Bug 1754585 - Update pdf.js to version 2.13.133 r=pdfjs-reviewers,calixte

Differential Revision: https://phabricator.services.mozilla.com/D138359

browser/locales/en-US/pdfviewer/viewer.properties

@@ -168,9 +168,10 @@ find_previous.title=Find the previous occurrence of the phrase
 find_previous_label=Previous
 find_next.title=Find the next occurrence of the phrase
 find_next_label=Next
-find_highlight=Highlight all
+find_highlight=Highlight All
-find_match_case_label=Match case
+find_match_case_label=Match Case
-find_entire_word_label=Whole words
+find_match_diacritics_label=Match Diacritics
+find_entire_word_label=Whole Words
 find_reached_top=Reached top of document, continued from bottom
 find_reached_bottom=Reached end of document, continued from top
 # LOCALIZATION NOTE (find_match_count): The supported plural forms are

toolkit/components/pdfjs/README.mozilla

@@ -1,5 +1,5 @@
 This is the PDF.js project output, https://github.com/mozilla/pdf.js
 
-Current extension version is: 2.13.93
+Current extension version is: 2.13.133
 
-Taken from upstream commit: 6b9cc24d4
+Taken from upstream commit: f8b2a99dd

toolkit/components/pdfjs/content/build/pdf.js

@@ -1344,7 +1344,7 @@ async function _fetchDocument(worker, source, pdfDataRangeTransport, docId) {
 
   const workerId = await worker.messageHandler.sendWithPromise("GetDocRequest", {
     docId,
-    apiVersion: '2.13.93',
+    apiVersion: '2.13.133',
     source: {
       data: source.data,
       url: source.url,
@@ -1871,14 +1871,12 @@ class PDFPageProxy {
   }
 
   streamTextContent({
-    normalizeWhitespace = false,
     disableCombineTextItems = false,
     includeMarkedContent = false
   } = {}) {
     const TEXT_CONTENT_CHUNK_SIZE = 100;
     return this._transport.messageHandler.sendWithStream("GetTextContent", {
       pageIndex: this._pageIndex,
-      normalizeWhitespace: normalizeWhitespace === true,
       combineTextItems: disableCombineTextItems !== true,
       includeMarkedContent: includeMarkedContent === true
     }, {
@@ -2736,9 +2734,13 @@ class WorkerTransport {
 
       if (loadingTask.onPassword) {
         const updatePassword = password => {
-          this._passwordCapability.resolve({
+          if (password instanceof Error) {
-            password
+            this._passwordCapability.reject(password);
-          });
+          } else {
+            this._passwordCapability.resolve({
+              password
+            });
+          }
         };
 
         try {
@@ -3345,9 +3347,9 @@ class InternalRenderTask {
 
 }
 
-const version = '2.13.93';
+const version = '2.13.133';
 exports.version = version;
-const build = '6b9cc24d4';
+const build = 'f8b2a99dd';
 exports.build = build;
 
 /***/ }),
@@ -12382,8 +12384,8 @@ var _svg = __w_pdfjs_require__(22);
 
 var _xfa_layer = __w_pdfjs_require__(20);
 
-const pdfjsVersion = '2.13.93';
+const pdfjsVersion = '2.13.133';
-const pdfjsBuild = '6b9cc24d4';
+const pdfjsBuild = 'f8b2a99dd';
 ;
 })();
 

toolkit/components/pdfjs/content/build/pdf.scripting.js

@@ -4937,8 +4937,8 @@ Object.defineProperty(exports, "initSandbox", ({
 
 var _initialization = __w_pdfjs_require__(1);
 
-const pdfjsVersion = '2.13.93';
+const pdfjsVersion = '2.13.133';
-const pdfjsBuild = '6b9cc24d4';
+const pdfjsBuild = 'f8b2a99dd';
 })();
 
