#!/bin/sh
# Runs the TopSpeed sidecar front-end (build/TSM2) over a corpus of
# .DEF/.MOD files and reports parse coverage.
#
# A *syntax* error is a Coco/R error (`'x' expected`, `invalid X`); the
# sidecar's own `not supported yet` (230) markers are the deliberate
# parse-now/lower-later discipline and do NOT count as failures.
#
# Usage: ./corpus.sh [corpus-dir]   (default: the TS-V3 tree)
set -e
cd "$(dirname "$0")"
TSM2="$(pwd)/build/TSM2"
CORPUS="${1:-$HOME/bin/Dos/M2/TS-V3}"
WORK="/tmp/opencode/tscorpus"
[ -x "$TSM2" ] || { echo "build/TSM2 missing -- run ./build.sh"; exit 1; }
[ -d "$CORPUS" ] || { echo "corpus not found: $CORPUS"; exit 1; }

rm -rf "$WORK"; mkdir -p "$WORK"
: > "$WORK/results.txt"
n=0
find "$CORPUS" \( -iname '*.DEF' -o -iname '*.MOD' \) 2>/dev/null | sort |
while read -r f; do
  n=$((n+1))
  w="$WORK/f$n"; mkdir -p "$w"; base=$(basename "$f")
  cp "$f" "$w/$base"
  ( cd "$w" && "$TSM2" "$base" >/dev/null 2>&1 )
  lst="$w/$(echo "$base" | sed 's/\.[^.]*$//').LST"
  if [ ! -f "$lst" ]; then echo "NOLST|$f" >> "$WORK/results.txt"; continue; fi
  if grep -qE "^\*\*\*\*\*.*(expected|invalid [A-Z])" "$lst"; then
    m=$(grep -E "^\*\*\*\*\*.*(expected|invalid [A-Z])" "$lst" | head -1 |
        sed -E 's/^\*\*\*\*\* *[^ ]* *//; s/^\*\*\*\*\* *//')
    echo "SYN|$m|$f" >> "$WORK/results.txt"
  else
    echo "OK|0|$f" >> "$WORK/results.txt"
  fi
done

total=$(wc -l < "$WORK/results.txt")
ok=$(grep -c '^OK' "$WORK/results.txt" || true)
syn=$(grep -c '^SYN' "$WORK/results.txt" || true)
echo "corpus: $CORPUS"
echo "total                      : $total"
echo "parse-OK (semantic/230 only): $ok"
echo "syntax-bad                 : $syn"
echo "=== syntax-error kinds ==="
grep '^SYN' "$WORK/results.txt" | cut -d'|' -f2 |
  sed -E 's/[0-9]+/N/g' | sort | uniq -c | sort -rn
echo "=== syntax-bad files ==="
grep '^SYN' "$WORK/results.txt" | cut -d'|' -f3 | sed "s|$CORPUS/||"