mimeopen with UTF-8 characters in name

[Paul-Gilmartin]

Hello, Michiel,

Earlier I submitted this by email to pardus@cpan.org. I now see I should be using GitHub.

Hello, Jaap,

I have in mimeopen:

No known bugs, please mail the author if you find one.

See File::MimeInfo::Applications(3) and File::DesktopEntry(3)
for some limitations.
...
Copyright (c) 2002, 2012 Jaap G Karssenberg. All rights reserved.

And then a strangely UTF-8k named file:

515 $ Movie="L'Odyssée Cartier.flv"
516 $ ls -l "$Movie"
-rwxrwx--- 1 root vboxsf 13225664 Mar 26 2012 L'Odyssée Cartier.flv
517 $
517 $ # So far so good. But:
518 $
518 $ mimeopen -L -n "$Movie"
Opening "L'Odyssée Cartier.flv" with VLC media player (video/x-flv)
VLC media player 2.0.3 Twoflower (revision 2.0.2-93-g77aa89e)
[0x8f26908] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface.
[0x925bf00] filesystem access error: cannot open file /media/sf_Shared/L'Odysse??e Cartier.flv (No such file or directory)
[0x9255930] main input error: open of `file:///media/sf_Shared/L'Odyssée Cartier.flv' failed

I notice the 0x925bf00 message has the filename corrupted, but 0x9255930
renders it correctly.

Have I stumbled into one of the limitations mentioned?

If I simply click on the file's icon in Thunar File Manager, Thunar opens it correctly with VLC.

Thanks,
gil

[Paul-Gilmartin]

(more) The version of mimeopen delivered with Crunchbang Linux is 0.15.
I downloaded and installed 0.27. It shows the same misbehavior

The offending character is 'COMBINING ACUTE ACCENT' lower-case "e".:
589 $ printf é | od -tx1
0000000 65 cc 81
... if I delete it from the pathname, mimeopen handles it OK.

Thanks,
gil

[mbeijen]

Hi Paul,

If I create such a file on my system I can open it correctly.

I notice you open the file from /media/sf_Shared/ - can you please tell me
what kind of file system it uses?

Michiel

On Thu, Jun 25, 2015 at 4:02 AM, Paul-Gilmartin notifications@github.com
wrote:

(more) The version of mimeopen delivered with Crunchbang Linux is 0.15.
I downloaded and installed 0.27. It shows the same misbehavior

The offending character is 'COMBINING ACUTE ACCENT' lower-case "e".:
589 $ printf é | od -tx1
0000000 65 cc 81
... if I delete it from the pathname, mimeopen handles it OK.

Thanks,
gil

—
Reply to this email directly or view it on GitHub
#15 (comment)
.

[Paul-Gilmartin]

Hello, Michiel,
Thanks for your quick reply.

I thought I had eliminated VLC as a suspect with:

619 $ vlc "/media/sf_Shared/L'Odyssée Cartier.flv"

with successful open. But, trying the command generated by mimeopen:

616 $ vlc "file:///media/sf_Shared/L'Odyssée Cartier.flv"
VLC media player 2.0.3 Twoflower (revision 2.0.2-93-g77aa89e)
[0x2105108] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface.
[0x23f82a8] filesystem access error: cannot open file /media/sf_Shared/L'Odysse??e Cartier.flv (No such file or directory)
[0x2136738] main input error: open of `file:///media/sf_Shared/L'Odyssée Cartier.flv' failed

So, in RFC 1738"

URLs are written only with the graphic printable characters of the
US-ASCII coded character set. The octets 80-FF hexadecimal are not
used in US-ASCII, and the octets 00-1F and 7F hexadecimal represent
control characters; these must be encoded.

So I try:

622 $ vlc "file:///media/sf_Shared/L'Odysse%cc%81e Cartier.flv"
VLC media player 2.0.3 Twoflower (revision 2.0.2-93-g77aa89e)
[0x1d8f108] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface.
[flv @ 0x7f4e10026460] Estimating duration from bitrate, this may be inaccurate
[flv @ 0x2c0f1a0] Estimating duration from bitrate, this may be inaccurate
[0x2c396e8] xcb_xv vout display error: no available XVideo adaptor

The message is harsh, but the movie opens OK. VLC isn't adhering
to Postel's principle very well.

Since you asked, I'm running Crunchbang Linux as a guest under
VirtualBox and the filesystem is a "shared folder" managed by
the VirtualBox hypervisor. But that doesn't appear to be a
contributing factor.

623 $ df -kT
Filesystem Type 1K-blocks Used Available Use% Mounted on
rootfs rootfs 19513212 5064736 13457244 28% /
udev devtmpfs 10240 0 10240 0% /dev
tmpfs tmpfs 50880 628 50252 2% /run
/dev/mapper/gil--crunchbang-root ext4 19513212 5064736 13457244 28% /
tmpfs tmpfs 5120 0 5120 0% /run/lock
tmpfs tmpfs 101740 84 101656 1% /run/shm
/dev/sda1 ext2 233191 18379 202371 9% /boot
none vboxsf 78381056 63319508 15061548 81% /media/sf_Shared

Thanks again,
gil

[GSI]

I confirm that opening files with umlauts in their names works with https://github.com/skalkoto/File-MimeInfo/commit/60d390cf1b1dfbcf039e850093e49d5f6e590d21. Please consider merging.

Without the patch (mimeopen 0.29):

Opening "bücher.pdf" with MuPDF  (application/pdf)
error: cannot open bücher.pdf: No such file or directory
mupdf: error: cannot open document

With the patch:

Opening "bcher.pdf" with MuPDF  (application/pdf)

Notice that the ü is missing in the command line output, but the file still opens as intended.