@reportwright/pdf 0.1.0-beta.3 → 0.1.0-beta.5

package/README.md

@@ -322,7 +322,7 @@ PDF/A levels and the reader. The most used:
 | `createPdf(sink, o)` / `toBytes(build, o)` | `title`, `author`, `creationDate`, `compress`, `metadata`, `pdfa`, `tagged`, `encrypt`, `sign`, `form.flatten`, `signal`, `timeoutMs`, `id` ([table](#createpdfsink-options)) |
 | `pdf.embedFont(bytes, o)` | `subset`, `shaper`, `textMapping` ([Fonts](#fonts), [Shaping](#shaping)) |
 | `page.text` / `page.textBox` | `x`, `y`, `font`, `size`, `color`, `width`, `align`, `maxLines`, `fallback`, `direction`, `kerning` ([Text options](#pages)) |
-| `pdf.embedImage` / `page.image` | `colorSpace`, `maxDecodedBytes` / `width`, `height`, `fit`, `opacity`, `alt` ([Images](#images)) |
+| `pdf.embedImage` / `page.image` | `colorSpace`, `maxDecodedBytes`, `compression` / `width`, `height`, `fit`, `opacity`, `alt` ([Images](#images)) |
 | `page.fill` / `stroke` | `fill`, `stroke`, `width`, `dash`, `cap`, `join`, `opacity`, `blendMode`, `rule` ([Vector graphics](#vector-graphics)) |
 | `encrypt` | `userPassword`, `ownerPassword`, `algorithm`, `permissions`, `encryptMetadata` ([Encryption](#encryption-1)) |
 | `pdf.sign(field, o)` | `signer`, `reason`, `location`, `name`, `certify`, `timestamp`, `subFilter`, `reserve` ([Signatures](#signatures)) |
@@ -573,10 +573,12 @@ const logo = await pdf.embedImage(fs.readFileSync('logo.png'));         // alpha
 for (const y of [400, 500, 600]) page.image(logo, { x: 40, y, height: 24, opacity: 0.6 });   // written once
 ```
 
-- `await pdf.embedImage(bytes, { colorSpace, maxDecodedBytes })` — a JPEG or PNG, written to the file at once as an image XObject;
+- `await pdf.embedImage(bytes, { colorSpace, maxDecodedBytes, compression })` — a JPEG or PNG, written to the file at once as an image XObject;
   returns a handle with `width` and `height` (pixels, after the EXIF orientation). `colorSpace`: an ICC colour space
   (`pdf.iccColorSpace`) with the image's number of components, instead of the device space. `maxDecodedBytes`: the
-  decoded size allowed for a PNG that needs decoding (default 64 MB, 67,108,864 bytes).
+  decoded size allowed for a PNG that needs decoding (default 64 MB, 67,108,864 bytes). `compression`: the zlib level
+  for a PNG that is re-compressed (alpha or interlacing): `'default'` is 6, or 3 above 2 megapixels (about half the
+  time for 2–3% more bytes on photos); `'fast'` is 1; or a level 1–9. JPEGs and passed-through PNGs are not affected.
   - **JPEG**: baseline and progressive, gray, RGB and CMYK, embedded unchanged (DCTDecode, no re-encoding). Adobe's
     inverted CMYK (an APP14 "Adobe" segment) is drawn with `/Decode [1 0 1 0 1 0 1 0]`. The EXIF orientation (1–8)
     is honoured. 12-bit, lossless, hierarchical and arithmetic-coded JPEGs are refused.
@@ -596,7 +598,7 @@ for (const y of [400, 500, 600]) page.image(logo, { x: 40, y, height: 24, opacit
   PDFs: `alt` makes the image a `Figure` with that alternate text; without `alt` it is an artifact (decoration).
 - Images are deduplicated automatically: `embedImage` called again with the same bytes (and the same options) in the
   same document returns the same handle and writes nothing new (keyed by a SHA-256 of the bytes; Node's `node:crypto`,
-  else WebCrypto, else a fast hash confirmed byte for byte). A different option (`colorSpace`, `maxDecodedBytes`)
+  else WebCrypto, else a fast hash confirmed byte for byte). A different option (`colorSpace`, `maxDecodedBytes`, `compression`)
   embeds it again.
 - An image drawn any number of times, on any pages, is one XObject in the file. Images are always XObjects, never
   inline images: an inline image would be repeated in every content stream and cannot have a soft mask.

