Merge pull request #20518 from calixteman/issue20516

Escape punctuation signs in search query

Author: timvandermeij

test/pdfs/.gitignore

@@ -765,3 +765,4 @@
 !two_paragraphs.pdf
 !paragraph_and_link.pdf
 !issue20225.pdf
+!issue20516.pdf

test/pdfs/issue20516.pdf

@@ -1158,6 +1158,26 @@ describe("pdf_find_controller", function () {
     });
   });
 
+  it("performs a search with a group of punctuation signs to escape", async () => {
+    const { eventBus, pdfFindController } =
+      await initPdfFindController("issue20516.pdf");
+
+    await testSearch({
+      eventBus,
+      pdfFindController,
+      state: {
+        query: `("client")`,
+      },
+      matchesPerPage: [1],
+      selectedMatch: {
+        pageIndex: 0,
+        matchIndex: 0,
+      },
+      pageMatches: [[6]],
+      pageMatchesLength: [[10]],
+    });
+  });
+
   describe("custom matcher", () => {
     it("calls to the matcher with the right arguments", async () => {
       const QUERY = "Foo  bar";

test/unit/pdf_find_controller_spec.js

@@ -77,8 +77,8 @@ const DIACRITICS_EXCEPTION = new Set([
 let DIACRITICS_EXCEPTION_STR; // Lazily initialized, see below.
 
 const DIACRITICS_REG_EXP = /\p{M}+/gu;
-const SPECIAL_CHARS_REG_EXP =
-  /([*+^${}()|[\]\\])|(\p{P}+)|(\s+)|(\p{M})|(\p{L})/gu;
+const SPECIAL_PUNCTUATION_CHARACTERS = /[.?*{}()[\]\\]/g;
+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;
 
@@ -739,7 +739,10 @@ class PDFFindController {
         }
         if (p2) {
           // Allow whitespaces around group of punctuation signs.
-          return addExtraWhitespaces(p2, p2.replaceAll(/[.?]/g, "\\$&"));
+          return addExtraWhitespaces(
+            p2,
+            p2.replaceAll(SPECIAL_PUNCTUATION_CHARACTERS, "\\$&")
+          );
         }
         if (p3) {
           // Replace spaces by \s+ to be sure to match any spaces.