// lib/parsers/plan-schema.mjs
// Plan schema parser — heading shape, per-step fields, ceremony, design part.
//
// The canonical step heading is `### Step N: <title>` (literal colon-space).
// Forbidden narrative drift formats (introduced in v1.8.0 to defend against
// Opus 4.7 schema-drift): `## Fase N`, `### Phase N`, `### Stage N`, `### Steg N`.
//
// plan_version 1.8 adds per-step field parsing (`parseStepFields`), the
// ceremony class the plan was made under (`small | bounded | architectural`,
// absent = architectural) and the design-part helpers (`splitSections`,
// `findModuleCycle`). plan_version 1.9 (the current target) adds the grammar
// of the `Files:` field and the `Test first:` `File:` line, the file first on
// every line (`entryPaths`). The rules that use them live in plan-validator.mjs.

import { ok, fail, issue } from '../util/result.mjs';
import { parseDocument } from '../util/frontmatter.mjs';
import { isTestPath } from '../util/test-path.mjs';

export const STEP_HEADING_REGEX = /^### Step (\d+):\s+(.+?)\s*$/m;
export const STEP_HEADING_GLOBAL = /^### Step (\d+):\s+(.+?)\s*$/gm;
export const FORBIDDEN_HEADING_REGEX = /^(?:##|###) (?:Fase|Phase|Stage|Steg) \d+/m;
export const FORBIDDEN_HEADING_GLOBAL = /^(?:##|###) (?:Fase|Phase|Stage|Steg) \d+/gm;
// Matches plan_version in either location the schema allows: at line start
// (frontmatter) or backtick-wrapped inside the prose "Generated by" metadata
// line the plan template emits (`> Generated by ... — `plan_version: 1.7``).
export const PLAN_VERSION_REGEX = /(?:^|`)plan_version:\s*['"]?([\d.]+)['"]?/m;

/**
 * Find all step heading positions in plan text.
 * @returns {Array<{n: number, title: string, line: number, offset: number}>}
 */
export function findSteps(text) {
  if (typeof text !== 'string') return [];
  const out = [];
  STEP_HEADING_GLOBAL.lastIndex = 0;
  let m;
  while ((m = STEP_HEADING_GLOBAL.exec(text)) !== null) {
    const offset = m.index;
    const line = text.slice(0, offset).split(/\r?\n/).length;
    out.push({ n: Number.parseInt(m[1], 10), title: m[2].trim(), line, offset });
  }
  return out;
}

/**
 * Find forbidden narrative-drift heading occurrences (Fase/Phase/Stage/Steg N).
 * @returns {Array<{form: string, line: number, offset: number, raw: string}>}
 */
export function findForbiddenHeadings(text) {
  if (typeof text !== 'string') return [];
  const out = [];
  FORBIDDEN_HEADING_GLOBAL.lastIndex = 0;
  let m;
  while ((m = FORBIDDEN_HEADING_GLOBAL.exec(text)) !== null) {
    const offset = m.index;
    const line = text.slice(0, offset).split(/\r?\n/).length;
    const raw = m[0];
    out.push({ form: raw, line, offset, raw });
  }
  return out;
}

/**
 * Slice plan text into per-step sections.
 * @returns {Array<{n: number, title: string, body: string, line: number}>}
 */
export function sliceSteps(text) {
  const heads = findSteps(text);
  const sections = [];
  for (let i = 0; i < heads.length; i++) {
    const start = heads[i].offset;
    const end = i + 1 < heads.length ? heads[i + 1].offset : text.length;
    const block = text.slice(start, end);
    sections.push({
      n: heads[i].n,
      title: heads[i].title,
      body: block,
      line: heads[i].line,
    });
  }
  return sections;
}

/**
 * Extract `plan_version: X.Y` from frontmatter or doc body.
 */
export function extractPlanVersion(text) {
  const m = typeof text === 'string' ? text.match(PLAN_VERSION_REGEX) : null;
  return m ? m[1] : null;
}

/**
 * Validate plan structure at the heading level.
 * Strict mode: forbidden-heading count > 0 → error. Step numbers must be 1..N contiguous.
 * @returns {import('../util/result.mjs').Result}
 */
export function validatePlanHeadings(text, opts = {}) {
  const strict = opts.strict !== false;
  const errors = [];
  const warnings = [];

  if (typeof text !== 'string') {
    return fail(issue('PLAN_INPUT', 'Plan text is not a string'));
  }

  const forbidden = findForbiddenHeadings(text);
  if (forbidden.length > 0) {
    const list = forbidden.map(f => `line ${f.line}: ${f.raw}`).join('; ');
    const errorIssue = issue(
      'PLAN_FORBIDDEN_HEADING',
      `Found ${forbidden.length} forbidden narrative-drift heading(s): ${list}`,
      'Use canonical "### Step N: <title>". Forbidden forms: Fase/Phase/Stage/Steg.',
    );
    if (strict) errors.push(errorIssue);
    else warnings.push(errorIssue);
  }

  const steps = findSteps(text);
  if (steps.length === 0) {
    errors.push(issue('PLAN_NO_STEPS', 'No step headings found', 'Expected at least one "### Step 1: <title>".'));
  } else {
    const numbers = steps.map(s => s.n);
    for (let i = 0; i < numbers.length; i++) {
      if (numbers[i] !== i + 1) {
        errors.push(issue(
          'PLAN_STEP_NUMBERING',
          `Step numbering breaks at position ${i + 1} (got Step ${numbers[i]})`,
          'Steps must be 1..N contiguous and ordered.',
        ));
        break;
      }
    }
  }

  return { valid: errors.length === 0, errors, warnings, parsed: { steps, forbidden } };
}

// ---- plan_version 1.8 ------------------------------------------------------

export const PLAN_VERSION_CURRENT = '1.9';
// 1.7 plans stay valid: the 1.8 rules apply only from 1.8 on, the 1.9 `Files:`
// grammar only from 1.9 on (see "plan_version 1.9" below).
export const PLAN_VERSIONS_SUPPORTED = Object.freeze(['1.7', '1.8', '1.9']);

/** True when dotted version `v` is >= `min` (numeric per part; null → false). */
export function versionAtLeast(v, min) {
  if (typeof v !== 'string' || !/^\d+(\.\d+)*$/.test(v)) return false;
  const a = v.split('.').map(Number);
  const b = min.split('.').map(Number);
  for (let i = 0; i < Math.max(a.length, b.length); i++) {
    const x = a[i] ?? 0;
    const y = b[i] ?? 0;
    if (x !== y) return x > y;
  }
  return true;
}

/** True when plan_version `v` reads `Files:` and the `Test first:` `File:` line by the 1.9 grammar. */
export const usesFilesGrammar = (v) => versionAtLeast(v, '1.9');

// ---- ceremony (the brief's size class) -------------------------------------

export const CEREMONY_VALUES = Object.freeze(['small', 'bounded', 'architectural']);
// Heaviest on doubt: a plan or brief that does not say is architectural.
export const CEREMONY_DEFAULT = 'architectural';
// A plan states its ceremony in exactly two places: frontmatter, or the
// "Generated by" header line (`> Generated by … · `ceremony: architectural``).
// Prose elsewhere ("the brief sets `ceremony: small`") never counts.
const HEADER_CEREMONY = /^> Generated by .*`ceremony:\s*([^`\s]+)\s*`/m;

/** @returns {{value: string, raw: string|null, defaulted: boolean, valid: boolean}} */
export function resolveCeremony(raw) {
  if (raw === null || raw === undefined || raw === '') {
    return { value: CEREMONY_DEFAULT, raw: null, defaulted: true, valid: true };
  }
  const s = String(raw).trim();
  return { value: s, raw: s, defaulted: false, valid: CEREMONY_VALUES.includes(s) };
}

function frontmatterCeremony(text) {
  const doc = typeof text === 'string' ? parseDocument(text) : null;
  const fm = doc?.parsed?.frontmatter;
  return fm && typeof fm === 'object' && fm.ceremony !== undefined ? fm.ceremony : null;
}

export function extractPlanCeremony(text) {
  if (typeof text !== 'string') return resolveCeremony(null);
  const fm = frontmatterCeremony(text);
  if (fm !== null) return resolveCeremony(fm);
  const m = text.match(HEADER_CEREMONY);
  return resolveCeremony(m ? m[1] : null);
}

/** Read `ceremony:` from a brief's frontmatter. Absent (or no frontmatter) → architectural. */
export function ceremonyFromBrief(briefText) {
  return resolveCeremony(frontmatterCeremony(briefText));
}

// ---- fences ------------------------------------------------------------------

/**
 * Tracks fenced code blocks the CommonMark way: a fence closes only on the same
 * character (` or ~), at least as long as the opener, with nothing after it.
 * `feed(line)` returns true when the line is inside a fence or is a fence line.
 * A fence line is indented up to three spaces; with `anyIndent` any indent
 * counts (the trailing-json readers, findings-schema `fencedBlocks`).
 */
export function fenceTracker({ anyIndent = false } = {}) {
  const fenceLine = anyIndent ? /^[ \t]*(`{3,}|~{3,})(.*)$/ : /^\s{0,3}(`{3,}|~{3,})(.*)$/;
  let open = null; // { ch, len }
  return {
    feed(line) {
      const m = line.match(fenceLine);
      if (!open) {
        if (m && !(m[1][0] === '`' && m[2].includes('`'))) {
          open = { ch: m[1][0], len: m[1].length };
          return true;
        }
        return false;
      }
      if (m && m[1][0] === open.ch && m[1].length >= open.len && m[2].trim() === '') open = null;
      return true;
    },
    get inside() { return open !== null; },
  };
}

// ---- per-step fields -------------------------------------------------------

// A top-level field bullet: `- **Name:** rest`, `- **Name**: rest` or `- Name: rest`.
const FIELD_LINE = /^- (?:\*\*)?([A-Z][A-Za-z -]*?)(?::\*\*|\*\*:|:)(?:\s+(.*))?$/;

/**
 * Parse a step body into its top-level `- **Field:**` bullets. A field's block
 * runs until the next top-level field bullet; indented lines (sub-bullets, the
 * Manifest YAML fence) belong to the field above them. Parsing stops at the
 * first heading after the step's own, outside fenced code: sliceSteps gives the
 * last step the rest of the plan, and a later section must not lend it fields.
 * @returns {Map<string, {name: string, inline: string, text: string}>}
 *   `inline` = text on the bullet line; `text` = inline + the indented block.
 */
export function parseStepFields(body) {
  const fields = new Map();
  if (typeof body !== 'string') return fields;
  const fence = fenceTracker();
  let current = null;
  let first = true;
  for (const line of body.split(/\r?\n/)) {
    const inFence = fence.feed(line);
    if (!inFence && /^#{1,6} /.test(line)) {
      if (!first) break;
      first = false;
      continue;
    }
    first = false;
    if (inFence) {
      if (current) current.lines.push(line);
      continue;
    }
    const m = line.match(FIELD_LINE);
    if (m) {
      current = { name: m[1].trim(), inline: (m[2] ?? '').trim(), lines: [] };
      if (!fields.has(current.name)) fields.set(current.name, current);
      continue;
    }
    if (current && /^\s/.test(line)) current.lines.push(line);
    else if (current && line.trim() !== '') current = null;
  }
  for (const f of fields.values()) {
    f.text = [f.inline, ...f.lines].join('\n').trim();
    delete f.lines;
  }
  return fields;
}

// ---- `On failure:` (K11) ---------------------------------------------------

// The recovery words /trekexecute reads from a step's `On failure:` ("Parse
// steps"). Free text that opens with none of them left the executor to guess:
// "retry once; if still failing, revert ... and STOP" met a ladder of fix rounds
// (voyage-lekerepo trial 2, 2026-10-08).
export const ON_FAILURE_ACTIONS = Object.freeze(['revert', 'retry', 'skip', 'escalate']);