 /******/ 	return __webpack_exports__;

toolkit/components/pdfjs/content/web/viewer.css

@@ -1260,7 +1260,7 @@ html[dir="rtl"] #outerContainer.sidebarOpen #loadingBar {
 .findbar.wrapContainers > div {
   clear: both;
 }
-.findbar.wrapContainers > div#findbarMessageContainer {
+.findbar.wrapContainers > div.findbarMessageContainer {
   height: auto;
 }
 html[dir="ltr"] .findbar {
@@ -1317,7 +1317,7 @@ html[dir="rtl"] .findbar .splitToolbarButton > .findNext {
 }
 
 .findbar label:hover,
-.findbar input:focus + label {
+.findbar input:focus-visible + label {
   color: var(--toggled-btn-color);
   background-color: var(--button-hover-color);
 }
@@ -1445,6 +1445,8 @@ html[dir="ltr"] .doorHangerRight:before {
 #findMsg {
   color: rgba(251, 0, 0, 1);
 }
+
+#findResultsCount:empty,
 #findMsg:empty {
   display: none;
 }
@@ -1507,12 +1509,12 @@ html[dir="rtl"] .splitToolbarButton > .toolbarButton {
 }
 
 .overlayButton:hover,
-.overlayButton:focus {
+.overlayButton:focus-visible {
   background-color: var(--overlay-button-hover-bg-color);
 }
 
 .overlayButton:hover > span,
-.overlayButton:focus > span {
+.overlayButton:focus-visible > span {
   color: var(--overlay-button-hover-color);
 }
 
@@ -1534,10 +1536,10 @@ html[dir="rtl"] .splitToolbarButton > .toolbarButton {
 }
 
 .splitToolbarButton > .toolbarButton:hover,
-.splitToolbarButton > .toolbarButton:focus,
+.splitToolbarButton > .toolbarButton:focus-visible,
 .dropdownToolbarButton:hover,
 .toolbarButton.textButton:hover,
-.toolbarButton.textButton:focus {
+.toolbarButton.textButton:focus-visible {
   background-color: var(--button-hover-color);
   z-index: 199;
 }
@@ -1601,11 +1603,11 @@ html[dir="rtl"] #toolbarViewerLeft > .toolbarButton:first-child {
   margin-right: 2px;
 }
 .toolbarButton:hover,
-.toolbarButton:focus {
+.toolbarButton:focus-visible {
   background-color: var(--button-hover-color);
 }
 .secondaryToolbarButton:hover,
-.secondaryToolbarButton:focus {
+.secondaryToolbarButton:focus-visible {
   background-color: var(--doorhanger-hover-bg-color);
   color: var(--doorhanger-hover-color);
 }
@@ -1670,7 +1672,7 @@ html[dir="rtl"] .dropdownToolbarButton > select {
   padding-right: 4px;
 }
 .dropdownToolbarButton > select:hover,
-.dropdownToolbarButton > select:focus {
+.dropdownToolbarButton > select:focus-visible {
   background-color: var(--button-hover-color);
   color: var(--toggled-btn-color);
 }
@@ -1711,7 +1713,7 @@ html[dir="rtl"] .dropdownToolbarButton > select {
 }
 
 .dropdownToolbarButton:hover::after,
-.dropdownToolbarButton:focus::after,
+.dropdownToolbarButton:focus-visible::after,
 .dropdownToolbarButton:active::after {
   background-color: var(--toolbar-icon-hover-bg-color);
 }
@@ -1723,9 +1725,9 @@ html[dir="rtl"] .dropdownToolbarButton > select {
 }
 
 .toolbarButton:hover::before,
-.toolbarButton:focus::before,
+.toolbarButton:focus-visible::before,
 .secondaryToolbarButton:hover::before,
-.secondaryToolbarButton:focus::before {
+.secondaryToolbarButton:focus-visible::before {
   background-color: var(--toolbar-icon-hover-bg-color);
 }
 

