Describe the bug Special characters in filenames are not recogniz

It's covered by tests though <div class="B

🎉 This issue has been resolved in v7.5.5 (<a href="h

[BUG] Special characters are not recognized anymore about junrar HOT 8 CLOSED

junrar commented on May 23, 2024

[BUG] Special characters are not recognized anymore

from junrar.

Comments (8)

gotson commented on May 23, 2024

It's covered by tests though

junrar/src/test/java/com/github/junrar/ArchiveTest.java

Line 388 in a26a8d2

public void unicodeFileNamesAreDecodedProperly() throws Exception {

from junrar.

AlexisJehan commented on May 23, 2024

It works with the unicode.rar file, effectively:

ウニコド.txt: 67
新建文本文档.txt: 10

Maybe the file I created is encoded differently? I made it with Winrar on Windows checking the RAR4 format.

from junrar.

gotson commented on May 23, 2024

I cannot reproduce on my macbook, it works fine. Could it be something with how your Java is configured?

from junrar.

gotson commented on May 23, 2024

Would you be able to produce a minimum reproducible repo on Github, showing that it works with a previous version and fails with 7.5.4?

from junrar.

AlexisJehan commented on May 23, 2024

I'm using Windows 10 64 bits and Adoptium JDK 11.0.16.1+1 (I also tried Adoptium JDK 18.0.2.1+1).

Here is the reproductible repo, with my outputs in the README file: https://github.com/AlexisJehan/JunrarIssue108
(The issue appears since Junrar 7.5.3)

from junrar.

gotson commented on May 23, 2024

Thanks, I have a fair idea where the bug could be, but I want to perform a repro first.

from junrar.

gotson commented on May 23, 2024

I managed to perform a repo on Gihub CI, it only fails on Windows with your file.

Archives have both a fileName and fileNameW, where the latter is the unicode one.

On Mac in CI, we get this:

07:00:29.751 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: ?.txt
07:00:29.752 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: テ.txt
07:00:29.752 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:29.752 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: テ.txt

And on Windows:

07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: ?.txt
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: ?.txt
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: ?.txt

For the existing unicode archive we had, we get something different:

On Mac:

07:00:29.770 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: ウニコド.txt
07:00:29.770 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: ウニコド.txt
07:00:29.770 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:29.770 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: ウニコド.txt
07:00:29.771 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: 新建文本文档.txt
07:00:29.775 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: 新建文本文档.txt
07:00:29.775 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:29.775 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: 新建文本文档.txt

On Windows:

07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: ã‚¦ãƒ‹ã‚³ãƒ‰.txt
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: ????.txt
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:47.789 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: ????.txt
07:00:47.805 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filename: æ–°å»ºæ–‡æœ¬æ–‡æ¡£.txt
07:00:47.805 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive filenameW: ??????.txt
07:00:47.805 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive isUnicode: true
07:00:47.805 [Test worker] WARN com.github.junrar.rarfile.FileHeader - Archive getFileName: ??????.txt

Additional validity checks were added in 7.5.3:

            if (!isFilenameValid(fileName)) {
                throw new CorruptHeaderException("Invalid filename: " + fileName);
            }
            if (!isFilenameValid(fileNameW)) {
                throw new CorruptHeaderException("Invalid filename: " + fileNameW);
            }

    private static boolean isFilenameValid(String filename) {
        try {
            String ignored = new File(filename).getCanonicalPath();
            return true;
        } catch (IOException e) {
            return false;
        }
    }

There's a few things i don't yet understand:

why isFilenameValid fails with ?.txt but not ????.txt
how the existing unicode archive manages to pass the test assertThat(names).containsExactlyInAnyOrder("新建文本文档.txt", "ウニコド.txt");

It might be an encoding issue in the console output, while internally in Java it works.

from junrar.

gotson commented on May 23, 2024

🎉 This issue has been resolved in v7.5.5 (Release Notes)

from junrar.

[BUG] Special characters are not recognized anymore about junrar HOT 8 CLOSED

Comments (8)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent