Merge pull request #16247 from Snuffleupagus/issue-7442

[api-minor] Add support, in `PDFFindController`, for mixing phrase/word searches (issue 7442)

test/unit/pdf_find_controller_spec.js

@@ -95,7 +95,6 @@ function testSearch({
         query: null,
         caseSensitive: false,
         entireWord: false,
-        phraseSearch: true,
         findPrevious: false,
         matchDiacritics: false,
       },
@@ -182,7 +181,6 @@ function testEmptySearch({ eventBus, pdfFindController, state }) {
         query: null,
         caseSensitive: false,
         entireWord: false,
-        phraseSearch: true,
         findPrevious: false,
         matchDiacritics: false,
       },
@@ -321,8 +319,7 @@ describe("pdf_find_controller", function () {
       eventBus,
       pdfFindController,
       state: {
-        query: "alternate solution",
+        query: ["alternate", "solution"],
-        phraseSearch: false,
       },
       matchesPerPage: [0, 0, 0, 0, 0, 1, 0, 0, 4, 0, 0, 0, 0, 0],
       selectedMatch: {
@@ -332,6 +329,25 @@ describe("pdf_find_controller", function () {
     });
   });
 
+  it("performs a multiple term (phrase) search", async function () {
+    // Page 9 contains 'alternate solution' and pages 6 and 9 contain
+    // 'solution'. Both should be found for multiple term (phrase) search.
+    const { eventBus, pdfFindController } = await initPdfFindController();
+
+    await testSearch({
+      eventBus,
+      pdfFindController,
+      state: {
+        query: ["alternate solution", "solution"],
+      },
+      matchesPerPage: [0, 0, 0, 0, 0, 1, 0, 0, 3, 0, 0, 0, 0, 0],
+      selectedMatch: {
+        pageIndex: 5,
+        matchIndex: 0,
+      },
+    });
+  });
+
   it("performs a normal search, where the text is normalized", async function () {
     const { eventBus, pdfFindController } = await initPdfFindController(
       "fraction-highlight.pdf"

web/app.js

@@ -2575,7 +2575,6 @@ function webViewerFindFromUrlHash(evt) {
     source: evt.source,
     type: "",
     query: evt.query,
-    phraseSearch: evt.phraseSearch,
     caseSensitive: false,
     entireWord: false,
     highlightAll: true,

web/firefoxcom.js

@@ -222,7 +222,6 @@ class MozL10n {
       source: window,
       type: type.substring(findLen),
       query: detail.query,
-      phraseSearch: true,
       caseSensitive: !!detail.caseSensitive,
       entireWord: !!detail.entireWord,
       highlightAll: !!detail.highlightAll,

web/pdf_find_bar.js

@@ -99,7 +99,6 @@ class PDFFindBar {
       source: this,
       type,
       query: this.findField.value,
-      phraseSearch: true,
       caseSensitive: this.caseSensitive.checked,
       entireWord: this.entireWord.checked,
       highlightAll: this.highlightAll.checked,

web/pdf_find_controller.js

@@ -387,6 +387,8 @@ function getOriginalIndex(diffs, pos, len) {
  * Provides search functionality to find a given string in a PDF document.
  */
 class PDFFindController {
+  #state = null;
+
   #updateMatchesCountOnProgress = true;
 
   #visitedPagesCount = 0;
@@ -421,7 +423,7 @@ class PDFFindController {
   }
 
   get state() {
-    return this._state;
+    return this.#state;
   }
 
   /**
@@ -445,13 +447,25 @@ class PDFFindController {
     if (!state) {
       return;
     }
+    if (
+      (typeof PDFJSDev === "undefined" || PDFJSDev.test("GENERIC")) &&
+      state.phraseSearch === false
+    ) {
+      console.error(
+        "The `phraseSearch`-parameter was removed, please provide " +
+          "an Array of strings in the `query`-parameter instead."
+      );
+      if (typeof state.query === "string") {
+        state.query = state.query.match(/\S+/g);
+      }
+    }
     const pdfDocument = this._pdfDocument;
     const { type } = state;
 
-    if (this._state === null || this.#shouldDirtyMatch(state)) {
+    if (this.#state === null || this.#shouldDirtyMatch(state)) {
       this._dirtyMatch = true;
     }
-    this._state = state;
+    this.#state = state;
     if (type !== "highlightallchange") {
       this.#updateUIState(FindState.PENDING);
     }
@@ -490,7 +504,7 @@ class PDFFindController {
 
         // When the findbar was previously closed, and `highlightAll` is set,
         // ensure that the matches on all active pages are highlighted again.
-        if (findbarClosed && this._state.highlightAll) {
+        if (findbarClosed && this.#state.highlightAll) {
           this.#updateAllPages();
         }
       } else if (type === "highlightallchange") {
@@ -537,7 +551,7 @@ class PDFFindController {
     this._pageMatches = [];
     this._pageMatchesLength = [];
     this.#visitedPagesCount = 0;
-    this._state = null;
+    this.#state = null;
     // Currently selected match.
     this._selected = {
       pageIdx: -1,
@@ -565,22 +579,44 @@ class PDFFindController {
   }
 
   /**
-   * @type {string} The (current) normalized search query.
+   * @type {string|Array} The (current) normalized search query.
    */
   get #query() {
-    if (this._state.query !== this._rawQuery) {
+    const { query } = this.#state;
-      this._rawQuery = this._state.query;
+    if (typeof query === "string") {
-      [this._normalizedQuery] = normalize(this._state.query);
+      if (query !== this._rawQuery) {
+        this._rawQuery = query;
+        [this._normalizedQuery] = normalize(query);
+      }
+      return this._normalizedQuery;
     }
-    return this._normalizedQuery;
+    // We don't bother caching the normalized search query in the Array-case,
+    // since this code-path is *essentially* unused in the default viewer.
+    return (query || []).filter(q => !!q).map(q => normalize(q)[0]);
   }
 
   #shouldDirtyMatch(state) {
     // When the search query changes, regardless of the actual search command
     // used, always re-calculate matches to avoid errors (fixes bug 1030622).
-    if (state.query !== this._state.query) {
+    const newQuery = state.query,
+      prevQuery = this.#state.query;
+    const newType = typeof newQuery,
+      prevType = typeof prevQuery;
+
+    if (newType !== prevType) {
       return true;
     }
+    if (newType === "string") {
+      if (newQuery !== prevQuery) {
+        return true;
+      }
+    } else {
+      // Array
+      if (JSON.stringify(newQuery) !== JSON.stringify(prevQuery)) {
+        return true;
+      }
+    }
+
     switch (state.type) {
       case "again":
         const pageNumber = this._selected.pageIdx + 1;
@@ -670,7 +706,7 @@ class PDFFindController {
   }
 
   #convertToRegExpString(query, hasDiacritics) {
-    const { matchDiacritics } = this._state;
+    const { matchDiacritics } = this.#state;
     let isUnicode = false;
     query = query.replaceAll(
       SPECIAL_CHARS_REG_EXP,
@@ -741,36 +777,31 @@ class PDFFindController {
 
   #calculateMatch(pageIndex) {
     let query = this.#query;
-    if (!query) {
+    if (query.length === 0) {
-      // Do nothing: the matches should be wiped out already.
+      return; // Do nothing: the matches should be wiped out already.
-      return;
     }
-
+    const { caseSensitive, entireWord } = this.#state;
-    const { caseSensitive, entireWord, phraseSearch } = this._state;
     const pageContent = this._pageContents[pageIndex];
     const hasDiacritics = this._hasDiacritics[pageIndex];
 
     let isUnicode = false;
-    if (phraseSearch) {
+    if (typeof query === "string") {
       [isUnicode, query] = this.#convertToRegExpString(query, hasDiacritics);
     } else {
       // Words are sorted in reverse order to be sure that "foobar" is matched
       // before "foo" in case the query is "foobar foo".
-      const match = query.match(/\S+/g);
+      query = query
-      if (match) {
+        .sort()
-        query = match
+        .reverse()
-          .sort()
+        .map(q => {
-          .reverse()
+          const [isUnicodePart, queryPart] = this.#convertToRegExpString(
-          .map(q => {
+            q,
-            const [isUnicodePart, queryPart] = this.#convertToRegExpString(
+            hasDiacritics
-              q,
+          );
-              hasDiacritics
+          isUnicode ||= isUnicodePart;
-            );
+          return `(${queryPart})`;
-            isUnicode ||= isUnicodePart;
+        })
-            return `(${queryPart})`;
+        .join("|");
-          })
-          .join("|");
-      }
     }
 
     const flags = `g${isUnicode ? "u" : ""}${caseSensitive ? "" : "i"}`;
@@ -780,7 +811,7 @@ class PDFFindController {
 
     // When `highlightAll` is set, ensure that the matches on previously
     // rendered (and still active) pages are correctly highlighted.
-    if (this._state.highlightAll) {
+    if (this.#state.highlightAll) {
       this.#updatePage(pageIndex);
     }
     if (this._resumePageIdx === pageIndex) {
@@ -876,7 +907,7 @@ class PDFFindController {
   }
 
   #nextMatch() {
-    const previous = this._state.findPrevious;
+    const previous = this.#state.findPrevious;
     const currentPageIndex = this._linkService.page - 1;
     const numPages = this._linkService.pagesCount;
 
@@ -911,7 +942,8 @@ class PDFFindController {
     }
 
     // If there's no query there's no point in searching.
-    if (!this.#query) {
+    const query = this.#query;
+    if (query.length === 0) {
       this.#updateUIState(FindState.FOUND);
       return;
     }
@@ -948,7 +980,7 @@ class PDFFindController {
   #matchesReady(matches) {
     const offset = this._offset;
     const numMatches = matches.length;
-    const previous = this._state.findPrevious;
+    const previous = this.#state.findPrevious;
 
     if (numMatches) {
       // There were matches for the page, so initialize `matchIdx`.
@@ -1021,7 +1053,7 @@ class PDFFindController {
       }
     }
 
-    this.#updateUIState(state, this._state.findPrevious);
+    this.#updateUIState(state, this.#state.findPrevious);
     if (this._selected.pageIdx !== -1) {
       // Ensure that the match will be scrolled into view.
       this._scrollMatches = true;
@@ -1106,7 +1138,7 @@ class PDFFindController {
       state,
       previous,
       matchesCount: this.#requestMatchesCount(),
-      rawQuery: this._state?.query ?? null,
+      rawQuery: this.#state?.query ?? null,
     });
   }
 }

web/pdf_link_service.js

@@ -350,10 +350,12 @@ class PDFLinkService {
     if (hash.includes("=")) {
       const params = parseQueryString(hash);
       if (params.has("search")) {
+        const query = params.get("search").replaceAll('"', ""),
+          phrase = params.get("phrase") === "true";
+
         this.eventBus.dispatch("findfromurlhash", {
           source: this,
-          query: params.get("search").replaceAll('"', ""),
+          query: phrase ? query : query.match(/\S+/g),
-          phraseSearch: params.get("phrase") === "true",
         });
       }
       // borrowing syntax from "Parameters for Opening PDF Files"