/**
 * The recovery word an `On failure:` text opens with, or null. Case-insensitive;
 * leading markdown (`**`, backticks, a sub-bullet's `-`) is stripped.
 */
export function onFailureAction(text) {
  const s = String(text ?? '').trim().replace(/^[*_`\s-]+/, '').toLowerCase();
  const m = s.match(/^([a-z]+)\b/);
  return m && ON_FAILURE_ACTIONS.includes(m[1]) ? m[1] : null;
}

// ---- a step that holds nothing but tests -----------------------------------

/**
 * The paths a step's `Files:` field names, in order, for the tests-only check
 * below (the K6 sweep keeps v6.2.0's own reader: plan-scope-sweep.mjs,
 * operator 2026-10-11). The convention in the plan template is a backticked
 * path. Measured on the plan.md files outside worktrees under the marketplace
 * and the test repo: 520 Files: fields in 48 plan files at rounds 7-8 (round 7
 * wrote "182 plans" here; round 8's count of the files is 48), and 560 fields
 * in 48 plan files at round 9 (2026-10-11; the same at round 10), 548 of
 * them with a backticked path. In 3 of those a plain path-like word stands
 * beside it, each inside a description, and none of them is read as a file:
 *
 * - A field with a backticked path names its backticked spans that pass the
 *   path test (a `.`, `/` or `\`; no URL, option or call like `parse()`),
 *   wherever they sit (`a.mjs`/`b.mjs`, inside a note), a backticked build
 *   file (`Makefile`, `Dockerfile`), and a backticked capitalised name with no
 *   extension that stands alone in its entry before any description. Beside
 *   them a plain word counts only when it is clearly a file or a directory,
 *   and only before the entry's description starts (a dash, or a colon after
 *   a path, backticked or plain, wherever it stands: "src/cli.mjs: wires
 *   parse", "the entry cli.mjs: wires parse"; a colon after a span that is no
 *   path, `x()`:, starts none): a path with a slash that ends in a file or a
 *   `/` anywhere there ("the CLI entry src/cli.mjs"), and a name with an
 *   extension, a dotfile or a build file only where an entry or a list starts:
 *   first in its entry, or after "and", "or", "+" or "&" (", Dockerfile", "and
 *   Makefile", ", package.json"). A description runs on past a comma and onto
 *   the next line until a new entry starts: a list item on a new line (`-`,
 *   `1.` or `1)`), a backticked path, or a new line that opens with a plain
 *   file with a slash, or with one alone on the line or followed by a colon or
 *   a dash ("src/cli.mjs: the fix", "package.json", "Makefile — a target").
 *   Round 7: such files were dropped, and a step that builds something read as
 *   tests-only. Rounds 8-9: "— mocks fs.readFileSync", "runs on Node.js 22.x",
 *   "no change to package.json", "— mocks fs, child_process.spawnSync" or
 *   "src/cli.mjs: wires parse and README.md" read as files, and a step that
 *   holds only its test did not. Round 10: a plain file on the next line, a
 *   numbered item, a `Parser` inside a description, `x()`: and a mid-entry
 *   "cli.mjs:" were each misread. "covers pass/fail", "JSON, YAML", "v6.3.0
 *   format" or a next line "index.js and the CLI" is no path.
 * - A field with no backticked path names every path-like word: a name with an
 *   extension, a dotfile, a path with a slash, a directory (`docs/`), a glob
 *   (`src/*.mjs`) or a build file; a parenthesised note is not read. Another
 *   word with no slash and no extension counts only alone in its entry and
 *   capitalised; "format" or "v6.3.0" in a description is no path.
 *
 * A backslash is read as a slash, a leading `./` is dropped. Release review of
 * 6.3.0, rounds 1-10.
 *
 * Three answers 6.3.0 gave and 6.3.1 lost are given back, each in its narrow
 * form, and the round 7-10 neighbours of each keep 6.3.1's answer:
 *
 * - 6.4 / regression A: inside a description, a comma entry that opens with a
 *   plain file with a slash, or with one followed by a colon or a dash, starts
 *   a new entry ("`a.test.mjs` - new test, src/a.mjs - the fix" names both).
 *   That is the new-line test without its "alone on the line" branch
 *   (opensChunk), so "— mocks fs, child_process.spawnSync and path.join" stays
 *   description (round 9).
 * - 6.4 / regression B: in an entry opened by a backticked path, a plain path
 *   with a slash after a word of description ("`a.test.mjs` covers the parser
 *   in src/parse.mjs") counts only where a list starts (after "and", "or", "+"
 *   or "&"). Right after the backticked path ("`a.test.mjs` src/b.mjs"), and in
 *   an entry not opened by one (", the CLI entry src/cli.mjs", round 7), it is
 *   read as before. A list joiner is no word of description: "`a.mjs` +
 *   `b.mjs` src/c.mjs" reads all three.
 * - 6.4 / regression C: a new line that is no list item and holds only a
 *   backticked capitalised name ("`LICENSE`") ends the description, and the
 *   name is a file. On the description's own line it stays no file (round 10).
 */
const SEG = '[\\w@.*?()-]';
const PATH_WORD = new RegExp(`^(?:${SEG}+\\/)+${SEG}*$|^[\\w@*?-][\\w@.*?-]*\\.[A-Za-z*][A-Za-z0-9*]*$|^\\.[\\w-][\\w.-]*$`);
// Abbreviations that PATH_WORD would read as a name with an extension.
const NOT_A_PATH = /^(?:e\.g|i\.e|n\.b|cf|etc|vs)$/i;
const BARE_NAME = /^[A-Z][\w.-]*$/;
// The last segment is a dotfile or a name with an extension, or the path is a directory.
const FILE_OR_DIR = /(?:^|\/)(?:\.[\w-][\w.-]*|[^/]*\.[A-Za-z*][A-Za-z0-9*]*)$|\/$/;
// A build file named without an extension: a capitalised name that ends in
// `file` (Makefile, Dockerfile, Gemfile, Jenkinsfile, ...), not the word Profile,
// and the lowercase names make, rake and just also read (round 8).
const BUILD_FILE = /^(?!Profile$)(?:[A-Z][a-z]+|GNUmake)file$|^(?:make|rake|just)file$/;
// Where an entry's description starts, and the words that start a list item.
const DESCRIPTION_DASH = /^(?:\u2014|\u2013|-|--)$/;
const LIST_JOINER = /^(?:and|or|\+|&)$/i;
const LIST_ITEM = /^\s*(?:[-*+]|\d+[.)])\s+/;
const lastSegment = (p) => p.replace(/\/$/, '').split('/').pop();
/** A plan path as the readers hold it: `/` separators, no leading `./`. */
export const normPlanPath = (p) => p.replace(/\\/g, '/').replace(/^\.\//, '');
const norm = normPlanPath;
// A backticked span that is a path, not a URL, an option or a call (`parse()`, `p.parse()`).
const tickedPath = (s) => /[./\\]/.test(s) && !/^https?:/.test(s) && !s.startsWith('-') && !/^[\w$.]+\(.*\)$/.test(s);
// A plain word without the brackets and punctuation around it, `\` read as `/`.
const plainWord = (w) => w.replace(/^[(\[]+/, '').replace(/[.,;:)\]]+$/, '').replace(/\\/g, '/');
const isPathWord = (w) => PATH_WORD.test(w) && !NOT_A_PATH.test(w);
// Clearly a file or a directory: what a plain word beside backticked paths must be.
const isFileWord = (w) => BUILD_FILE.test(w) || (isPathWord(w) && (FILE_OR_DIR.test(w) || BUILD_FILE.test(lastSegment(w))));
// A line inside a description that opens a new entry: its first word is a
// plain file with a slash, or a plain file alone on the line or followed by a
// colon or a dash ("src/cli.mjs", "package.json", "Makefile — a target"), not
// "child_process.spawnSync, path.join" or "Node.js 22 only" (round 10).
// A comma entry inside a description that opens a new entry: opensEntry's test
// without "alone on the line", so "path.join" after a comma stays description
// (6.4 / regression A).
const opensChunk = (tokens) => {
  const w = plainWord(tokens[0]);
  return isFileWord(w) && (w.includes('/') || tokens[0].endsWith(':') || DESCRIPTION_DASH.test(tokens[1]));
};
const opensEntry = (tokens) => opensChunk(tokens) || (tokens.length === 1 && isFileWord(plainWord(tokens[0])));