package/dist/index.js

@@ -957,29 +957,32 @@ var init_browser = __esm({
 });
 
 // src/core/deflate.js
-async function deflate(bytes) {
+async function deflate(bytes, level = 6) {
   if (slots.free > 0) slots.free--;
   else await new Promise((ok) => slots.wait.push(
     /** @type {() => void} */
     ok
   ));
   try {
-    return await now(bytes);
+    return await now(bytes, level);
   } finally {
     const next = slots.wait.shift();
     if (next) next();
     else slots.free++;
   }
 }
-async function now(bytes) {
-  if (NODE_ZLIB) return new Promise((ok, fail2) => NODE_ZLIB.deflate(bytes, { level: 6 }, (e, z) => e ? fail2(e) : ok(z)));
-  if (typeof CompressionStream === "undefined") return zlibSync(bytes, { level: 6 });
+async function now(bytes, level) {
+  if (NODE_ZLIB) return new Promise((ok, fail2) => NODE_ZLIB.deflate(bytes, { level }, (e, z) => e ? fail2(e) : ok(z)));
+  if (typeof CompressionStream === "undefined" || level !== 6) return zlibSync(bytes, { level: (
+    /** @type {any} */
+    level
+  ) });
   return new Uint8Array(await new Response(new Blob([
     /** @type {any} */
     bytes
   ]).stream().pipeThrough(new CompressionStream("deflate"))).arrayBuffer());
 }
-function deflateStream() {
+function deflateStream(level = 6) {
   const out = [];
   const join = () => {
     let n = 0;
@@ -993,7 +996,7 @@ function deflateStream() {
     return z;
   };
   if (NODE_ZLIB) {
-    const z = NODE_ZLIB.createDeflate({ level: 6 });
+    const z = NODE_ZLIB.createDeflate({ level, chunkSize: 1 << 20 });
     const ended = new Promise((ok, fail2) => {
       z.on("data", (c) => out.push(c));
       z.on("end", ok);
@@ -1017,7 +1020,19 @@ function deflateStream() {
   if (typeof CompressionStream === "undefined") return { async push(b) {
     out.push(b.slice());
   }, async end() {
-    return zlibSync(join(), { level: 6 });
+    return zlibSync(join(), { level: (
+      /** @type {any} */
+      level
+    ) });
+  }, abort() {
+  } };
+  if (level !== 6) return { async push(b) {
+    out.push(b.slice());
+  }, async end() {
+    return zlibSync(join(), { level: (
+      /** @type {any} */
+      level
+    ) });
   }, abort() {
   } };
   const cs = new CompressionStream("deflate"), w = cs.writable.getWriter();
@@ -3432,6 +3447,7 @@ var init_links = __esm({
 
 // src/images/png.js
 function crc32(b, from, to) {
+  if (NODE_ZLIB2?.crc32) return NODE_ZLIB2.crc32(b.subarray(from, to));
   if (!CRC_TABLE) {
     CRC_TABLE = new Uint32Array(256);
     for (let n = 0; n < 256; n++) {
@@ -3572,16 +3588,24 @@ function unfilterRow(f, c, p, bpp) {
   }
   for (let i2 = 0; i2 < bpp; i2++) c[i2] += p[i2];
   for (let i2 = bpp; i2 < rb; i2++) {
-    const a = c[i2 - bpp], u = p[i2], cc = p[i2 - bpp], pp = a + u - cc, pa = Math.abs(pp - a), pb = Math.abs(pp - u), pc = Math.abs(pp - cc);
-    c[i2] += pa <= pb && pa <= pc ? a : pb <= pc ? u : cc;
+    const a = c[i2 - bpp], u = p[i2], cc = p[i2 - bpp];
+    let pa = u - cc, pb = a - cc, pc = pa + pb, k;
+    k = pa >> 31;
+    pa = (pa ^ k) - k;
+    k = pb >> 31;
+    pb = (pb ^ k) - k;
+    k = pc >> 31;
+    pc = (pc ^ k) - k;
+    c[i2] += a ^ (a ^ (u ^ (u ^ cc) & pc - pb >> 31)) & (pb - pa | pc - pa) >> 31;
   }
 }
 function filterRow(x2, up, bpp, out, o, t) {
   const rb = x2.length, m = Math.min(bpp, rb);
-  let best = 0, sum = 0;
+  let best = 0, sum = 0, s, k;
   for (let i2 = 0; i2 < rb; i2++) {
-    const v = x2[i2];
-    sum += v < 128 ? v : 256 - v;
+    s = x2[i2] << 24 >> 24;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   out[o] = 0;
   out.set(x2, o + 1);
@@ -3590,12 +3614,15 @@ function filterRow(x2, up, bpp, out, o, t) {
   for (let i2 = 0; i2 < m; i2++) {
     const v = x2[i2];
     t[i2] = v;