toolkit/components/pdfjs/content/web/viewer.html

@@ -101,17 +101,21 @@ See https://github.com/adobe-type-tools/cmap-resources
 
           <div id="findbarOptionsOneContainer">
             <input type="checkbox" id="findHighlightAll" class="toolbarField" tabindex="94">
-            <label for="findHighlightAll" class="toolbarLabel" data-l10n-id="find_highlight">Highlight all</label>
+            <label for="findHighlightAll" class="toolbarLabel" data-l10n-id="find_highlight">Highlight All</label>
             <input type="checkbox" id="findMatchCase" class="toolbarField" tabindex="95">
-            <label for="findMatchCase" class="toolbarLabel" data-l10n-id="find_match_case_label">Match case</label>
+            <label for="findMatchCase" class="toolbarLabel" data-l10n-id="find_match_case_label">Match Case</label>
           </div>
           <div id="findbarOptionsTwoContainer">
-            <input type="checkbox" id="findEntireWord" class="toolbarField" tabindex="96">
+            <input type="checkbox" id="findMatchDiacritics" class="toolbarField" tabindex="96">
-            <label for="findEntireWord" class="toolbarLabel" data-l10n-id="find_entire_word_label">Whole words</label>
+            <label for="findMatchDiacritics" class="toolbarLabel" data-l10n-id="find_match_diacritics_label">Match Diacritics</label>
-            <span id="findResultsCount" class="toolbarLabel hidden"></span>
+            <input type="checkbox" id="findEntireWord" class="toolbarField" tabindex="97">
+            <label for="findEntireWord" class="toolbarLabel" data-l10n-id="find_entire_word_label">Whole Words</label>
           </div>
 
-          <div id="findbarMessageContainer">
+          <div class="findbarMessageContainer">
+            <span id="findResultsCount" class="toolbarLabel"></span>
+          </div>
+          <div class="findbarMessageContainer">
             <span id="findMsg" class="toolbarLabel"></span>
           </div>
         </div>  <!-- findbar -->

toolkit/components/pdfjs/content/web/viewer.js

@@ -2448,7 +2448,8 @@ function webViewerFindFromUrlHash(evt) {
     caseSensitive: false,
     entireWord: false,
     highlightAll: true,
-    findPrevious: false
+    findPrevious: false,
+    matchDiacritics: true
   });
 }
 
@@ -3105,8 +3106,8 @@ function removeNullCharacters(str, replaceInvisible = false) {
   return str.replace(NullCharactersRegExp, "");
 }
 
-function binarySearchFirstItem(items, condition) {
+function binarySearchFirstItem(items, condition, start = 0) {
-  let minIndex = 0;
+  let minIndex = start;
   let maxIndex = items.length - 1;
 
   if (maxIndex < 0 || !condition(items[maxIndex])) {
@@ -4632,14 +4633,14 @@ class PasswordPrompt {
     this._isViewerEmbedded = isViewerEmbedded;
     this.updateCallback = null;
     this.reason = null;
-    this.submitButton.addEventListener("click", this.verify.bind(this));
+    this.submitButton.addEventListener("click", this.#verify.bind(this));
-    this.cancelButton.addEventListener("click", this.close.bind(this));
+    this.cancelButton.addEventListener("click", this.#cancel.bind(this));
     this.input.addEventListener("keydown", e => {
       if (e.keyCode === 13) {
-        this.verify();
+        this.#verify();
       }
     });
-    this.overlayManager.register(this.overlayName, this.container, this.close.bind(this), true);
+    this.overlayManager.register(this.overlayName, this.container, this.#cancel.bind(this), true);
   }
 
   async open() {
@@ -4659,7 +4660,7 @@ class PasswordPrompt {
     });
   }
 
-  verify() {
+  #verify() {
     const password = this.input.value;
 
     if (password?.length > 0) {
@@ -4668,6 +4669,11 @@ class PasswordPrompt {
     }
   }
 