export function fieldPaths(text) {
  const spans = [];
  const masked = String(text ?? '')
    .replace(/`([^`\n]+)`/g, (_, p) => `\u0000${spans.push(p.trim()) - 1}\u0000`)
    .replace(/(^|\s)\(([^()\n]*)\)/g, (m, pre, inner) => (inner.includes('\u0000') ? m : `${pre} `));
  const isTicked = spans.map((p) => tickedPath(p) || BUILD_FILE.test(p));
  const tickedField = isTicked.some(Boolean);
  const out = [];
  // A description runs on past a comma and onto the next line until a new
  // entry starts: a list item on a new line (`-`, `1.`, `1)`), a backticked
  // path, or a line that opens with a plain file (rounds 9-10).
  let described = false;
  for (const line of masked.split('\n')) {
    const item = LIST_ITEM.exec(line);
    const body = item ? line.slice(item[0].length) : line;
    if (item || (described && body.trim() !== '' && opensEntry(body.trim().split(/\s+/)))) described = false;
    // A new line that holds only a backticked capitalised name ends the
    // description (6.4 / regression C).
    const alone = /^\u0000(\d+)\u0000$/.exec(body.trim());
    if (alone && BARE_NAME.test(spans[Number(alone[1])])) described = false;
    for (const raw of body.split(',')) {
      const entry = raw.trim().replace(/^[-*]\s+/, '');
      if (entry === '') continue;
      const tokens = entry.split(/\s+/);
      const first = /^\u0000(\d+)\u0000/.exec(entry);
      // An entry opened by a backticked path: after a word of description a
      // plain path with a slash counts only where a list starts (6.4 / regression B).
      const tickedEntry = Boolean(first && isTicked[Number(first[1])]);
      if (tickedEntry) described = false;
      if (described && opensChunk(tokens)) described = false;
      let prose = false;
      // A backticked capitalised name alone in its entry, before any
      // description, is a file (rounds 6, 10).
      if (first && first[0] === entry && !described && BARE_NAME.test(spans[Number(first[1])])) {
        out.push(spans[Number(first[1])]);
        continue;
      }
      if (tickedField) {
        // In the order they stand: the backticked paths, and before the
        // description a plain word that is clearly a file where it may stand
        // (rounds 7-8). A path followed by a colon starts its description,
        // backticked or plain, wherever it stands; a span that is no path
        // does not (rounds 9-10).
        for (let i = 0; i < tokens.length; i++) {
          const t = tokens[i];
          if (t.includes('\u0000')) {
            const ticked = [...t.matchAll(/\u0000(\d+)\u0000/g)].map((m) => Number(m[1])).filter((n) => isTicked[n]);
            out.push(...ticked.map((n) => spans[n]));
            if (ticked.length > 0 && t.endsWith(':')) described = true;
            continue;
          }
          if (DESCRIPTION_DASH.test(t)) described = true;
          const w = plainWord(t);
          const file = isFileWord(w);
          if (!file && !LIST_JOINER.test(t)) prose = true;
          if (described || !file) continue;
          if ((w.includes('/') && !(tickedEntry && prose)) || i === 0 || LIST_JOINER.test(tokens[i - 1])) out.push(w);
          if (t.endsWith(':')) described = true;
        }
        continue;
      }
      const words = tokens.filter((w) => !w.includes('\u0000')).map(plainWord).filter(Boolean);
      const found = words.filter((w) => isPathWord(w) || BUILD_FILE.test(w));
      if (found.length === 0 && words.length === 1 && tokens.length === 1 && BARE_NAME.test(words[0])) found.push(words[0]);
      out.push(...found);
    }
  }
  return out.filter(Boolean).map(norm);
}

