MOPAC/internal/config/toml.go

// Package config implements the harness.toml surface for the MOPAC harness.
//
// The TOML parser in toml.go is a deliberate stdlib-only subset: tables,
// bare keys, basic strings, integers, booleans, and (single- or multi-line)
// string/int/bool arrays. Anything richer is a parse error so misconfigured
// files fail loudly instead of silently.
package config

import (
	"fmt"
	"os"
	"sort"
	"strings"
)

// TOMLDoc is the parsed representation of harness.toml: nested tables are
// TOMLDoc values; leaves are string, int64, bool, or []any of those.
type TOMLDoc map[string]any

// Table returns the nested table at the given key path, or an empty doc.
func (d TOMLDoc) Table(keys ...string) TOMLDoc {
	cur := any(d)
	for _, k := range keys {
		m, ok := cur.(TOMLDoc)
		if !ok {
			return TOMLDoc{}
		}
		cur = m[k]
	}
	if m, ok := cur.(TOMLDoc); ok {
		return m
	}
	return TOMLDoc{}
}

func (d TOMLDoc) String(key string) (string, bool) {
	v, ok := d[key].(string)
	return v, ok
}

func (d TOMLDoc) Int(key string) (int64, bool) {
	v, ok := d[key].(int64)
	return v, ok
}

func (d TOMLDoc) Bool(key string) (bool, bool) {
	v, ok := d[key].(bool)
	return v, ok
}

func (d TOMLDoc) StringList(key string) ([]string, bool) {
	raw, ok := d[key].([]any)
	if !ok {
		return nil, false
	}
	out := make([]string, 0, len(raw))
	for _, e := range raw {
		s, ok := e.(string)
		if !ok {
			return nil, false
		}
		out = append(out, s)
	}
	return out, true
}

// Keys returns the table's key names in sorted order.
func (d TOMLDoc) Keys() []string {
	out := make([]string, 0, len(d))
	for k := range d {
		out = append(out, k)
	}
	sort.Strings(out)
	return out
}

// ParseTOML parses the harness.toml subset.
func ParseTOML(src string) (TOMLDoc, error) {
	lines, err := logicalLines(src)
	if err != nil {
		return nil, err
	}
	root := TOMLDoc{}
	cur := root
	for _, ln := range lines {
		if strings.HasPrefix(ln.text, "[") {
			path, err := parseTableName(ln.text, ln.no)
			if err != nil {
				return nil, err
			}
			cur, err = descend(root, path, ln.no)
			if err != nil {
				return nil, err
			}
			continue
		}
		if err := parseKeyValue(cur, ln.text, ln.no); err != nil {
			return nil, err
		}
	}
	return root, nil
}

type logicalLine struct {
	no   int
	text string
}

// logicalLines joins physical lines into logical ones: comments are stripped
// (quote-aware), blank lines dropped, and unbalanced array brackets pull the
// next line in so multi-line arrays work. Strings may not span lines.
func logicalLines(src string) ([]logicalLine, error) {
	var out []logicalLine
	var buf strings.Builder
	curNo := 0
	depth := 0
	for i, raw := range strings.Split(src, "\n") {
		no := i + 1
		line := stripComment(strings.TrimSuffix(raw, "\r"))
		if buf.Len() == 0 {
			trimmed := strings.TrimSpace(line)
			if trimmed == "" {
				continue
			}
			curNo = no
			buf.WriteString(trimmed)
		} else {
			buf.WriteString(" ")
			buf.WriteString(strings.TrimSpace(line))
		}
		inStr, esc := false, false
		d := 0
		for _, r := range line {
			switch {
			case esc:
				esc = false
			case inStr && r == '\\':
				esc = true
			case r == '"':
				inStr = !inStr
			case !inStr && r == '[':
				d++
			case !inStr && r == ']':
				d--
			}
		}
		if inStr {
			return nil, fmt.Errorf("harness.toml:%d: unterminated string", no)
		}
		if depth+d < 0 {
			return nil, fmt.Errorf("harness.toml:%d: unbalanced ']'", no)
		}
		depth += d
		if depth == 0 {
			out = append(out, logicalLine{no: curNo, text: buf.String()})
			buf.Reset()
		}
	}
	if depth != 0 {
		return nil, fmt.Errorf("harness.toml:%d: unterminated array (missing ']')", curNo)
	}
	return out, nil
}

