GNOME Bugzilla – Bug 586703
Japanese kana ignored (treated as punctuation) in sorts
Last modified: 2009-06-25 22:20:50 UTC
Originally filed here: https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/389739 File name sorting has a nice feature where it ignores characters like [ and * at the beginning of file names, and sorts according to the first sortable character. i.e. [a *a ^a [b ^b *c etc...This is nice, however: Japanese kana characters are incorrectly treated the same way, which causes them to be sorted in a completely arbitrary way. This does not happen when the locale is Japanese, but does for all other locales that I've tried. Kanji seem not to be affected, only kana. Note in the attached picture (there was one in the original post on the Ubuntu tracker, you can look there but it doesn't really add anything that the description doesn't) the way that the katakana "A" characters are treated exactly the same way as the brackets, asterisk, etc. Also, the two folder names at the top are sorted incorrectly in relation to each other; which gets much worse if you have a lot of real file names (the sorting is completely arbitrary and useless). The way files are sorted in the Japanese localisation is correct, though.
Likely a problem somewhere in the depths of the UTF-8 collation key code. Reassigning to GLib for their opinion on this.
GLib just uses libc's collation routines internally.
Thanks Owen.
Any idea who this should be reported to then? libc's page says to report bugs to the distros if applicable, and it's locale-specific anyway.
Sounds like you should revisit the Ubuntu bug, then, and get it reassigned to the appropriate package.