// ---- plan_version 1.9 ------------------------------------------------------
//
// From plan_version 1.9 a step's `Files:` field has a grammar, so no reader has
// to tell a path from a description in free text (decision 16, operator
// 2026-10-11: only a 1.9 plan is refused; an older plan is read by fieldPaths
// as before and gets one warning, PLAN_STEP_PATH_UNTICKED):
//
// - The entries are the field's first non-empty line and each list-item line
//   (`-`, `*`, `+`, `1.`, `1)`). Any other non-empty line continues the entry
//   above it: when the entry so far ends with a comma it is appended with one
//   space before the entry is read, so a list wrapped after a comma
//   ("`tests/a.test.mjs`," and "`src/a.mjs`" on the next line) names both
//   files; otherwise it and every later wrapped line of that entry are
//   description and never read (operator decision 29, 14.10; reviews
//   `d21ff81e`, `eaeb398e`). An entry with nothing on its own line (a `File:`
//   line with nothing after it) takes its first wrapped line. An empty first
//   line before a nested list is no entry. An issue carries the line number
//   of the entry's first line (review `b4fa1927`).
// - An entry's files are the backticked spans it opens with, separated only by
//   commas or whitespace. Everything from the first other text on is
//   description and is never read ("`a.test.mjs` - covers `src/parse.mjs`"
//   names one file).
// - An entry that does not open with a backticked span is refused
//   (PLAN_STEP_PATH_NOT_FIRST); so are `none` and a field that names no file.
//   A leading span that fails isPathShape (`parse()`, an option, a URL) is
//   refused as no path (PLAN_STEP_PATH_NOT_A_PATH).
//
// - A `Test first:` `File:` line is a line that opens with `File:` (or
//   `**File:**`) after an optional list marker of the same set (`-`, `*`, `+`,
//   `1.`, `1)`); its rest is one entry under this grammar. Any other
//   non-empty line right after it continues it by the same rule: when the
//   entry so far ends with a comma it is appended with one space, so a list
//   wrapped after a comma names both files; otherwise it and every later
//   wrapped line of that entry are description and never read (operator
//   decision 29, 14.10; reviews `d21ff81e`, `eaeb398e`). A `File:` line with
//   nothing after it takes its first wrapped line. Any list item (a
//   `Verifies:` or `Pattern:` line too) ends it, and a non-list line after
//   such an item belongs to that item, never to the `File:` entry. A line that
//   opens with `File:` always starts a new entry; a `File:` mid-line names no
//   test (review `4b6460e5`).
//
// isPathShape is the 1.9-only test on a span in the leading position of an
// entry (operator decision 22, 13.10): a placeholder (none, TBD, TODO, TBA,
// N/A, NA, `...`) is no file; a path, a build file or an all-capitals name
// such as `LICENSE` or `README` is; a mixed-case bare name (`Parser`) is not,
// and `./Name` names such a file. The readers below 1.9 keep BARE_NAME.

const LEADING_SPAN = /^`([^`\n]+)`/;
// Operator decision 22: these words are the input a 1.9 plan refuses, not unfinished code.
const PLACEHOLDER = /^(?:none|n\/?a|tbd|tba|todo|\.{3})$/i;
const CAPS_NAME = /^[A-Z][A-Z0-9_-]*$/;

/** A backticked span in the leading position of an entry that names a file. */
export function isPathShape(s) {
  const span = String(s ?? '').trim();
  if (PLACEHOLDER.test(span)) return false;
  return tickedPath(span) || BUILD_FILE.test(span) || CAPS_NAME.test(span);
}

/**
 * The trimmed contents of the backticked spans `text` opens with, each
 * separated from the next only by commas or whitespace; empty when `text` does
 * not open with a backticked span. The caller judges each span.
 * @param {string} text
 * @returns {string[]}
 */
function leadingSpans(text) {
  const out = [];
  let rest = String(text ?? '');
  let m = LEADING_SPAN.exec(rest);
  while (m) {
    out.push(m[1].trim());
    rest = rest.slice(m[0].length);
    const sep = /^[\s,]+/.exec(rest);
    m = sep ? LEADING_SPAN.exec(rest.slice(sep[0].length)) : null;
    if (m) rest = rest.slice(sep[0].length);
  }
  return out;
}

/**
 * Whether a wrapped line (a non-empty line that is no list item) joins the
 * 1.9 entry above it (operator decision 29, 14.10; reviews `d21ff81e`,
 * `eaeb398e`). An entry with nothing on its own line (a `File:` line with
 * nothing after it) takes its first wrapped line; an entry whose text so far
 * ends with a comma gets the line appended with one space, so a list wrapped
 * after a comma names both files. Otherwise the line is description: null,
 * and the caller closes the entry, so every later wrapped line of it is
 * description too.
 * @param {string} text - the entry so far
 * @param {string} line - the wrapped line
 * @returns {string|null} the entry's new text, or null when the line is description
 */
function wrapLine(text, line) {
  const so = String(text ?? '');
  if (so.trim() === '') return line.trim();
  if (so.trimEnd().endsWith(',')) return `${so} ${line.trim()}`.trim();
  return null;
}

/**
 * The files a 1.9 `Files:` field names, by the grammar above.
 * @param {string} text - the field's text (parseStepFields(body).get('Files').text)
 * @returns {{files: string[], issues: Array<{line: number, entry: string,
 *   code: 'PLAN_STEP_PATH_NOT_FIRST'|'PLAN_STEP_PATH_NOT_A_PATH'}>}}
 *   files normalised, in order, deduplicated; `line` is 1-based in `text`.
 */
export function entryPaths(text) {
  const files = [];
  const issues = [];
  const entries = [];
  String(text ?? '').split('\n').forEach((line, i) => {
    const item = LIST_ITEM.exec(line);
    const rest = line.trim();
    if (item) entries.push({ line: i + 1, entry: line.slice(item[0].length).trim(), closed: false });
    else if (rest === '') return;
    else if (entries.length === 0) entries.push({ line: i + 1, entry: rest, closed: false });
    else {
      // A wrapped line joins the entry above it only after a comma, or when
      // the entry is empty (wrapLine); otherwise it closes the entry, and it
      // and every later wrapped line of that entry are description, never read
      // (operator decision 29).
      const last = entries[entries.length - 1];
      if (last.closed) return;
      const joined = wrapLine(last.entry, rest);
      if (joined === null) last.closed = true;
      else last.entry = joined;
    }
  });
  if (entries.length === 0) {
    issues.push({ line: 1, entry: '', code: 'PLAN_STEP_PATH_NOT_FIRST' });
    return { files, issues };
  }
  for (const { line, entry } of entries) {
    const spans = leadingSpans(entry);
    if (spans.length === 0) {
      issues.push({ line, entry, code: 'PLAN_STEP_PATH_NOT_FIRST' });
      continue;
    }
    for (const span of spans) {
      if (isPathShape(span)) {
        const p = norm(span);
        if (!files.includes(p)) files.push(p);
      } else {
        issues.push({ line, entry, code: 'PLAN_STEP_PATH_NOT_A_PATH' });
      }
    }
  }
  return { files, issues };
}

