voyage/lib/review/critic-diff.mjs

// lib/review/critic-diff.mjs
// Rule A and the rule-B record for the plan gate (decision 43).
//
// Why: from round 2 on, the critic loop counts a new MAJOR that cites plan text
// outside the diff since the last reviewed round as minor (rule A), so a loop of
// critic rounds converges on what changed instead of re-litigating text that was
// already reviewed. That needs a line diff between two operative plan texts
// that is deterministic: the same two texts always give the same hunks, so the
// same findings always give the same verdict.
//
// lineDiff(oldText, newText) -> { inside, oldToNew, hunks }
//   Lines are 1-based operative lines; CRLF is normalized first.
//   inside   Set of every inserted or changed new line; for a pure-deletion
//            hunk also the new lines on both sides of the deletion point.
//   oldToNew oldToNew[o] = the new line an unchanged old line moved to, else
//            null (index 0 is null).
//   hunks    [{oldStart, oldEnd, newStart, newEnd}], inclusive ranges in order;
//            an empty side has end = start - 1.
//   How: trim the common prefix and suffix; anchor on lines that occur exactly
//   once in both texts, keep the longest increasing subsequence of anchors and
//   recurse between them; a gap with no unique line uses an LCS table, ties
//   resolved toward matching the earliest old line.
//
// Pure: no I/O, and it never reads the plan gate (one dependency direction).
//
// Credit: patience-style anchoring (unique-line anchors kept by their longest
// increasing subsequence), the idea behind Bram Cohen's patience diff; the code
// here is written fresh.

const toLines = (text) => {
  const t = String(text ?? '').replace(/\r\n?/g, '\n');
  if (t === '') return [];
  const out = t.split('\n');
  if (out[out.length - 1] === '') out.pop();
  return out;
};

/** Indexes (into `items`) of a longest strictly increasing subsequence of items[k].b; earliest on ties. */
function lisByB(items) {
  const tails = []; // tails[len-1] = index into items of the smallest tail
  const prev = new Array(items.length).fill(-1);
  for (let k = 0; k < items.length; k++) {
    const v = items[k].b;
    let lo = 0;
    let hi = tails.length;
    while (lo < hi) {
      const mid = (lo + hi) >> 1;
      if (items[tails[mid]].b < v) lo = mid + 1; else hi = mid;
    }
    if (lo > 0) prev[k] = tails[lo - 1];
    tails[lo] = k;
  }
  const out = [];
  for (let k = tails.length ? tails[tails.length - 1] : -1; k !== -1; k = prev[k]) out.push(k);
  return out.reverse();
}

/** LCS over a[aLo..aHi) x b[bLo..bHi); pushes matched [i, j] pairs in order. */
function lcsGap(a, aLo, aHi, b, bLo, bHi, pairs) {
  const n = aHi - aLo;
  const m = bHi - bLo;
  const w = m + 1;
  // L[i*w + j] = LCS length of a[aLo+i..aHi) and b[bLo+j..bHi)
  const L = new Uint32Array((n + 1) * w);
  for (let i = n - 1; i >= 0; i--) {
    for (let j = m - 1; j >= 0; j--) {
      L[i * w + j] = a[aLo + i] === b[bLo + j]
        ? L[(i + 1) * w + j + 1] + 1
        : Math.max(L[(i + 1) * w + j], L[i * w + j + 1]);
    }
  }
  let i = 0;
  let j = 0;
  while (i < n && j < m) {
    if (a[aLo + i] === b[bLo + j]) {
      pairs.push([aLo + i, bLo + j]);
      i += 1;
      j += 1;
    } else if (L[i * w + j + 1] >= L[(i + 1) * w + j]) {
      j += 1; // keep old line i for the earliest match
    } else {
      i += 1;
    }
  }
}

/** Match a[aLo..aHi) against b[bLo..bHi); pushes matched [i, j] pairs (0-based) in order. */
function matchRange(a, aLo, aHi, b, bLo, bHi, pairs) {
  while (aLo < aHi && bLo < bHi && a[aLo] === b[bLo]) {
    pairs.push([aLo, bLo]);
    aLo += 1;
    bLo += 1;
  }
  const tail = [];
  while (aLo < aHi && bLo < bHi && a[aHi - 1] === b[bHi - 1]) {
    aHi -= 1;
    bHi -= 1;
    tail.push([aHi, bHi]);
  }
  if (aLo < aHi && bLo < bHi) {
    const count = new Map(); // line -> {na, nb, a, b}
    for (let i = aLo; i < aHi; i++) {
      const c = count.get(a[i]) ?? { na: 0, nb: 0, a: -1, b: -1 };
      c.na += 1;
      c.a = i;
      count.set(a[i], c);
    }
    for (let j = bLo; j < bHi; j++) {
      const c = count.get(b[j]);
      if (!c) continue;
      c.nb += 1;
      c.b = j;
    }
    const anchors = [];
    for (let i = aLo; i < aHi; i++) {
      const c = count.get(a[i]);
      if (c.na === 1 && c.nb === 1) anchors.push({ a: c.a, b: c.b });
    }
    if (anchors.length > 0) {
      let pa = aLo;
      let pb = bLo;
      for (const k of lisByB(anchors)) {
        const { a: ia, b: ib } = anchors[k];
        matchRange(a, pa, ia, b, pb, ib, pairs);
        pairs.push([ia, ib]);
        pa = ia + 1;
        pb = ib + 1;
      }
      matchRange(a, pa, aHi, b, pb, bHi, pairs);
    } else {
      lcsGap(a, aLo, aHi, b, bLo, bHi, pairs);
    }
  }
  for (let k = tail.length - 1; k >= 0; k--) pairs.push(tail[k]);
}

/**
 * Deterministic line diff of two operative texts.
 * @param {string} oldText
 * @param {string} newText
 * @returns {{inside: Set<number>, oldToNew: Array<number|null>, hunks: Array<{oldStart: number, oldEnd: number, newStart: number, newEnd: number}>}}
 */
export function lineDiff(oldText, newText) {
  const a = toLines(oldText);
  const b = toLines(newText);
  const pairs = [];
  matchRange(a, 0, a.length, b, 0, b.length, pairs);
  const oldToNew = new Array(a.length + 1).fill(null);
  for (const [i, j] of pairs) oldToNew[i + 1] = j + 1;
  const hunks = [];
  const inside = new Set();
  let pi = -1;
  let pj = -1;
  for (const [qi, qj] of [...pairs, [a.length, b.length]]) {
    if (qi > pi + 1 || qj > pj + 1) {
      const h = { oldStart: pi + 2, oldEnd: qi, newStart: pj + 2, newEnd: qj };
      hunks.push(h);
      if (h.newEnd >= h.newStart) {
        for (let n = h.newStart; n <= h.newEnd; n++) inside.add(n);
      } else {
        if (h.newStart - 1 >= 1) inside.add(h.newStart - 1);
        if (h.newStart <= b.length) inside.add(h.newStart);
      }
    }
    pi = qi;
    pj = qj;
  }
  return { inside, oldToNew, hunks };
}