-    sum += v < 128 ? v : 256 - v;
+    s = v << 24 >> 24;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   for (let i2 = m; i2 < rb; i2++) {
-    const v = x2[i2] - x2[i2 - bpp] & 255;
-    t[i2] = v;
-    sum += v < 128 ? v : 256 - v;
+    s = x2[i2] - x2[i2 - bpp] << 24 >> 24;
+    t[i2] = s;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   if (sum < best) {
     best = sum;
@@ -3605,9 +3632,10 @@ function filterRow(x2, up, bpp, out, o, t) {
   if (!up) return;
   sum = 0;
   for (let i2 = 0; i2 < rb; i2++) {
-    const v = x2[i2] - up[i2] & 255;
-    t[i2] = v;
-    sum += v < 128 ? v : 256 - v;
+    s = x2[i2] - up[i2] << 24 >> 24;
+    t[i2] = s;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   if (sum < best) {
     best = sum;
@@ -3616,15 +3644,25 @@ function filterRow(x2, up, bpp, out, o, t) {
   }
   sum = 0;
   for (let i2 = 0; i2 < m; i2++) {
-    const v = x2[i2] - up[i2] & 255;
-    t[i2] = v;
-    sum += v < 128 ? v : 256 - v;
+    s = x2[i2] - up[i2] << 24 >> 24;
+    t[i2] = s;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   for (let i2 = m; i2 < rb; i2++) {
-    const a = x2[i2 - bpp], u = up[i2], c = up[i2 - bpp], pa = Math.abs(u - c), pb = Math.abs(a - c), pc = Math.abs(a + u - c - c);
-    const v = x2[i2] - (pa <= pb && pa <= pc ? a : pb <= pc ? u : c) & 255;
-    t[i2] = v;
-    sum += v < 128 ? v : 256 - v;
+    const a = x2[i2 - bpp], u = up[i2], c = up[i2 - bpp];
+    let pa = u - c, pb = a - c, pc = pa + pb;
+    k = pa >> 31;
+    pa = (pa ^ k) - k;
+    k = pb >> 31;
+    pb = (pb ^ k) - k;
+    k = pc >> 31;
+    pc = (pc ^ k) - k;
+    const notA = (pb - pa | pc - pa) >> 31, uc = u ^ (u ^ c) & pc - pb >> 31;
+    s = x2[i2] - (a ^ (a ^ uc) & notA) << 24 >> 24;
+    t[i2] = s;
+    k = s >> 31;
+    sum += (s ^ k) - k;
   }
   if (sum < best) {
     out[o] = 4;
@@ -3857,13 +3895,14 @@ async function embedImage(doc, bytes, opt = {}) {
     /** @type {any} */
     opt[k2]
   ) : void 0;
-  const o = { colorSpace: own2("colorSpace"), maxDecodedBytes: own2("maxDecodedBytes") ?? MAX_DECODED };
+  const o = { colorSpace: own2("colorSpace"), maxDecodedBytes: own2("maxDecodedBytes") ?? MAX_DECODED, compression: own2("compression") ?? "default" };
+  if (!(o.compression === "default" || o.compression === "fast" || Number.isInteger(o.compression) && o.compression >= 1 && o.compression <= 9)) throw new TypeError("embedImage: compression is 'default', 'fast' or a zlib level 1–9");
   if (!(Number.isSafeInteger(o.maxDecodedBytes) && o.maxDecodedBytes > 0)) throw new TypeError("embedImage: maxDecodedBytes is a whole number of bytes > 0");
   if (b.length > MAX_INPUT) throw new RangeError(`embedImage: at most ${MAX_INPUT >> 20} MB`);
   let seen = SEEN.get(doc);
   if (!seen) SEEN.set(doc, seen = /* @__PURE__ */ new Map());
   const { key, copy } = await digest(b);
-  const k = `${key} ${o.colorSpace == null ? "-" : idOf(o.colorSpace)} ${o.maxDecodedBytes}`;
+  const k = `${key} ${o.colorSpace == null ? "-" : idOf(o.colorSpace)} ${o.maxDecodedBytes} ${o.compression}`;
   const hit = seen.get(k);
   if (hit && (!hit.copy || same(hit.copy, b))) return hit.img;
   const img = embed(doc, b, o);
@@ -3921,7 +3960,8 @@ async function embed(doc, b, o) {
     let color, a = null;
     if (p.filtered) color = [p.data, parms(ch, p.bpc)];
     else {
-      const ce = encoder(h, rb, bpp, parms(ch, p.bpc)), ae = p.alpha ? encoder(h, arb, Math.max(1, abpc >> 3), parms(1, abpc)) : null;
+      const lv = o.compression === "fast" ? 1 : o.compression === "default" ? w * h > 2e6 ? 3 : 6 : o.compression;
+      const ce = encoder(h, rb, bpp, parms(ch, p.bpc), lv), ae = p.alpha ? encoder(h, arb, Math.max(1, abpc >> 3), parms(1, abpc), lv) : null;
       try {
         await p.rows(async (c, al) => {