/**
 * The paths fieldPaths reads from `text` that do not stand in backticks: the
 * paths below plan_version 1.9 that a 1.9 plan would refuse.
 * @returns {string[]}
 */
export function untickedPaths(text) {
  const ticked = new Set([...String(text ?? '').matchAll(/`([^`\n]+)`/g)].map((m) => norm(m[1].trim())));
  return fieldPaths(text).filter((p) => !ticked.has(p));
}

// A document or a data file outside tests/ is never the test here:
// `docs/specs/api.md` and `api/specs/openapi.yaml` sit under a `specs/` folder,
// which lib/util/test-path.mjs reads as tests for the floor guard. Under tests/
// either one is a fixture, and a fixture is a test file.
const DOC_RE = /\.(md|markdown|mdx|txt|rst|adoc)$/i;
const DATA_RE = /\.(ya?ml|json|toml|xml|csv)$/i;
const UNDER_TESTS_RE = /(^|\/)(tests?|__tests__)\//;

/** A path that is clearly a test: a test by its path, and under tests/ or neither a document nor a data file. */
function clearlyTest(p) {
  return isTestPath(p) && (UNDER_TESTS_RE.test(p) || !(DOC_RE.test(p) || DATA_RE.test(p)));
}

/**
 * Whether every file in a step's `Files:` is a test: the step's own `Test
 * first:` file, or a path that is clearly a test (lib/util/test-path.mjs). Such
 * a step can never be proven red - the behaviour its tests cover already
 * exists at its start point (an earlier step built it, or it was there before
 * the plan), so they already pass there (voyage-lekerepo
 * trial 3, 2026-10-09: `not-red-at-base`, stop class 4). The plan validator
 * warns on it (PLAN_STEP_TESTS_ONLY) and red-proof.mjs reports it
 * (`tests_only`), so the executor sends the step up to the plan. A step
 * without `Files:` is not such a step. From plan_version 1.9 `Files:` is read
 * by its grammar (entryPaths); below 1.9 (or without a version) by fieldPaths.
 * @param {string} body - the step's section (sliceSteps)
 * @param {string|null} [version] - the plan's plan_version
 * @returns {{testsOnly: boolean, files: string[]}}
 */
export function stepTestsOnly(body, version) {
  return stepTestsOnlyFromFields(parseStepFields(body), version);
}

// A line of the `Test first:` field that names its test files (6.3.1's
// selector); global, so read it with matchAll only: one match per `File:`.
const TEST_FIRST_FILE_LINE = /\bFile:\**[ \t]*(?=`)/gi;
// From plan_version 1.9: a line that opens with `File:` (after an optional
// list marker of the set LIST_ITEM accepts, and `**`), backticked or not.
const TEST_FIRST_FILE_LINE_19 = /^\s*(?:(?:[-*+]|\d+[.)])\s+)?(?:\*\*)?File:(?:\*\*)?/i;
// Below plan_version 1.9: one glue token at the start of the rest of a `File:`
// run (operator decisions 28, 13.10, and 31, 14.10): whitespace or a comma
// (group 1); a note in brackets holding no bracket, backtick or newline,
// optionally in one `*` or `_` (`(new)`, `*(new)*`, `(existing, the parser)`);
// the whole word `and` or `plus`, any case (`android` is no glue); or `+`.
const OLDER_GLUE = /^(?:(\s+|,)|[*_]?\([^()`\n]*\)[*_]?|(?:and|plus)(?!\w)|\+)/i;

/**
 * The files of one below-1.9 `File:` run, read from just after the marker:
 * the leading backticked span, then each span after one or more glue tokens
 * (OLDER_GLUE). While every glue token so far was whitespace or a comma, a
 * span with the shape of a path is kept and any other span is skipped
 * (operator decision 20). Once a glue token was a note or a word, a span is
 * kept only when it has the shape of a path and is clearly a test, and the
 * first span that is not ends the run (operator decisions 28 and 31). Any text
 * that is neither a span nor glue ends the run.
 * @param {string} text - the line after one `File:` marker
 * @returns {string[]} the kept spans, trimmed, not normalised
 */
function olderFileRun(text) {
  const out = [];
  let rest = String(text ?? '');
  let loose = false;
  let m = LEADING_SPAN.exec(rest);
  while (m) {
    const s = m[1].trim();
    if (!loose) {
      if (tickedPath(s) || BUILD_FILE.test(s)) out.push(s);
    } else if (tickedPath(s) && clearlyTest(norm(s))) {
      out.push(s);
    } else {
      break;
    }
    rest = rest.slice(m[0].length);
    let glued = false;
    let g = OLDER_GLUE.exec(rest);
    while (g) {
      glued = true;
      if (g[1] === undefined) loose = true;
      rest = rest.slice(g[0].length);
      g = OLDER_GLUE.exec(rest);
    }
    m = glued ? LEADING_SPAN.exec(rest) : null;
  }
  return out;
}

