Theia-Modula-2/grammars/P6/README.md

P6 Pascal validator pilot

p6_pascal.atg describes Wirth's Pascal-P6 (1980) compiler input. It shares the P4 skeleton (same repair pass: quoted ";" literals, terminated productions, no "" empties, no ident/token double definitions, no reserved words as productions, start production matching COMPILER, left-recursion removal, filled-in undefined nonterminals, merged duplicate branches, lenient list shapes) plus its own deltas, all diagnosed against fpc -Miso as oracle:

• EnumType ((red, green, blue)) alongside the other types.
• ParamType (array[expr] of T, expression bounds) merged with ArrayType into a single Bound form (SimpleType for array[boolean]-style index types, Expr [".." Expr] for ranges and computed bounds) — the two array[-led productions generated duplicate CASE labels and would not compile separately.
• ProcSection (proc-led procedural-type form) kept reachable via StructType.
• Label = pnumber | ident (enum case labels), with statement labels kept numeric-only — an optional [Label ":"] prefix commits on ident and then chokes on :=, so i := 1 would break; fpc rejects ident-statement-labels anyway.
• P4's preal/CONTEXT("..")/subrange TypeIdent upgrade carried over (floats, 0..255, N..M, dotted names).

The generated parser gets a strict EOF check (IF sym # 0 THEN SynError(0) after the start symbol in Parse, patched in by build.sh with a loud assertion): without it, the Coco/R driver never scans past the program's final ., silently ignoring trailing garbage — and worse, masking any failure that leaves a . behind (e.g. unhandled float syntax). This makes the validator marginally stricter than fpc, which accepts trailing junk after END. — degenerate input only, documented here.

Two benign LL(1) warning families remain (dangling-ELSE, loop-exit ";" greediness — both resolve correctly).

if/while/for branches take a single Statement (standard Pascal — plural branches greedily swallow the ; separating case items), and begin..end compounds are statements in their own right.

Build

./build.sh        # -> build/P6 (build/ is git-ignored)

Needs CR (Coco/R), gm2 -fiso, the parser.frm/scanner.frm frames and FileIO.def/.mod — all resolved from the environment with sensible defaults (CRFRAMES, M2LIB, CR, GM2).

Differential results (P6 vs fpc -Miso)

Accept battery (transferred from P4) + tests/deep.pas (enums, records, subrange-free shapes, nested routines, labels/goto, all statements): 0 false positives. Broken mutants: 0 missed, positions close.