           await Promise.all([ce.row(c), ae && al ? ae.row(al) : null]);
@@ -3957,8 +3997,8 @@ async function embed(doc, b, o) {
     { ref, w: turned ? h : w, h: turned ? w : h, orient: ORIENT[orientation], alpha }
   );
 }
-function encoder(rows, rb, bpp, dp) {
-  const z = deflateStream(), per = Math.max(1, Math.floor((1 << 20) / (rb + 1))), t = new Uint8Array(rb);
+function encoder(rows, rb, bpp, dp, level) {
+  const z = deflateStream(level), per = Math.max(1, Math.floor((1 << 20) / (rb + 1))), t = new Uint8Array(rb);
   const plain2 = rows * rb <= 256 << 10 ? new Uint8Array(rows * rb) : null;
   const strips = [new Uint8Array(per * (rb + 1)), new Uint8Array(per * (rb + 1))], busy = [Promise.resolve(), Promise.resolve()];
   let s = 0, strip = strips[0], k = 0, y = 0;
@@ -3982,7 +4022,7 @@ function encoder(rows, rb, bpp, dp) {
       await busy[s ^ 1];
       const f = await z.end();
       if (!plain2) return [f, dp];
-      const p = await deflate(plain2);
+      const p = await deflate(plain2, level);
       return p.length <= f.length + dp.length ? [p, ""] : [f, dp];
     },
     abort: () => z.abort()
@@ -12364,13 +12404,13 @@ function createForm(doc, opt = {}) {
       if (!Array.isArray(list) || list.length > MAX_OPTIONS) throw new RangeError(`${who}: options is an array of at most ${MAX_OPTIONS} options`);
       const P = placeOf(o, who), L = lookOf(o, who, { border: 1 }), tip = tipOf(o, who);
       const { t, da } = textOf(P.page, o, who), q = qOf(o.align, who);
-      let chars2 = 0;
+      let chars3 = 0;
       const index = /* @__PURE__ */ new Map();
       const opts = list.map((it, i2) => {
         const pair = Array.isArray(it) ? it : [it, it];
         if (pair.length !== 2) throw new TypeError(`${who}: an option is a string or [export, display]`);
         const [ex, dp] = pair.map((v) => str2(v, "an option", MAX_TOOLTIP, who));
-        if ((chars2 += ex.length + dp.length) > MAX_OPTION_CHARS) throw new RangeError(`${who}: the options have more than ${MAX_OPTION_CHARS} characters in all`);
+        if ((chars3 += ex.length + dp.length) > MAX_OPTION_CHARS) throw new RangeError(`${who}: the options have more than ${MAX_OPTION_CHARS} characters in all`);
         if (index.has(ex)) throw new TypeError(`${who}: two options have the export value ${JSON.stringify(ex.slice(0, 80))}`);
         index.set(ex, i2);
         return { ex, dp };
@@ -13233,8 +13273,9 @@ function indexOf(buf, needle, from, to = buf.length) {
   to = Math.min(to, buf.length);
   if (to - from > 64) {
     if (NodeBuffer) {
-      const i2 = NodeBuffer.from(buf.buffer, buf.byteOffset, to).indexOf(needle, from, "latin1");
-      return i2;
+      let v = VIEWS.get(buf);
+      if (!v) VIEWS.set(buf, v = NodeBuffer.from(buf.buffer, buf.byteOffset, buf.length));
+      return (to === v.length ? v : v.subarray(0, to)).indexOf(needle, from, "latin1");
     }
     for (let i2 = buf.indexOf(n0, from); i2 >= 0 && i2 + L <= to; i2 = buf.indexOf(n0, i2 + 1)) {
       let j = 1;
@@ -13258,7 +13299,7 @@ function lastIndexOf(buf, needle, to, span = to) {
   }
   return -1;
 }
-var MAX_DEPTH4, MAX_ITEMS, Ref, PdfString, PdfStream, Op, CLASS, T, HEX, Lexer, parseValueEnd, NodeBuffer;
+var MAX_DEPTH4, MAX_ITEMS, Ref, PdfString, PdfStream, Op, CLASS, T, HEX, Lexer, parseValueEnd, NodeBuffer, VIEWS;
 var init_lexer = __esm({
   "src/reader/lexer.js"() {
     init_budget();
@@ -13429,7 +13470,7 @@ var init_lexer = __esm({
         }
         this.budget.tick(1 + (n >> 10));
         this.pos = i2;
-        this.val = new PdfString(out.slice(0, n), false);
+        this.val = new PdfString(n === out.length ? out : out.slice(0, n), false);
         return T.STR;
       }
       /** @param {number} i after "<" */