/**
 * plan_version 1.9: the rest of each `Test first:` `File:` line (after `File:`
 * and any `**`), one entry per `File:` line, in order. Any other non-empty
 * line right after it continues the entry above it: when the entry so far
 * ends with a comma it is appended with one space, so a list wrapped after a
 * comma names both files; otherwise it and every later wrapped line of that
 * entry are description and never read (operator decision 29, 14.10; reviews
 * `d21ff81e`, `eaeb398e`). An entry with nothing on its own line (a `File:`
 * line with nothing after it) takes its first wrapped line. Any list item ends
 * the entry; a line that opens with `File:` starts a new one. The one
 * selector both testFirstFiles and the plan validator's grammar check read.
 * @param {Map<string, {text: string}>} fields - parseStepFields(body)
 * @returns {string[]}
 */
export function testFirstFileEntries(fields) {
  if (!fields || !fields.has('Test first')) return [];
  const out = [];
  // Whether the line above is the current `File:` line or its continuation.
  let open = false;
  for (const line of String(fields.get('Test first').text ?? '').split('\n')) {
    const at = TEST_FIRST_FILE_LINE_19.exec(line);
    if (at) {
      out.push(line.slice(at[0].length).trim());
      open = true;
    } else if (LIST_ITEM.test(line) || line.trim() === '') {
      open = false;
    } else if (open) {
      const joined = wrapLine(out[out.length - 1], line);
      if (joined === null) open = false;
      else out[out.length - 1] = joined;
    }
  }
  return out;
}

/**
 * The step's test files: every file on the `Test first:` `File:` line.
 *
 * Below plan_version 1.9 (or without a version) a line counts as that line
 * exactly when TEST_FIRST_FILE_LINE matches it. Each `File:` marker on the
 * line opens a run: the backticked spans that stand right after it, separated
 * only by commas or whitespace (olderFileRun), and every span in that run with
 * the shape of a path is a test file (a call `parse()`, an option, a URL is
 * skipped and the run goes on). A path in the description after the run is no
 * test file (operator decision 20, 12.10; review `6c31212c`: "covers `parse()`
 * in `src/parser.mjs`" made the source a test and the step tests-only). The
 * markers are matched globally because 6.3.1 read the first span after each
 * `File:`, so a line with two markers keeps at least 6.3.1's files. Below 1.9
 * the run also goes on past a note in brackets (`(new)`, `*(new)*`,
 * `(existing ...)`) and the words `and`, `plus` and `+`, to a further path
 * only when it is clearly a test, and a source path there ends it
 * (olderFileRun; operator decisions 28 and 31, 13.10 and 14.10; review
 * `bd2520b0`; review `6c31212c` for the guard).
 *
 * From 1.9 a line counts when it opens with `File:` (testFirstFileEntries),
 * backticked or not, after an optional list marker (`1.` and `1)` included);
 * its rest is one entry read by entryPaths: the leading backticked paths are
 * the test files, the rest is description. Any other non-empty line right
 * after it continues the entry: when the entry so far ends with a comma it is
 * appended with one space, so a list wrapped after a comma names both files;
 * otherwise it and every later wrapped line of that entry are description and
 * never read (operator decision 29, 14.10; reviews `d21ff81e`, `eaeb398e`). A
 * `File:` line with nothing after it takes its first wrapped line.
 *
 * Normalised, in order, deduplicated. A `Verifies:` line without a `File:` is
 * never read.
 * @param {Map<string, {text: string}>} fields - parseStepFields(body)
 * @param {string|null} [version] - the plan's plan_version
 * @returns {string[]}
 */
export function testFirstFiles(fields, version) {
  if (!fields || !fields.has('Test first')) return [];
  const out = [];
  if (usesFilesGrammar(version)) {
    for (const entry of testFirstFileEntries(fields)) {
      for (const p of entryPaths(entry).files) if (!out.includes(p)) out.push(p);
    }
    return out;
  }
  for (const line of String(fields.get('Test first').text ?? '').split('\n')) {
    for (const at of line.matchAll(TEST_FIRST_FILE_LINE)) {
      for (const s of olderFileRun(line.slice(at.index + at[0].length))) {
        const p = norm(s);
        if (!out.includes(p)) out.push(p);
      }
    }
  }
  return out;
}

/**
 * Whether a path counts as a test for this step: its own `Test first:` files
 * (testFirstFiles), or a path that is clearly a test.
 * @returns {(path: string) => boolean}
 */
function stepTestFile(f, version) {
  const testFirst = testFirstFiles(f, version);
  return (p) => testFirst.includes(p) || clearlyTest(p);
}

/** `stepTestsOnly` for a caller that already holds the step's parsed fields. */
export function stepTestsOnlyFromFields(f, version) {
  const text = f.has('Files') ? f.get('Files').text : null;
  const files = text === null ? [] : usesFilesGrammar(version) ? entryPaths(text).files : fieldPaths(text);
  return { testsOnly: files.length > 0 && files.every(stepTestFile(f, version)), files };
}

// ---- sections (fence-aware) ------------------------------------------------

/**
 * Split text at headings of exactly `level` (#-count), ignoring headings inside
 * fenced code blocks. Text before the first heading has title null.
 * @returns {Array<{title: string|null, start: number, end: number, text: string}>}
 *   start/end are line indexes (end exclusive).
 */
export function splitSections(text, level = 2) {
  const lines = typeof text === 'string' ? text.split('\n') : [];
  const prefix = '#'.repeat(level) + ' ';
  const fence = fenceTracker();
  const out = [];
  let cur = { title: null, start: 0 };
  for (let i = 0; i < lines.length; i++) {
    const l = lines[i];
    if (fence.feed(l)) continue;
    if (l.startsWith(prefix)) {
      out.push({ ...cur, end: i });
      cur = { title: l.slice(prefix.length).trim(), start: i };
    }
  }
  out.push({ ...cur, end: lines.length });
  return out
    .filter(s => s.end > s.start || s.title !== null)
    .map(s => ({ ...s, text: lines.slice(s.start, s.end).join('\n') }));
}

/** First section of `level` whose title equals `title` (case-insensitive), or null. */
export function findSection(text, title, level = 2) {
  const want = title.toLowerCase();
  return splitSections(text, level).find(s => s.title && s.title.toLowerCase() === want) ?? null;
}