+  #cancel() {
+    this.close();
+    this.updateCallback(new Error("PasswordPrompt cancelled."));
+  }
+
   setUpdateCallback(updateCallback, reason) {
     this.updateCallback = updateCallback;
     this.reason = reason;
@@ -5261,6 +5267,7 @@ class PDFFindBar {
     this.findField = options.findField;
     this.highlightAll = options.highlightAllCheckbox;
     this.caseSensitive = options.caseSensitiveCheckbox;
+    this.matchDiacritics = options.matchDiacriticsCheckbox;
     this.entireWord = options.entireWordCheckbox;
     this.findMsg = options.findMsg;
     this.findResultsCount = options.findResultsCount;
@@ -5303,6 +5310,9 @@ class PDFFindBar {
     this.entireWord.addEventListener("click", () => {
       this.dispatchEvent("entirewordchange");
     });
+    this.matchDiacritics.addEventListener("click", () => {
+      this.dispatchEvent("diacriticmatchingchange");
+    });
 
     this.eventBus._on("resize", this._adjustWidth.bind(this));
   }
@@ -5320,7 +5330,8 @@ class PDFFindBar {
       caseSensitive: this.caseSensitive.checked,
       entireWord: this.entireWord.checked,
       highlightAll: this.highlightAll.checked,
-      findPrevious: findPrev
+      findPrevious: findPrev,
+      matchDiacritics: this.matchDiacritics.checked
     });
   }
 
@@ -5381,7 +5392,6 @@ class PDFFindBar {
 
     matchCountMsg.then(msg => {
       this.findResultsCount.textContent = msg;
-      this.findResultsCount.classList.toggle("hidden", !total);
 
       this._adjustWidth();
     });
@@ -5452,12 +5462,12 @@ Object.defineProperty(exports, "__esModule", ({
 }));
 exports.PDFFindController = exports.FindState = void 0;
 
+var _ui_utils = __webpack_require__(3);
+
 var _pdfjsLib = __webpack_require__(5);
 
 var _pdf_find_utils = __webpack_require__(16);
 
-var _ui_utils = __webpack_require__(3);
-
 const FindState = {
   FOUND: 0,
   NOT_FOUND: 1,
@@ -5482,51 +5492,105 @@ const CHARACTERS_TO_NORMALIZE = {
   "\u00BD": "1/2",
   "\u00BE": "3/4"
 };
+const DIACRITICS_EXCEPTION = new Set([0x3099, 0x309a, 0x094d, 0x09cd, 0x0a4d, 0x0acd, 0x0b4d, 0x0bcd, 0x0c4d, 0x0ccd, 0x0d3b, 0x0d3c, 0x0d4d, 0x0dca, 0x0e3a, 0x0eba, 0x0f84, 0x1039, 0x103a, 0x1714, 0x1734, 0x17d2, 0x1a60, 0x1b44, 0x1baa, 0x1bab, 0x1bf2, 0x1bf3, 0x2d7f, 0xa806, 0xa82c, 0xa8c4, 0xa953, 0xa9c0, 0xaaf6, 0xabed, 0x0c56, 0x0f71, 0x0f72, 0x0f7a, 0x0f7b, 0x0f7c, 0x0f7d, 0x0f80, 0x0f74]);
+const DIACRITICS_EXCEPTION_STR = [...DIACRITICS_EXCEPTION.values()].map(x => String.fromCharCode(x)).join("");
+const DIACRITICS_REG_EXP = /\p{M}+/gu;
+const SPECIAL_CHARS_REG_EXP = /([.*+?^${}()|[\]\\])|(\p{P})|(\s+)|(\p{M})|(\p{L})/gu;
+const NOT_DIACRITIC_FROM_END_REG_EXP = /([^\p{M}])\p{M}*$/u;
+const NOT_DIACRITIC_FROM_START_REG_EXP = /^\p{M}*([^\p{M}])/u;
 let normalizationRegex = null;
 
 function normalize(text) {
   if (!normalizationRegex) {
     const replace = Object.keys(CHARACTERS_TO_NORMALIZE).join("");
-    normalizationRegex = new RegExp(`[${replace}]`, "g");
+    normalizationRegex = new RegExp(`([${replace}])|(\\S-\\n)|(\\n)|(\\p{M}+)`, "gum");
   }
 
-  let diffs = null;
+  const rawDiacriticsPositions = [];
-  const normalizedText = text.replace(normalizationRegex, function (ch, index) {
+  let m;
-    const normalizedCh = CHARACTERS_TO_NORMALIZE[ch],
-          diff = normalizedCh.length - ch.length;
 