@@ -13451,7 +13492,7 @@ var init_lexer = __esm({
         if (hi >= 0) out[n++] = hi << 4;
         this.budget.tick(1 + (n >> 10));
         this.pos = j < e ? j + 1 : j;
-        this.val = new PdfString(out.slice(0, n), true);
+        this.val = new PdfString(n === out.length ? out : out.slice(0, n), true);
         return T.STR;
       }
       /** @param {number} i after "/" */
@@ -13483,6 +13524,8 @@ var init_lexer = __esm({
     parseValueEnd = () => void 0;
     NodeBuffer = /** @type {any} */
     globalThis.Buffer;
+    VIEWS = /** @type {WeakMap<Uint8Array, any>} */
+    /* @__PURE__ */ new WeakMap();
   }
 });
 
@@ -14644,14 +14687,26 @@ function decodeText(v) {
   if (!(v instanceof PdfString)) return typeof v === "string" ? v : v == null ? "" : String(v);
   const b = v.bytes;
   if (b.length >= 2 && b[0] === 254 && b[1] === 255) {
-    let s2 = "";
-    for (let i2 = 2; i2 + 1 < b.length; i2 += 2) s2 += String.fromCharCode(b[i2] << 8 | b[i2 + 1]);
-    return s2;
+    const u2 = new Uint16Array(b.length - 2 >> 1);
+    for (let i3 = 0; i3 < u2.length; i3++) u2[i3] = b[2 + 2 * i3] << 8 | b[3 + 2 * i3];
+    return chars2(u2);
   }
   if (b.length >= 3 && b[0] === 239 && b[1] === 187 && b[2] === 191) return new TextDecoder().decode(b.subarray(3));
-  let s = "";
-  for (const x2 of b) s += String.fromCharCode(DOC.get(x2) ?? x2);
-  return s;
+  let i2 = 0;
+  while (i2 < b.length && DOC_T[b[i2]] === b[i2]) i2++;
+  if (i2 === b.length) return latin12(b);
+  const u = new Uint16Array(b.length);
+  for (i2 = 0; i2 < b.length; i2++) u[i2] = DOC_T[b[i2]];
+  return chars2(u);
+}
+function chars2(u) {
+  const parts = [];
+  for (let i2 = 0; i2 < u.length; i2 += 8192) parts.push(String.fromCharCode.apply(
+    null,
+    /** @type {any} */
+    u.subarray(i2, i2 + 8192)
+  ));
+  return parts.join("");
 }
 function glyphText(n) {
   if (Object.hasOwn(NAMES2, n)) return NAMES2[n];
@@ -14689,11 +14744,11 @@ function parseCMap(data, budget, held = { n: 0 }) {
     for (let i2 = 0; i2 + 1 < b.length; i2 += 2) u.push(b[i2] << 8 | b[i2 + 1]);
     return String.fromCharCode(...u);
   };
-  let chars2 = 0;
+  let chars3 = 0;
   const set2 = (c, t) => {
-    chars2 += t.length;
+    chars3 += t.length;
     held.n += 8 + t.length;
-    if (chars2 > MAX_CMAP_CHARS) fail("E_BUDGET", `A CMap maps codes to more than ${MAX_CMAP_CHARS.toLocaleString("en")} characters`);
+    if (chars3 > MAX_CMAP_CHARS) fail("E_BUDGET", `A CMap maps codes to more than ${MAX_CMAP_CHARS.toLocaleString("en")} characters`);
     if (held.n > MAX_HELD) fail("E_BUDGET", `A page's fonts hold more than ${MAX_HELD.toLocaleString("en")} units of code mappings (an entry 8, a character 1)`);
     map.set(c, t);
   };
@@ -15012,12 +15067,30 @@ function pageText(src, pageDict, resources, raw = false) {
   const PX = [], PY = [], PE = [], PS = [], PG = [];
   const PA = [], PAE = [], PC = [], PD = [];
   const PT = [];
-  let chars2 = 0;
+  let chars3 = 0;
   let moved = true;
   const fonts = /* @__PURE__ */ new Map();
-  const add2 = (x2, y, end, size, text3, d0, d1, d2, d3) => {
+  let run2 = [];
+  let runs = false;
+  const flush = () => {
+    if (run2.length) {
+      PT[PT.length - 1] += run2.join("");
+      run2 = [];
+    }
+  };
+  const add2 = (x2, y, end, size, text3, d0, d1, d2, d3, glyph = false) => {
+    const n = PT.length - 1;
+    if (glyph && runs && !moved && !(PG[n] & 1) && PD[n] === d3 && PC[n] === d2 && PS[n] === size && d0 >= PA[n] && d0 >= PAE[n] - 0.01 && d0 - PAE[n] <= Math.max(size, 1) * 0.25 && !RTL2.test(text3)) {
+      if ((chars3 += text3.length) > MAX_TEXT_CHARS) throw new PdfReadError("E_BUDGET", `A page makes more than ${MAX_TEXT_CHARS.toLocaleString("en")} characters of text`);