// stripComment removes a trailing # comment, ignoring # inside strings.
func stripComment(line string) string {
	inStr, esc := false, false
	for i, r := range line {
		switch {
		case esc:
			esc = false
		case inStr && r == '\\':
			esc = true
		case r == '"':
			inStr = !inStr
		case r == '#' && !inStr:
			return line[:i]
		}
	}
	return line
}

func parseTableName(text string, no int) ([]string, error) {
	if !strings.HasSuffix(text, "]") {
		return nil, fmt.Errorf("harness.toml:%d: malformed table header %q", no, text)
	}
	inner := strings.TrimSuffix(strings.TrimPrefix(text, "["), "]")
	if inner == "" {
		return nil, fmt.Errorf("harness.toml:%d: empty table name", no)
	}
	parts := strings.Split(inner, ".")
	for _, p := range parts {
		if !isBareKey(p) {
			return nil, fmt.Errorf("harness.toml:%d: invalid table name segment %q", no, p)
		}
	}
	return parts, nil
}

func descend(root TOMLDoc, path []string, no int) (TOMLDoc, error) {
	cur := root
	for _, k := range path {
		if existing, ok := cur[k]; ok {
			sub, ok := existing.(TOMLDoc)
			if !ok {
				return nil, fmt.Errorf("harness.toml:%d: key %q already has a value, cannot become table", no, k)
			}
			cur = sub
			continue
		}
		sub := TOMLDoc{}
		cur[k] = sub
		cur = sub
	}
	return cur, nil
}

func parseKeyValue(m TOMLDoc, text string, no int) error {
	eq := strings.IndexByte(text, '=')
	if eq < 0 {
		return fmt.Errorf("harness.toml:%d: expected key = value, got %q", no, text)
	}
	key := strings.TrimSpace(text[:eq])
	if strings.HasPrefix(key, `"`) {
		// Quoted key (e.g. `"In Progress" = "Done"` in a status map):
		// the key must be exactly one quoted string.
		q, err := parseStringValue(key, no)
		if err != nil {
			return fmt.Errorf("harness.toml:%d: invalid quoted key: %w", no, err)
		}
		key = q
	} else if !isBareKey(key) {
		return fmt.Errorf("harness.toml:%d: invalid key %q", no, key)
	}
	valStr := strings.TrimSpace(text[eq+1:])
	if valStr == "" {
		return fmt.Errorf("harness.toml:%d: missing value for key %q", no, key)
	}
	val, err := parseValue(valStr, no)
	if err != nil {
		return err
	}
	if _, exists := m[key]; exists {
		return fmt.Errorf("harness.toml:%d: duplicate key %q", no, key)
	}
	m[key] = val
	return nil
}

func isBareKey(s string) bool {
	if s == "" {
		return false
	}
	for _, r := range s {
		switch {
		case r >= 'a' && r <= 'z', r >= 'A' && r <= 'Z', r >= '0' && r <= '9', r == '_', r == '-':
		default:
			return false
		}
	}
	return true
}

func parseValue(v string, no int) (any, error) {
	switch {
	case strings.HasPrefix(v, `"`):
		return parseStringValue(v, no)
	case strings.HasPrefix(v, "["):
		return parseArrayValue(v, no)
	default:
		return parseScalar(v, no)
	}
}

func parseScalar(v string, no int) (any, error) {
	switch v {
	case "true":
		return true, nil
	case "false":
		return false, nil
	}
	n, err := parseTOMLInt(v)
	if err != nil {
		return nil, fmt.Errorf("harness.toml:%d: unsupported value %q (want string, int, bool, or array)", no, v)
	}
	return n, nil
}

func parseTOMLInt(v string) (int64, error) {
	s := strings.ReplaceAll(v, "_", "")
	return parseSignedInt(s)
}