-    if (diff !== 0) {
+  while ((m = DIACRITICS_REG_EXP.exec(text)) !== null) {
-      (diffs ||= []).push([index, diff]);
+    rawDiacriticsPositions.push([m[0].length, m.index]);
+  }
+
+  let normalized = text.normalize("NFD");
+  const positions = [[0, 0]];
+  let k = 0;
+  let shift = 0;
+  let shiftOrigin = 0;
+  let eol = 0;
+  let hasDiacritics = false;
+  normalized = normalized.replace(normalizationRegex, (match, p1, p2, p3, p4, i) => {
+    i -= shiftOrigin;
+
+    if (p1) {
+      const replacement = CHARACTERS_TO_NORMALIZE[match];
+      const jj = replacement.length;
+
+      for (let j = 1; j < jj; j++) {
+        positions.push([i - shift + j, shift - j]);
+      }
+
+      shift -= jj - 1;
+      return replacement;
     }
 
-    return normalizedCh;
+    if (p2) {
+      positions.push([i - shift, 1 + shift]);
+      shift += 1;
+      shiftOrigin += 1;
+      eol += 1;
+      return p2.charAt(0);
+    }
+
+    if (p3) {
+      positions.push([i - shift + 1, shift - 1]);
+      shift -= 1;
+      shiftOrigin += 1;
+      eol += 1;
+      return " ";
+    }
+
+    hasDiacritics = true;
+    let jj = p4.length;
+
+    if (i + eol === rawDiacriticsPositions[k]?.[1]) {
+      jj -= rawDiacriticsPositions[k][0];
+      ++k;
+    }
+
+    for (let j = 1; j < jj + 1; j++) {
+      positions.push([i - 1 - shift + j, shift - j]);
+    }
+
+    shift -= jj;
+    shiftOrigin += jj;
+    return p4;
   });
-  return [normalizedText, diffs];
+  positions.push([normalized.length, shift]);
+  return [normalized, positions, hasDiacritics];
 }
 
-function getOriginalIndex(matchIndex, diffs = null) {
+function getOriginalIndex(diffs, pos, len) {
   if (!diffs) {
-    return matchIndex;
+    return [pos, len];
   }
 
-  let totalDiff = 0;
+  const start = pos;
+  const end = pos + len;
+  let i = (0, _ui_utils.binarySearchFirstItem)(diffs, x => x[0] >= start);
 
-  for (const [index, diff] of diffs) {
+  if (diffs[i][0] > start) {
-    const currentIndex = index + totalDiff;
+    --i;
-
-    if (currentIndex >= matchIndex) {
-      break;
-    }
-
-    if (currentIndex + diff > matchIndex) {
-      totalDiff += matchIndex - currentIndex;
-      break;
-    }
-
-    totalDiff += diff;
   }
 
-  return matchIndex - totalDiff;
+  let j = (0, _ui_utils.binarySearchFirstItem)(diffs, x => x[0] >= end, i);
+
+  if (diffs[j][0] > end) {
+    --j;
+  }
+
+  return [start + diffs[i][1], len + diffs[j][1] - diffs[i][1]];
 }
 
 class PDFFindController {
@@ -5682,6 +5746,7 @@ class PDFFindController {
     this._extractTextPromises = [];
     this._pageContents = [];
     this._pageDiffs = [];
+    this._hasDiacritics = [];
     this._matchesCountTotal = 0;
     this._pagesToSearch = null;
     this._pendingFindMatches = new Set();
@@ -5724,65 +5789,23 @@ class PDFFindController {
     return true;
   }
 
