#24 Extended character support

[Anonymous]

Currently, when scanning or repairing mp3s it can't read files characters with various international symbols such as umlauts, tildes, accent marks, etc. thus requiring the user to remove those files from the scan/repair, abort, and start over. I'm doing this on Windows 7.

[Mel]

I also encountered this. Win 8.1. Front end 0.1.1. It hangs indefinitely. Not all characters, some french characters seem ok, but it hung on some Cyrillic characters, Dvorak spelled with accents and Zoe with an umlaut on the e. Aha! A clue! I noticed that sometimes Noël is OK and sometimes hangs. The e with umlaut as a single character works but sometimes it is actually multiple characters... ie when I back space to delete, the e disappears then the umlaut. See Mary Blige "A Mary Christmas" album.

[zWore]

Yes, there is a problem with the following signs:

█
Æ
ã
ā
À
å
è
ê
ī
ñ
û
ŵ

At last on Windows 7 x64 with the version mp3val-0.1.8_with_frontend-0.1.1-bin-win32

But it does not have problems with some other diacritic signs like for example:

ä
é
ź
ö

[Mohammad Sanati]

I faced to such issue. the file name looks like "10. Don't Tread Оn Me.mp3" in ascii format.
In windows explorer the filename shows as "10. Don't Tread Оn Me.mp3"
The scan stops and says "Cannot open input file "U:\OneDrive\My Music\Metallica\Greatest Hits\Disc 1\10. Don't Tread ?n Me.mp3" or it is empty"

I have to stop the scan and remove the file ans restart scan.
I am using Windows 10

I'm also wondering why the scanning process stops! In case of any fatal problem, it should continue for the rest of the list and show status as cannot read file.

[gabeweb]

Hi, I'm also experiencing this issue. It's practically with any file where the file name (not the content of the tags or other data) has non-Latin characters.

What I ended up doing was simply replacing the file name with Latin characters, and then changing it back to the original name.

This way, I have been able to verify and fix MP3 files (it doesn't alter the content of the tags or cause other errors).

[nichtich]

I'm having the same problem under Windows 11.

[nichtich]

Ok, I looked into it. All characters of the Latin-1 (ISO/IEC 8859-1) charset seem to work fine, including ä, é, ß etc. The following filenames couldn't be processed:
Smoke Hour ★ Willie Nelson - Beyoncé.mp3
Sweet ★ Honey ★ Buckiin' - Beyoncé.mp3
Jesus ♥'s Me - CocoRosie.mp3
Johnny And Mary - √c.mp3
Up And Away - Σtella.mp3
Bal - Derya Yıldırım & Grup Şimşek.mp3

I noticed that in the frontend's file list the non-Latin chacters are being substituted or omitted:
★ and ♥ are substituted by "?"
Σ is substituted by "S"
ı, Ş and ş are substituted by i, S and s
√ is simply dropped

A folder named Σtella couldn't be processed at all - instead of the MP3s inside a single entry "l" is listed with the error message "Cannot open input file "l" or it is empty".

I then tried to run mp3val from the command line but encountered the same problems. So I guess that there are two issues:

• mp3val itself can't handle non-Latin-1 characters in filenames.
• The frontend seems to be substituting non-Latin-1 characters with replacement characters by some automatic conversion and so mp3val cant’t find these files.

Can these issues be fixed? This is such a great piece of software and it shouldn't be too hard to implement proper handling of Unicode in both mp3val and the frontend.