Beanz/pdf.js
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix the regex string used to find the chars to normalize with NFKC when searching

Browse Source 
ICU has been updated in Firefox (see https://bugzilla.mozilla.org/show_bug.cgi?id=2000225).
The char `0xA7F1` is now alphabetic and can be normalized as a "S".
This commit is contained in:
Calixte Denizet 2025-11-27 14:13:57 +01:00
parent ec71e4ed65
commit eee20cf138
No known key found for this signature in database
GPG Key ID: 0C5442631EE0691F
1 changed files with 27 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
web/pdf_find_utils.js
View File

@ -13,6 +13,8 @@
* limitations under the License. * limitations under the License.
*/ */
import { FeatureTest } from "pdfjs-lib";
const CharacterType = { const CharacterType = {
SPACE: 0, SPACE: 0,
ALPHA_LETTER: 1, ALPHA_LETTER: 1,
@ -114,8 +116,13 @@ function getCharacterType(charCode) {
let NormalizeWithNFKC; let NormalizeWithNFKC;
function getNormalizeWithNFKC() { function getNormalizeWithNFKC() {
if (
(typeof PDFJSDev === "undefined" && FeatureTest.platform.isFirefox) ||
(typeof PDFJSDev !== "undefined" && PDFJSDev.test("MOZCENTRAL"))
) {
/* eslint-disable no-irregular-whitespace */ /* eslint-disable no-irregular-whitespace */
NormalizeWithNFKC ||= ` ¨ª¯²-µ¸-º¼-¾IJ-ijĿ-ŀʼnſDŽ-njDZ-dzʰ-ʸ˘-˝ˠ-ˤʹͺ;΄-΅·ϐ-ϖϰ-ϲϴ-ϵϹևٵ-ٸक़-य़ড়-ঢ়য়ਲ਼ਸ਼ਖ਼-ਜ਼ਫ਼ଡ଼-ଢ଼ำຳໜ-ໝ༌གྷཌྷདྷབྷཛྷཀྵჼᴬ-ᴮᴰ-ᴺᴼ-ᵍᵏ-ᵪᵸᶛ-ᶿẚ-ẛάέήίόύώΆ᾽-῁ΈΉ῍-῏ΐΊ῝-῟ΰΎ῭-`ΌΏ´-῾ - ‑‗․-… ″-‴‶-‷‼‾⁇-⁉⁗ ⁰-ⁱ⁴-₎ₐ-ₜ₨℀-℃℅-ℇ℉-ℓℕ-№ℙ-ℝ℠-™ℤΩℨK-ℭℯ-ℱℳ-ℹ℻-⅀ⅅ-ⅉ⅐-ⅿ↉∬-∭∯-∰〈-〉①-⓪⨌⩴-⩶⫝̸ⱼ-ⱽⵯ⺟⻳⼀-⿕　〶〸-〺゛-゜ゟヿㄱ-ㆎ㆒-㆟㈀-㈞㈠-㉇㉐-㉾㊀-㏿ꚜ-ꚝꝰꟲ-ꟴꟸ-ꟹꭜ-ꭟꭩ豈-嗀塚晴凞-羽蘒諸逸-都飯-舘並-龎ff-stﬓ-ﬗיִײַ-זּטּ-לּמּנּ-סּףּ-פּצּ-ﮱﯓ-ﴽﵐ-ﶏﶒ-ﷇﷰ-﷼︐-︙︰-﹄﹇-﹒﹔-﹦﹨-﹫ﹰ-ﹲﹴﹶ-ﻼ！-ﾾￂ-ￇￊ-ￏￒ-ￗￚ-ￜ￠-￦`; NormalizeWithNFKC ||= `\xA0¨ª¯²-µ¸-º¼-¾IJ-ijĿ-ŀʼnſDŽ-njDZ-dzʰ-ʸ˘-˝ˠ-ˤʹͺ;΄-΅·ϐ-ϖϰ-ϲϴ-ϵϹևٵ-ٸक़-य़ড়-ঢ়য়ਲ਼ਸ਼ਖ਼-ਜ਼ਫ਼ଡ଼-ଢ଼ำຳໜ-ໝ༌གྷཌྷདྷབྷཛྷཀྵჼᴬ-ᴮᴰ-ᴺᴼ-ᵍᵏ-ᵪᵸᶛ-ᶿẚ-ẛάέήίόύώΆ᾽-῁ΈΉ῍-῏ΐΊ῝-῟ΰΎ῭-`ΌΏ´-῾ - ‑‗․-… ″-‴‶-‷‼‾⁇-⁉⁗ ⁰-ⁱ⁴-₎ₐ-ₜ₨℀-℃℅-ℇ℉-ℓℕ-№ℙ-ℝ℠-™ℤΩℨK-ℭℯ-ℱℳ-ℹ℻-⅀ⅅ-ⅉ⅐-ⅿ↉∬-∭∯-∰〈-〉①-⓪⨌⩴-⩶⫝̸ⱼ-ⱽⵯ⺟⻳⼀-⿕　〶〸-〺゛-゜ゟヿㄱ-ㆎ㆒-㆟㈀-㈞㈠-㉇㉐-㉾㊀-㏿ꚜ-ꚝꝰ꟱-ꟴꟸ-ꟹꭜ-ꭟꭩ豈-嗀塚晴凞-羽蘒諸逸-都飯-舘並-龎ff-stﬓ-ﬗיִײַ-זּטּ-לּמּנּ-סּףּ-פּצּ-ﮱﯓ-ﴽﵐ-ﶏﶒ-ﷇﷰ-﷼︐-︙︰-﹄﹇-﹒﹔-﹦﹨-﹫ﹰ-ﹲﹴﹶ-ﻼ！-ﾾￂ-ￇￊ-ￏￒ-ￗￚ-ￜ￠-￦`;
}
if (typeof PDFJSDev === "undefined" || PDFJSDev.test("TESTING")) { if (typeof PDFJSDev === "undefined" || PDFJSDev.test("TESTING")) {
const ranges = []; const ranges = [];
@ -145,7 +152,24 @@ function getNormalizeWithNFKC() {
} }
} }
} }
if (ranges.join("") !== NormalizeWithNFKC) {
const rangesStr = ranges.join("");
if (!NormalizeWithNFKC) {
NormalizeWithNFKC = rangesStr;
} else if (rangesStr !== NormalizeWithNFKC) {
for (let i = 1; i < rangesStr.length; i++) {
if (rangesStr[i] !== NormalizeWithNFKC[i]) {
console.log(
`Difference at index ${i}: ` +
`U+${rangesStr.charCodeAt(i).toString(16).toUpperCase().padStart(4, "0")}` +
`!== U+${NormalizeWithNFKC.charCodeAt(i)
.toString(16)
.toUpperCase()
.padStart(4, "0")}`
);
break;
}
}
throw new Error( throw new Error(
"getNormalizeWithNFKC - update the `NormalizeWithNFKC` string." "getNormalizeWithNFKC - update the `NormalizeWithNFKC` string."
); );