-  _prepareMatches(matchesWithLength, matches, matchesLength) {
-    function isSubTerm(currentIndex) {
-      const currentElem = matchesWithLength[currentIndex];
-      const nextElem = matchesWithLength[currentIndex + 1];
-
-      if (currentIndex < matchesWithLength.length - 1 && currentElem.match === nextElem.match) {
-        currentElem.skipped = true;
-        return true;
-      }
-
-      for (let i = currentIndex - 1; i >= 0; i--) {
-        const prevElem = matchesWithLength[i];
-
-        if (prevElem.skipped) {
-          continue;
-        }
-
-        if (prevElem.match + prevElem.matchLength < currentElem.match) {
-          break;
-        }
-
-        if (prevElem.match + prevElem.matchLength >= currentElem.match + currentElem.matchLength) {
-          currentElem.skipped = true;
-          return true;
-        }
-      }
-
-      return false;
-    }
-
-    matchesWithLength.sort(function (a, b) {
-      return a.match === b.match ? a.matchLength - b.matchLength : a.match - b.match;
-    });
-
-    for (let i = 0, len = matchesWithLength.length; i < len; i++) {
-      if (isSubTerm(i)) {
-        continue;
-      }
-
-      matches.push(matchesWithLength[i].match);
-      matchesLength.push(matchesWithLength[i].matchLength);
-    }
-  }
-
   _isEntireWord(content, startIdx, length) {
-    if (startIdx > 0) {
+    let match = content.slice(0, startIdx).match(NOT_DIACRITIC_FROM_END_REG_EXP);
+
+    if (match) {
       const first = content.charCodeAt(startIdx);
-      const limit = content.charCodeAt(startIdx - 1);
+      const limit = match[1].charCodeAt(0);
 
       if ((0, _pdf_find_utils.getCharacterType)(first) === (0, _pdf_find_utils.getCharacterType)(limit)) {
         return false;
       }
     }
 
-    const endIdx = startIdx + length - 1;
+    match = content.slice(startIdx + length).match(NOT_DIACRITIC_FROM_START_REG_EXP);
 
-    if (endIdx < content.length - 1) {
+    if (match) {
-      const last = content.charCodeAt(endIdx);
+      const last = content.charCodeAt(startIdx + length - 1);
-      const limit = content.charCodeAt(endIdx + 1);
+      const limit = match[1].charCodeAt(0);
 
       if ((0, _pdf_find_utils.getCharacterType)(last) === (0, _pdf_find_utils.getCharacterType)(limit)) {
         return false;
@@ -5792,96 +5815,114 @@ class PDFFindController {
     return true;
   }
 
-  _calculatePhraseMatch(query, pageIndex, pageContent, pageDiffs, entireWord) {
+  _calculateRegExpMatch(query, entireWord, pageIndex, pageContent) {
     const matches = [],
           matchesLength = [];
-    const queryLen = query.length;
+    const diffs = this._pageDiffs[pageIndex];
-    let matchIdx = -queryLen;
+    let match;
 
-    while (true) {
+    while ((match = query.exec(pageContent)) !== null) {
-      matchIdx = pageContent.indexOf(query, matchIdx + queryLen);
+      if (entireWord && !this._isEntireWord(pageContent, match.index, match[0].length)) {
-
-      if (matchIdx === -1) {
-        break;
-      }
-
-      if (entireWord && !this._isEntireWord(pageContent, matchIdx, queryLen)) {
         continue;
       }
 
-      const originalMatchIdx = getOriginalIndex(matchIdx, pageDiffs),
+      const [matchPos, matchLen] = getOriginalIndex(diffs, match.index, match[0].length);
-            matchEnd = matchIdx + queryLen - 1,
+
-            originalQueryLen = getOriginalIndex(matchEnd, pageDiffs) - originalMatchIdx + 1;
+      if (matchLen) {
-      matches.push(originalMatchIdx);
+        matches.push(matchPos);
-      matchesLength.push(originalQueryLen);
+        matchesLength.push(matchLen);
+      }
     }
 
     this._pageMatches[pageIndex] = matches;
     this._pageMatchesLength[pageIndex] = matchesLength;
   }
 
-  _calculateWordMatch(query, pageIndex, pageContent, pageDiffs, entireWord) {
+  _convertToRegExpString(query, hasDiacritics) {
-    const matchesWithLength = [];
+    const {
-    const queryArray = query.match(/\S+/g);
+      matchDiacritics
+    } = this._state;
+    let isUnicode = false;
+    query = query.replace(SPECIAL_CHARS_REG_EXP, (match, p1, p2, p3, p4, p5) => {
+      if (p1) {
+        return `[ ]*\\${p1}[ ]*`;
+      }
 
