Wrong mime type for the files with non-ascii characters

atzengi25atn · n00b Joined: 22 Jan 2012 Posts: 14

Hello everyone,

Facing a strange problem in Dolphin recently.
I don't know since when it started.

For some pdf files for instance, when they have some specific non-ascii character in the filename, Dolphin does not detect the mime type correctly.
For example I have made 10 copies of the same pdf file and renamed them as they include some of the characters like "İ", "Ş", etc.
The most problematic one in my experience is the dotted capital I "İ".

When I list them in the folder they all look normal as follows,

Fitzcarraldo · Posted: Fri Feb 03, 2017 12:52 am Post subject:

Strange indeed. I don't experience that in my installation. I wonder if something is messed up in the file associations in your installation? Have you checked 'System Settings' > 'Applications' > 'File Associations'? If you enter 'pascal' in the box 'Find file type or filename pattern' you should see the following under 'Known Types' if you expand the '>-text' entry:

atzengi25atn · n00b Joined: 22 Jan 2012 Posts: 14

Thank you @Fitzcarraldo.

I can see the x-pascal as you mentioned having the extensions *.p and *.pas
The applications listed are Kate, Kwrite, Libreoffice Writer, etc.

The "pdf" mime type, it also appears under the "application" category and has the *.pdf extension only.
The applications associated are Okular, gimp, etc.

Generally I don't have a problem with these file types.
Those which have the problem turn OK when I change their name to an ascii character.

Can you rename a file in your system as shown in my first message please?
Does it react the same?

Fitzcarraldo · Posted: Fri Feb 03, 2017 9:28 am Post subject:

Before making my first post I renamed a PDF file to AIA.pdf and then to AİA.pdf (I with a dot above it). The 'file' command and Dolphin showed both files as being PDF files.

However, as you requested me to rename a file as shown in your first post, I have renamed a PDF file to AIIA.pdf and then to AİİA.pdf (two 'I's with a dot above each). And, yes, I see the same as you! Dolphin indicates the file type is 'Pascal source code', but the 'file' command shows the PDF file type correctly:

atzengi25atn · n00b Joined: 22 Jan 2012 Posts: 14

I have found other file types affected with this. Most probably there are more.

Try these please;

ABC.pdf -> AİİA.pdf
ABC.odt -> AİİA.odt
ABC.docx -> AİA.docx

Renamed pdf file appears as Pascal Source Code, odt file appears as Object Code, and docx file appears as Zip Archive. Really weird.
docx is affected only with one "İ". It seems OK with double "İ"s. Another weirdness.

By the way, I have filed the bug here : https://bugs.kde.org/show_bug.cgi?id=375956

Fitzcarraldo · Posted: Fri Feb 03, 2017 3:37 pm Post subject:

1. ABC.pdf -> AİİA.pdf
2. ABC.odt -> AİİA.odt
3. ABC.docx -> AİA.docx
4. ABC.docx -> AİİA.docx

In my case, Dolphin shows the original files above as, respectively, the following:

1. PDF document.
2. ODT document.
3. Word 2007 document.
4. Word 2007 document.

In my case, Dolphin shows the renamed files above as, respectively, the following:

1. Pascal source code.
2. object code.
3. Zip archive.
4. Word 2007 document.

In other words, exactly the same as you see.
_________________
Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC systemd-utils[udev] elogind KDE on both.
My blog

Fitzcarraldo · Posted: Sat Feb 04, 2017 10:54 am Post subject:

I see that KDE developer Elvis Angelaccio has acted on your bug report and filed three bug reports in the freedesktop.org bugzilla regarding the x11-misc/shared-mime-info package:

https://bugs.freedesktop.org/show_bug.cgi?id=99672 (PDF file with non-ASCII filename detected as Pascal source code)
https://bugs.freedesktop.org/show_bug.cgi?id=99673 (ODT file with non-ASCII filename detected as application/x-object)
https://bugs.freedesktop.org/show_bug.cgi?id=99674 (docx file with non-ASCII filename detected as application/zip)
_________________
Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC systemd-utils[udev] elogind KDE on both.
My blog

krinn · Watchman Joined: 02 May 2003 Posts: 7470

Fitzcarraldo · Posted: Tue Feb 28, 2017 12:33 am Post subject:

Apparently it is a bug in Qt5 and is fixed in Qt 5.8.1:

https://bugreports.qt.io/browse/QTBUG-58822
_________________
Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC systemd-utils[udev] elogind KDE on both.
My blog