daml-syntax

daml-syntax is the shared parsed-source surface for Daml tools in this workspace. It owns source presentation around daml-parser: diagnostics, line mapping, UTF-16 offsets, token/trivia access, and parser span conversion.

Public dependencies

daml-parser and text-size are intentional public dependencies: parser AST, lexer, and layout types plus TextRange/TextSize appear in this crate's public API. SemVer for daml-syntax includes compatible major versions of those crates when upgrading.

Coordinate safety

Line and column APIs use distinct coordinate newtypes: LineNumber, ByteColumn, CharColumn, ByteOffset, and Utf16Offset. One-based coordinates reject zero through try_new and TryFrom<usize>, while raw coordinate extraction uses usize::from(coordinate). UTF-16 column lookup returns CoordinateRangeError instead of clamping lines or columns past the source. UTF-16 slices are returned as Utf16Range values with named start()/end() accessors. Diagnostic end columns use DiagnosticEndColumn: non-empty same-line diagnostic spans carry an exclusive 1-based Unicode-scalar end column, while multi-line and empty spans are named separately so callers cannot collapse them accidentally.

Quick start

use daml_syntax::{LineNumber, SourceFile};

let source = "module M where\nfoo : Int\nfoo = 1\n";
let file = SourceFile::parse(source);

assert!(file.diagnostics().is_empty());
assert_eq!(file.module().name, "M");
assert_eq!(file.line_index().line_col(0.into()).line, LineNumber::new(1));

For callers that only need lexer, trivia, or laid-out token facts, use the lighter tokenized surface:

use daml_syntax::SourceTokens;

let source = "module M where\nfoo : Int\nfoo = 1\n";
let tokens = SourceTokens::lex(source);

assert!(tokens.lex_errors().is_empty());
assert!(!tokens.laid_out_tokens().is_empty());

README examples are included in hidden rustdoc items and compile-tested via cargo test -p daml-syntax --doc.