func parseSignedInt(s string) (int64, error) {
	var n int64
	i := 0
	neg := false
	if i < len(s) && (s[i] == '+' || s[i] == '-') {
		neg = s[i] == '-'
		i++
	}
	if i >= len(s) {
		return 0, fmt.Errorf("not an integer: %q", s)
	}
	for ; i < len(s); i++ {
		if s[i] < '0' || s[i] > '9' {
			return 0, fmt.Errorf("not an integer: %q", s)
		}
		n = n*10 + int64(s[i]-'0')
		if n < 0 {
			return 0, fmt.Errorf("integer overflow: %q", s)
		}
	}
	if neg {
		n = -n
	}
	return n, nil
}

func parseStringValue(v string, no int) (string, error) {
	if len(v) < 2 {
		return "", fmt.Errorf("harness.toml:%d: unterminated string %q", no, v)
	}
	var b strings.Builder
	i := 1
	for i < len(v) {
		c := v[i]
		if c == '"' {
			if rest := strings.TrimSpace(v[i+1:]); rest != "" {
				return "", fmt.Errorf("harness.toml:%d: unexpected trailing content after string: %q", no, rest)
			}
			return b.String(), nil
		}
		if c != '\\' {
			b.WriteByte(c)
			i++
			continue
		}
		if i+1 >= len(v) {
			return "", fmt.Errorf("harness.toml:%d: dangling escape in %q", no, v)
		}
		switch v[i+1] {
		case '"':
			b.WriteByte('"')
		case '\\':
			b.WriteByte('\\')
		case 'n':
			b.WriteByte('\n')
		case 't':
			b.WriteByte('\t')
		case 'r':
			b.WriteByte('\r')
		default:
			return "", fmt.Errorf("harness.toml:%d: unsupported escape \\%c", no, v[i+1])
		}
		i += 2
	}
	return "", fmt.Errorf("harness.toml:%d: unterminated string", no)
}

func parseArrayValue(v string, no int) ([]any, error) {
	if !strings.HasSuffix(v, "]") {
		return nil, fmt.Errorf("harness.toml:%d: unterminated array %q", no, v)
	}
	inner := strings.TrimSpace(v[1 : len(v)-1])
	if inner == "" {
		return []any{}, nil
	}
	var out []any
	elems := splitTopLevel(inner, ',')
	if len(elems) > 0 && strings.TrimSpace(elems[len(elems)-1]) == "" {
		elems = elems[:len(elems)-1] // tolerate a trailing comma
	}
	for _, elem := range elems {
		elem = strings.TrimSpace(elem)
		if elem == "" {
			return nil, fmt.Errorf("harness.toml:%d: empty array element in %q", no, v)
		}
		val, err := parseValue(elem, no)
		if err != nil {
			return nil, err
		}
		out = append(out, val)
	}
	return out, nil
}

// splitTopLevel splits on sep outside of double-quoted strings.
func splitTopLevel(s string, sep rune) []string {
	var out []string
	var cur strings.Builder
	inStr, esc := false, false
	for _, r := range s {
		switch {
		case esc:
			esc = false
			cur.WriteRune(r)
		case inStr && r == '\\':
			esc = true
			cur.WriteRune(r)
		case r == '"':
			inStr = !inStr
			cur.WriteRune(r)
		case r == sep && !inStr:
			out = append(out, cur.String())
			cur.Reset()
		default:
			cur.WriteRune(r)
		}
	}
	out = append(out, cur.String())
	return out
}

// Load reads, parses, and validates harness.toml at path.
func Load(path string) (*Config, error) {
	data, err := os.ReadFile(path)
	if err != nil {
		return nil, fmt.Errorf("config: read %s: %w", path, err)
	}
	doc, err := ParseTOML(string(data))
	if err != nil {
		return nil, fmt.Errorf("config: %w", err)
	}
	cfg := Default()
	if err := cfg.apply(doc); err != nil {
		return nil, fmt.Errorf("config: %w", err)
	}
	if err := cfg.Validate(); err != nil {
		return nil, fmt.Errorf("config: %w", err)
	}
	return cfg, nil
}