/** Every fenced block whose info string is `lang`, in order. */
export function fencedBlocks(text, lang) {
  const out = [];
  const lines = typeof text === 'string' ? text.split('\n') : [];
  const fence = fenceTracker();
  let cur = null;
  for (const l of lines) {
    const wasInside = fence.inside;
    fence.feed(l);
    if (!wasInside && fence.inside) {
      cur = new RegExp(`^\\s{0,3}(\`{3,}|~{3,})\\s*${lang}\\s*$`).test(l) ? [] : null;
    } else if (wasInside && !fence.inside) {
      if (cur) out.push(cur.join('\n'));
      cur = null;
    } else if (cur) {
      cur.push(l);
    }
  }
  return out;
}

// ---- module map ------------------------------------------------------------

// Directives that are not node statements: the keyword must stand alone or be
// followed by whitespace, so a node id like `end-user` or `graph.mjs` is kept.
const DIRECTIVE = /^(graph|flowchart|subgraph|end|classDef|class|style|linkStyle|direction|click|accTitle|accDescr)(\s|$)/;
// Directed and undirected links, after labels are stripped:
//   -->  --->  ==>  ===>  -.->  -..->  with an optional leading < (both ways);
//   ---  ===  -.-  ~~~  (no direction: not a dependency).
const LINK = /\s*(<?-{2,}>|<?={2,}>|<?-\.+->|-{3,}|={3,}|-\.+-|~{3,})\s*/;
const NODE_ID = /^[A-Za-z0-9_][\w.\-/]*$/;
const ARROWISH = /-->|==>|-\.+->|<--|<==/;

function stripLabels(line) {
  return line
    .replace(/"[^"]*"/g, '')
    .replace(/:::[\w-]+/g, '')
    .replace(/\[\[[^\]]*\]\]|\[\([^\]]*\)\]|\[\/[^\]]*\/\]|\[\\[^\]]*\\\]|\[[^\]]*\]/g, '')
    .replace(/\(\(\([^)]*\)\)\)|\(\([^)]*\)\)|\(\[[^\]]*\]\)|\([^)]*\)/g, '')
    .replace(/\{\{[^}]*\}\}|\{[^}]*\}/g, '')
    .replace(/>[^\]\s][^\]]*\]/g, '')
    .replace(/\|[^|]*\|/g, '')
    // link text: `-- text -->`, `-. text .->`, `== text ==>`
    .replace(/--\s+[^-<>=]+?\s+(-{2,}>)/g, '$1')
    .replace(/-\.\s+[^.<>]+?\s+\.(-+>)/g, '-.$1')
    .replace(/==\s+[^=<>]+?\s+(={2,}>)/g, '$1');
}

/**
 * Parse a mermaid flowchart into directed edges.
 * Handles chains (a --> b --> c), `&` groups, `;`-separated statements,
 * `<-->` (one edge each way), long arrows, labels and `:::class`.
 * @returns {{edges: Array<[string, string]>, unparsed: string[]}}
 *   unparsed = statements that look like links but yielded no edge.
 */
export function parseMermaid(mermaid) {
  const edges = [];
  const unparsed = [];
  if (typeof mermaid !== 'string') return { edges, unparsed };
  for (const raw of mermaid.split(/\r?\n/)) {
    for (const stmt0 of raw.split(';')) {
      const stmt = stmt0.trim();
      if (!stmt || stmt.startsWith('%%') || DIRECTIVE.test(stmt)) continue;
      const parts = stripLabels(stmt).split(LINK).map(p => p.trim());
      if (parts.length < 3) {
        if (ARROWISH.test(stmt)) unparsed.push(stmt);
        continue;
      }
      let ok = true;
      const groups = [];
      for (let i = 0; i < parts.length; i += 2) {
        const ids = parts[i].split('&').map(s => s.trim());
        if (ids.length === 0 || ids.some(id => !NODE_ID.test(id))) { ok = false; break; }
        groups.push(ids);
      }
      if (!ok) { unparsed.push(stmt); continue; }
      for (let i = 1, g = 0; i < parts.length; i += 2, g++) {
        const link = parts[i];
        const forward = link.endsWith('>');
        const backward = link.startsWith('<');
        if (!forward) continue; // undirected: not a dependency
        for (const a of groups[g]) {
          for (const b of groups[g + 1]) {
            edges.push([a, b]);
            if (backward) edges.push([b, a]);
          }
        }
      }
    }
  }
  return { edges, unparsed };
}

/** Directed edges [from, to] from mermaid flowchart text (see parseMermaid). */
export function parseMermaidEdges(mermaid) {
  return parseMermaid(mermaid).edges;
}

/**
 * Find one directed cycle in a module map (mermaid text or an edge list).
 * Iterative depth-first search: a deep chain cannot overflow the stack.
 * @returns {string[]|null} the cycle as a closed path (first === last), or null.
 */
export function findModuleCycle(input) {
  const edges = Array.isArray(input) ? input : parseMermaidEdges(input);
  const adj = new Map();
  for (const [a, b] of edges) {
    if (!adj.has(a)) adj.set(a, []);
    if (!adj.has(b)) adj.set(b, []);
    adj.get(a).push(b);
  }
  const state = new Map(); // unset = unvisited, 1 = on the path, 2 = done
  for (const root of adj.keys()) {
    if (state.has(root)) continue;
    const path = [root];
    const iters = [0];
    state.set(root, 1);
    while (path.length > 0) {
      const n = path[path.length - 1];
      const next = adj.get(n);
      const i = iters[iters.length - 1];
      if (i >= next.length) {
        state.set(n, 2);
        path.pop();
        iters.pop();
        continue;
      }
      iters[iters.length - 1] = i + 1;
      const nx = next[i];
      if (state.get(nx) === 1) return [...path.slice(path.indexOf(nx)), nx];
      if (!state.has(nx)) {
        state.set(nx, 1);
        path.push(nx);
        iters.push(0);
      }
    }
  }
  return null;
}