+      run2.push(text3);
+      if (run2.length >= 4096) flush();
+      PAE[n] = Math.max(PAE[n], d1);
+      return;
+    }
+    flush();
+    runs = glyph && !raw && !RTL2.test(text3);
     if (PT.length >= budget.maxTextItems) throw new PdfReadError("E_BUDGET", `A page has more than ${budget.maxTextItems.toLocaleString("en")} pieces of text (budget.maxTextItems)`);
-    if ((chars2 += text3.length) > MAX_TEXT_CHARS) throw new PdfReadError("E_BUDGET", `A page makes more than ${MAX_TEXT_CHARS.toLocaleString("en")} characters of text`);
+    if ((chars3 += text3.length) > MAX_TEXT_CHARS) throw new PdfReadError("E_BUDGET", `A page makes more than ${MAX_TEXT_CHARS.toLocaleString("en")} characters of text`);
     if (raw) {
       PX.push(x2);
       PY.push(y);
@@ -15081,7 +15154,7 @@ function pageText(src, pageDict, resources, raw = false) {
       }
       const n = PT.length - 1;
       if (text3.includes("⁠")) text3 = text3.replaceAll("⁠", "");
-      if (text3) add2(x2, y, end, size, text3, d0, d1, d2, d3);
+      if (text3) add2(x2, y, end, size, text3, d0, d1, d2, d3, true);
       else if (n >= 0 && PAE[n] < d1 && PD[n] === d3 && Math.abs(PC[n] - d2) < 0.01) {
         if (raw) PE[n] = end;
         PAE[n] = d1;
@@ -15247,6 +15320,7 @@ function pageText(src, pageDict, resources, raw = false) {
       default:
     }
   }
+  flush();
   if (raw) return PT.map((text3, i2) => ({ x: PX[i2], y: PY[i2], end: PE[i2], size: PS[i2], text: text3, gap: !!(PG[i2] & 1) }));
   const out = [];
   for (let i2 = 0; i2 < PT.length; ) {
@@ -15273,7 +15347,7 @@ function pageText(src, pageDict, resources, raw = false) {
   }
   return out.join("");
 }
-var MAX_FORMS, MAX_CMAP, MAX_MARKS, MAX_ACTUAL, MAX_TEXT_CHARS, MAX_DST, MAX_CMAP_CHARS, MAX_HELD, MAX_W, MAX_DO, TEXT_TIMEOUT_MS, DOC, WIN, WIN_T, MAC_HIGH, BASE, STD_HIGH, NAMES2, ACCENT, own, STD_ALIAS, NONE, stops, MAX_CMAP_FONTS, mul6, mul2, ID, matrixOf, C0, G0, G1, RTL2, STRONG_L, NUM, rev2;
+var MAX_FORMS, MAX_CMAP, MAX_MARKS, MAX_ACTUAL, MAX_TEXT_CHARS, MAX_DST, MAX_CMAP_CHARS, MAX_HELD, MAX_W, MAX_DO, TEXT_TIMEOUT_MS, DOC, DOC_T, WIN, WIN_T, MAC_HIGH, BASE, STD_HIGH, NAMES2, ACCENT, own, STD_ALIAS, NONE, stops, MAX_CMAP_FONTS, mul6, mul2, ID, matrixOf, C0, G0, G1, RTL2, STRONG_L, NUM, rev2;
 var init_text2 = __esm({
   "src/reader/text.js"() {
     init_lexer();
@@ -15333,6 +15407,7 @@ var init_text2 = __esm({
       [158, 382],
       [160, 8364]
     ]);
+    DOC_T = new Uint16Array(256).map((_, c) => DOC.get(c) ?? c);
     WIN = /* @__PURE__ */ new Map();
     {
       const p = ENCODINGS.WinAnsi;
@@ -15732,9 +15807,15 @@ function numText(v) {
   return Math.abs(v) < 1 ? v.toFixed(10).replace(/\.?0+$/, "") || "0" : v.toFixed(0);
 }
 function hexText(b) {
-  let s = "<";
-  for (const x2 of b) s += HEXD[x2 >> 4] + HEXD[x2 & 15];
-  return s + ">";
+  if (NodeBuffer2 && b.length > 64) return `<${NodeBuffer2.from(b.buffer, b.byteOffset, b.length).toString("hex")}>`;
+  const parts = ["<"];
+  for (let i2 = 0; i2 < b.length; i2 += 4096) {
+    let s = "";
+    for (let k = i2; k < Math.min(b.length, i2 + 4096); k++) s += HEXD[b[k] >> 4] + HEXD[b[k] & 15];
+    parts.push(s);
+  }
+  parts.push(">");
+  return parts.join("");
 }
 function createCopier(t) {
   const maps = /* @__PURE__ */ new Map();
@@ -16136,7 +16217,7 @@ function createCopier(t) {
     }
   };
 }
-var SHARE, GRAY, HEXD, Raw, Foreign, AUDIT, OPEN_A, CLOSE_A, OPEN_D, CLOSE_D;
+var SHARE, GRAY, HEXD, NodeBuffer2, Raw, Foreign, AUDIT, OPEN_A, CLOSE_A, OPEN_D, CLOSE_D;
 var init_copy = __esm({
   "src/reader/copy.js"() {
     init_lexer();
@@ -16147,6 +16228,8 @@ var init_copy = __esm({