-    for (let i = 0, len = queryArray.length; i < len; i++) {
+      if (p2) {
-      const subquery = queryArray[i];
+        return `[ ]*${p2}[ ]*`;
-      const subqueryLen = subquery.length;
+      }
-      let matchIdx = -subqueryLen;
 
-      while (true) {
+      if (p3) {
-        matchIdx = pageContent.indexOf(subquery, matchIdx + subqueryLen);
+        return "[ ]+";
+      }
 
-        if (matchIdx === -1) {
+      if (matchDiacritics) {
-          break;
+        return p4 || p5;
-        }
+      }
 
-        if (entireWord && !this._isEntireWord(pageContent, matchIdx, subqueryLen)) {
+      if (p4) {
-          continue;
+        return DIACRITICS_EXCEPTION.has(p4.charCodeAt(0)) ? p4 : "";
-        }
+      }
 
-        const originalMatchIdx = getOriginalIndex(matchIdx, pageDiffs),
+      if (hasDiacritics) {
-              matchEnd = matchIdx + subqueryLen - 1,
+        isUnicode = true;
-              originalQueryLen = getOriginalIndex(matchEnd, pageDiffs) - originalMatchIdx + 1;
+        return `${p5}\\p{M}*`;
-        matchesWithLength.push({
+      }
-          match: originalMatchIdx,
+
-          matchLength: originalQueryLen,
+      return p5;
-          skipped: false
+    });
-        });
+    const trailingSpaces = "[ ]*";
+
+    if (query.endsWith(trailingSpaces)) {
+      query = query.slice(0, query.length - trailingSpaces.length);
+    }
+
+    if (matchDiacritics) {
+      if (hasDiacritics) {
+        isUnicode = true;
+        query = `${query}(?=[${DIACRITICS_EXCEPTION_STR}]|[^\\p{M}]|$)`;
       }
     }
 
-    this._pageMatchesLength[pageIndex] = [];
+    return [isUnicode, query];
-    this._pageMatches[pageIndex] = [];
-
-    this._prepareMatches(matchesWithLength, this._pageMatches[pageIndex], this._pageMatchesLength[pageIndex]);
   }
 
   _calculateMatch(pageIndex) {
-    let pageContent = this._pageContents[pageIndex];
-    const pageDiffs = this._pageDiffs[pageIndex];
     let query = this._query;
-    const {
-      caseSensitive,
-      entireWord,
-      phraseSearch
-    } = this._state;
 
     if (query.length === 0) {
       return;
     }
 
-    if (!caseSensitive) {
+    const {
-      pageContent = pageContent.toLowerCase();
+      caseSensitive,
-      query = query.toLowerCase();
+      entireWord,
-    }
+      phraseSearch
+    } = this._state;
+    const pageContent = this._pageContents[pageIndex];
+    const hasDiacritics = this._hasDiacritics[pageIndex];
+    let isUnicode = false;
 
     if (phraseSearch) {
-      this._calculatePhraseMatch(query, pageIndex, pageContent, pageDiffs, entireWord);
+      [isUnicode, query] = this._convertToRegExpString(query, hasDiacritics);
     } else {
-      this._calculateWordMatch(query, pageIndex, pageContent, pageDiffs, entireWord);
+      const match = query.match(/\S+/g);
+
+      if (match) {
+        query = match.sort().reverse().map(q => {
+          const [isUnicodePart, queryPart] = this._convertToRegExpString(q, hasDiacritics);
+
+          isUnicode ||= isUnicodePart;
+          return `(${queryPart})`;
+        }).join("|");
+      }
     }
 
