parser::parse_message;

let result = parse_message("Hello, {$name}!");
assert!(result.diagnostics.is_empty());
assert!(result.cst.node_count() > 0);

Re-exports§

pub use snapshot::decode_snapshot;
pub use snapshot::decode_snapshot_owned;
pub use snapshot::parse_batch_result_to_snapshot;
pub use snapshot::parse_batch_to_snapshot;
pub use snapshot::parse_message_to_snapshot;
pub use snapshot::parse_result_to_snapshot;
pub use snapshot::parse_session_to_snapshot;
pub use snapshot::parse_source_to_snapshot;
pub use snapshot::BatchSnapshotResult;
pub use snapshot::DecodeError;
pub use snapshot::DecodeErrorCode;
pub use snapshot::RootId;
pub use snapshot::SectionKind;
pub use snapshot::SnapshotOptions;
pub use snapshot::SnapshotResult;
pub use snapshot::SnapshotSourceMetadata;
pub use snapshot::SnapshotView;
pub use snapshot::SnapshotViewOwned;
pub use snapshot::SnapshotWriteError;
pub use snapshot::SnapshotWriteErrorCode;
pub use snapshot::SourceTextUnavailable;

Modules§

snapshot: Binary AST snapshot writer, decoder, and accessor.

Structs§

BatchParseItem
BatchParseOptions
BatchParseResult
CstCapacity: Pre-sizing hint for CstTables::reserve.
CstChildren: Iterator over a contiguous edge slice yielding CstChild values.
CstNodeTokens: Convenience iterator that yields only the token children of a node.
CstNodeView: Lazy node view: holds the node id + a reference back to the owning view.
CstTables: Flat indexed CST tables — the Phase 1 parser’s primary output.
CstTokenView: Lazy token view.
CstTriviaRange: Iterator over a compact trivia range belonging to a token.
CstTriviaView: Lazy trivia view.
CstView: Lightweight accessor over a parsed CstTables.
Diagnostic: Owned public diagnostic.
DiagnosticIter
DiagnosticLabel
DiagnosticRef: Lightweight reference-style view over a single diagnostic record.
DiagnosticView: Borrowed diagnostics view tied to a crate::ParseWorkspace.
EdgeId: Identifier for an edge entry in crate::CstTables.
NodeId: Identifier for a node in crate::CstTables.
ParseCapacity: Pre-sizing hint passed to ParseWorkspace::with_capacity.
ParseInput: Single-source batch input.
ParseOptions: Parser knobs. See design/002 for the rationale.
ParseResult: Owned parse result. Detached from any workspace.
ParseSessionResult: Borrowed parse result. Lives until the next workspace clear() / reset().
ParseWorkspace: Reusable workspace for repeated parse, batch parse, benchmarks, and LSP.
SemanticModel: Optional semantic lowering result.
SemanticView: Borrowed view onto a SemanticModel.
SourceFile: Owned source file registered in SourceStore.
SourceFileInput: Public input used to register a source file with SourceStore.
SourceId: Identifier for a source file inside crate::SourceStore.
SourceLocation: Resolved 1-based line/column pair derived from SourceStore.
SourceStore: Source ownership for single parse, batch parse, diagnostics, and Phase 2 snapshot roots.
Span: UTF-8 byte span [start, end) into the source text.
TokenId: Identifier for a token record in crate::CstTables.
TriviaId: Identifier for a trivia record in crate::CstTables.

Enums§

BatchExecution: Batch execution mode. Phase 1 only implements BatchExecution::Sequential; requesting any other mode returns a BatchParseResult whose BatchParseResult::execution is Sequential and whose BatchParseResult::degraded flag is set so callers can observe the fallback. The Parallel variant is reserved for a future milestone.
BindingValidationErrorCode: Binding input validation failures not better expressed as TypeError / RangeError.
CstChild: Reference from an edge to either a node or a token child.
DiagnosticCode: Diagnostic code. The catalog is part of the parser’s snapshot compatibility surface — do not reuse a code for a different message.
DiagnosticSeverity: Severity of a diagnostic. Compact u8 representation; numeric ordering is used by tooling layers but the parser itself does not rely on it.
InitializationErrorCode: Binding runtime initialization failures.
MessageMode: Syntactic message mode (simple-message vs complex-message).
OxMf2ErrorDomain: Domain classification for a unified API error code.
SemanticMessageKind: Data-model message kind (PatternMessage vs SelectMessage).
SourceStoreError: Errors returned by SourceStore::try_add.
SourceTextErrorCode: Source text access failures after parsing or decoding.
SyntaxKind: Shared classification for nodes, tokens, trivia, errors, and missing productions. Compact u16 representation that is stored directly in node, token, and trivia records.

Constants§

NONE_U32: Sentinel value used for optional table references.
OX_MF2_API_ERROR_MIN: Minimum value for ox-mf2 API error codes. Values below this are reserved.
OX_MF2_BINDING_VALIDATION_ERROR_MAX
OX_MF2_BINDING_VALIDATION_ERROR_MIN
OX_MF2_DECODE_ERROR_MAX
OX_MF2_DECODE_ERROR_MIN
OX_MF2_INITIALIZATION_ERROR_MAX
OX_MF2_INITIALIZATION_ERROR_MIN
OX_MF2_SNAPSHOT_WRITE_ERROR_MAX
OX_MF2_SNAPSHOT_WRITE_ERROR_MIN
OX_MF2_SOURCE_TEXT_ERROR_MAX
OX_MF2_SOURCE_TEXT_ERROR_MIN

Functions§

ox_mf2_error_code_name: Stable programmatic name for an API error code.
ox_mf2_error_domain: Classify an API error code into its domain.
parse_batch: Parse inputs sequentially and return owned results in input order.
parse_message: One-shot convenience parser. Parses source directly and returns an owned ParseResult. The success path does not allocate a temporary SourceStore; malformed inputs build one only when diagnostics need line/column materialisation.
parse_source: Parse source_id from sources and return an owned ParseResult.
parse_source_session: Reuse workspace to parse source_id. Returns a ParseSessionResult borrowed from the workspace.

Type Aliases§

OxMf2ErrorCode: Unified numeric API error code exposed to language bindings.