     SHARE = /* @__PURE__ */ new Set(["Font", "FontDescriptor", "ExtGState", "Encoding", "Pattern", "Shading", "Halftone", "Function", "XObject", "Mask", "CMap", "OutputIntent", "Filespec", "EmbeddedFile"]);
     GRAY = -2;
     HEXD = "0123456789abcdef";
+    NodeBuffer2 = /** @type {any} */
+    globalThis.Buffer;
     Raw = class {
       /** @param {string} s */
       constructor(s) {
@@ -16438,7 +16521,8 @@ async function run(self, sink, o) {
   if (level && o.encrypt) throw new TypeError("save: PDF/A forbids encryption: pass pdfa: false to drop the PDF/A claim and encrypt");
   const keepStruct = single && o.keepStructure !== false && !self.edits.flatten;
   const info = self.info();
-  const edited = !!self.edits.info || level?.conformance !== self.pdfa()?.conformance;
+  const tooBig = ["title", "author", "subject", "keywords"].some((k) => typeof info[k] === "string" && info[k].length > INFO_MAX);
+  const edited = !!self.edits.info || tooBig || level?.conformance !== self.pdfa()?.conformance;
   const uaLost = !keepStruct && /pdfuaid:part/.test(self.xmp() ?? "");
   const srcXmp = self.edits.xmp ?? (edited || uaLost || !level && self.pdfa() ? null : self.xmp());
   const created = o.creationDate ?? (info.creationDate instanceof Date ? info.creationDate : /* @__PURE__ */ new Date());
@@ -16451,6 +16535,13 @@ async function run(self, sink, o) {
       warn("the source's output intent profile was not usable; the bundled sRGB profile was written instead");
     }
   }
+  const infoText = (k) => {
+    const v = info[k];
+    if (v == null) return void 0;
+    if (v.length <= INFO_MAX) return v;
+    warn(`the source's ${k} (${v.length.toLocaleString("en")} characters) was cut to ${INFO_MAX.toLocaleString("en")}`);
+    return v.slice(0, INFO_MAX);
+  };
   const pdf = (
     /** @type {any} */
     createPdf(sink, {
@@ -16461,10 +16552,10 @@ async function run(self, sink, o) {
       creationDate: created,
       metadata: !!(level || self.edits.xmp != null || self.xmp() != null),
       // an XMP packet only when the source had one or PDF/A needs one: a file split into pages is not a kilobyte of XMP each
-      title: info.title ?? void 0,
-      author: info.author ?? void 0,
-      subject: info.subject ?? void 0,
-      keywords: info.keywords ?? void 0,
+      title: infoText("title"),
+      author: infoText("author"),
+      subject: infoText("subject"),
+      keywords: infoText("keywords"),
       ...level ? { pdfa: { part: level.part, conformance: "b", ...intentOpt ? { outputIntent: intentOpt } : {} } } : {}
     })
   );
@@ -16775,7 +16866,7 @@ function sameBytes(a, b) {
   for (let i2 = 0; i2 < a.length; i2++) if (a[i2] !== b[i2]) return false;
   return true;
 }
-var MAX_OPEN, KEYS2, OWN;
+var INFO_MAX, MAX_OPEN, KEYS2, OWN;
 var init_save = __esm({
   "src/reader/save.js"() {
     init_index();
@@ -16789,6 +16880,7 @@ var init_save = __esm({
     init_pdf();
     init_text2();
     init_budget();
+    INFO_MAX = 65536;
     MAX_OPEN = 256;
     KEYS2 = ["compress", "keepActiveContent", "encrypt", "pdfa", "creationDate", "signal", "timeoutMs", "keepStructure"];
     OWN = /* @__PURE__ */ new Set(["Type", "Parent", "Contents", "Resources", "Annots", "MediaBox", "CropBox", "Rotate", "B"]);
@@ -16866,17 +16958,19 @@ async function saveIncremental(self, o = {}) {
     keepStruct: () => true
   });
   const override = async (n, v) => put(n, await copier.text(self, v), null, v instanceof Map ? v : null);
-  let present;
-  const objects = [];
-  try {
-    present = findActive(self, (n, v) => {
+  let scan = SCANS.get(src);
+  if (!scan) try {
+    const objects2 = [];
+    const present2 = findActive(self, (n, v) => {
       const d = v instanceof PdfStream ? v.dict : v;