+    const flags = `g${isUnicode ? "u" : ""}${caseSensitive ? "" : "i"}`;
+    query = new RegExp(query, flags);
+
+    this._calculateRegExpMatch(query, entireWord, pageIndex, pageContent);
+
     if (this._state.highlightAll) {
       this._updatePage(pageIndex);
     }
@@ -5913,23 +5954,26 @@ class PDFFindController {
       this._extractTextPromises[i] = extractTextCapability.promise;
       promise = promise.then(() => {
         return this._pdfDocument.getPage(i + 1).then(pdfPage => {
-          return pdfPage.getTextContent({
+          return pdfPage.getTextContent();
-            normalizeWhitespace: true
-          });
         }).then(textContent => {
           const textItems = textContent.items;
           const strBuf = [];
 
           for (let j = 0, jj = textItems.length; j < jj; j++) {
             strBuf.push(textItems[j].str);
+
+            if (textItems[j].hasEOL) {
+              strBuf.push("\n");
+            }
           }
 
-          [this._pageContents[i], this._pageDiffs[i]] = normalize(strBuf.join(""));
+          [this._pageContents[i], this._pageDiffs[i], this._hasDiacritics[i]] = normalize(strBuf.join(""));
           extractTextCapability.resolve(i);
         }, reason => {
           console.error(`Unable to get text content for page ${i + 1}`, reason);
           this._pageContents[i] = "";
           this._pageDiffs[i] = null;
+          this._hasDiacritics[i] = false;
           extractTextCapability.resolve(i);
         });
       });
@@ -9745,7 +9789,7 @@ class BaseViewer {
       throw new Error("Cannot initialize BaseViewer.");
     }
 
-    const viewerVersion = '2.13.93';
+    const viewerVersion = '2.13.133';
 
     if (_pdfjsLib.version !== viewerVersion) {
       throw new Error(`The API version "${_pdfjsLib.version}" does not match the Viewer version "${viewerVersion}".`);
@@ -12184,7 +12228,6 @@ class PDFPageView {
       return finishPaintTask(null).then(() => {
         if (textLayer) {
           const readableStream = pdfPage.streamTextContent({
-            normalizeWhitespace: true,
             includeMarkedContent: true
           });
           textLayer.setTextContentStream(readableStream);
@@ -13878,7 +13921,7 @@ class MozL10n {
 }
 
 (function listenFindEvents() {
-  const events = ["find", "findagain", "findhighlightallchange", "findcasesensitivitychange", "findentirewordchange", "findbarclose"];
+  const events = ["find", "findagain", "findhighlightallchange", "findcasesensitivitychange", "findentirewordchange", "findbarclose", "finddiacriticmatchingchange"];
   const findLen = "find".length;
 
   const handleEvent = function ({
@@ -13905,7 +13948,8 @@ class MozL10n {
       caseSensitive: !!detail.caseSensitive,
       entireWord: !!detail.entireWord,
       highlightAll: !!detail.highlightAll,
-      findPrevious: !!detail.findPrevious
+      findPrevious: !!detail.findPrevious,
+      matchDiacritics: !!detail.matchDiacritics
     });
   };
 
@@ -14593,8 +14637,8 @@ var _app_options = __webpack_require__(1);
 
 var _app = __webpack_require__(2);
 
-const pdfjsVersion = '2.13.93';
+const pdfjsVersion = '2.13.133';
-const pdfjsBuild = '6b9cc24d4';
+const pdfjsBuild = 'f8b2a99dd';
 window.PDFViewerApplication = _app.PDFViewerApplication;
 window.PDFViewerApplicationOptions = _app_options.AppOptions;
 ;
@@ -14679,6 +14723,7 @@ function getViewerConfiguration() {
       findField: document.getElementById("findInput"),
       highlightAllCheckbox: document.getElementById("findHighlightAll"),
       caseSensitiveCheckbox: document.getElementById("findMatchCase"),
+      matchDiacriticsCheckbox: document.getElementById("findMatchDiacritics"),
       entireWordCheckbox: document.getElementById("findEntireWord"),
       findMsg: document.getElementById("findMsg"),
       findResultsCount: document.getElementById("findResultsCount"),

toolkit/components/pdfjs/moz.yaml

@@ -20,7 +20,7 @@ origin:
 
   # Human-readable identifier for this version/release
   # Generally "version NNN", "tag SSS", "bookmark SSS"
-  release: version 2.13.93
+  release: version 2.13.133
 
   # The package's license, where possible using the mnemonic from
   # https://spdx.org/licenses/