# Oberon-07 validator pilot

`oberon07_oberon.atg` is the installed Oberon-07 grammar plus 5 fixes
needed to build a working validator with Coco/R V1.53 (`CR`) and GNU
Modula-2 (`gm2 -fiso`), all diagnosed against `voc` as oracle:

1. `statement`: `assignment | ProcedureCall` factored into
   `AssignmentOrProcedureCall` (the unfactored form generated duplicate
   `CASE` labels — uncompilable parser).
2. `designator`: dropped `"(" qualident ")"` (hijacked call parens;
   `Out.String("hi")` failed with `ident expected`).
3. `qualident`: `[ ident "." ] ident` rewritten as
   `ident { "." ident }` (optional-prefix lookahead trap; same rewrite
   the M2 grammar uses).
4. `real` requires post-dot digits + `CONTEXT ("..")` on integers
   (the `2..3` lexer trap; same as the M2 grammar's documented fix).
5. Added `ForStatement` (`voc` accepts `FOR` as an extension although
   strict Oberon-07 dropped it; without it every `FOR` loop is a
   false positive).

Two benign LL(1) warnings remain (`DeclarationSequence` PROCEDURE
conflict, `qualident` dot conflict); the differential below shows they
do not bite real code.

## Build

```sh
./build.sh        # -> build/O07 (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 (O07 vs `voc`)

23 accept probes + `tests/deep.Mod` (records, pointers, nested
recursion, multidim arrays, procedure types, CASE ranges, hex/reals):
**0 false positives**. 8 broken mutants: **0 missed**, positions
within ~3 columns. `voc` quirks found on the way: bare `WITH` fields
rejected (only qualified `p.x` accepted); `FOR` accepted as an
extension; module `END`-name mismatch unreported by O07 (accepted gap).