-      if (d instanceof Map && (d.has("ByteRange") || d.has("Perms") || d.has("Lock") || R(d.get("FT")) === "Sig" || d.has("V") || nestedSigs(d).length)) objects.push([n, v]);
+      if (d instanceof Map && (d.has("ByteRange") || d.has("Perms") || d.has("Lock") || R(d.get("FT")) === "Sig" || d.has("V") || nestedSigs(d).length)) objects2.push([n, v]);
     });
+    SCANS.set(src, scan = { present: present2, objects: objects2 });
   } catch (e) {
     throw new PdfReadError("E_ACTIVE", `${who}: the file could not be checked for active content (${/** @type {Error} */
     e.message}); use save() to rewrite the file without it (this invalidates existing signatures)`);
   }
+  const { present, objects } = scan;
   checkSigned(self, sig, o, objects, who);
   if (present.length && !o.keepActiveContent) throw new PdfReadError("E_ACTIVE", `${who}: the file has active content (${present.map((p) => `${p.what} ×${p.count}`).join(", ")}), which an incremental update cannot remove (the original bytes stay). Use save() to rewrite the file without it (this invalidates existing signatures), or pass keepActiveContent: true to keep it knowingly`);
   const fills = await fillReplacements(self, { pdfa: !!self.pdfa(), warn: (m) => warnings.push(m) });
@@ -17328,7 +17422,7 @@ function checkSigned(self, sig, o, objects, who) {
     }
   }
 }
-var SIGN_KEYS, MAX_WIDGET_SHARE, TRANSFORMS, isSigDict;
+var SIGN_KEYS, SCANS, MAX_WIDGET_SHARE, TRANSFORMS, isSigDict;
 var init_incremental = __esm({
   "src/reader/incremental.js"() {
     init_lexer();
@@ -17344,6 +17438,7 @@ var init_incremental = __esm({
     init_serialize();
     init_serialize();
     SIGN_KEYS = ["field", "signer", "reason", "location", "contactInfo", "name", "date", "timestamp", "subFilter", "reserve", "hash"];
+    SCANS = /* @__PURE__ */ new WeakMap();
     MAX_WIDGET_SHARE = 0.4;
     TRANSFORMS = /* @__PURE__ */ new Set(["DocMDP", "FieldMDP", "UR", "UR3"]);
     isSigDict = (d) => d instanceof Map && d.has("ByteRange") && (d.has("Contents") || d.has("Type") || d.has("Filter"));

package/index.d.ts

@@ -650,7 +650,7 @@ export interface Pdf {
    * textMapping: 'carrier' (default: Firefox, pdf.js, search) or 'actualText' (a shaped cluster's text only in its ActualText: MuPDF, AI pipelines) */
   embedFont(bytes: Uint8Array | ArrayBuffer, o?: { subset?: boolean | Subsetter; shaper?: Shaper; textMapping?: 'carrier' | 'actualText' }): Promise<Font>;
   /** a JPEG or PNG, written now as an image XObject. colorSpace: an ICC colour space with the image's component count */
-  embedImage(bytes: Uint8Array | ArrayBuffer, o?: { colorSpace?: IccColorSpace; /** the decoded size a PNG with alpha or interlacing may have; default 64 MB */ maxDecodedBytes?: number }): Promise<Image>;
+  embedImage(bytes: Uint8Array | ArrayBuffer, o?: { colorSpace?: IccColorSpace; /** the decoded size a PNG with alpha or interlacing may have; default 64 MB */ maxDecodedBytes?: number; /** zlib level for a re-compressed PNG (alpha, interlacing): 'default' = 6, or 3 above 2 megapixels; 'fast' = 1; or 1–9 */ compression?: 'default' | 'fast' | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 }): Promise<Image>;
   /** a spot colour: its name and the device colour (gray, RGB, CMYK) shown for it at full tint. One definition per name */
   spotColor(name: string, alternate: number | string | readonly number[]): SpotColor;
   /** an ICC-based colour space from a gray, RGB or CMYK profile (its header is checked) */

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@reportwright/pdf",
-  "version": "0.1.0-beta.3",
+  "version": "0.1.0-beta.5",
   "description": "A streaming PDF writer for Node and browsers with no runtime dependencies: flat memory, PDF/A-1b to 3a, PDF/UA-1, Factur-X, embedded TrueType with Unicode text.",
   "type": "module",
   "license": "MIT",