Unicode collation table for swedish do not sort correctly

Question

Level 1

0 points

Unicode collation table for swedish do not sort correctly

I have a problem when using locale unicode setting "sv_SE.UTF-8" in "usr/share/locale" (Mac OS X 10.6.2) for alphabetic sorting in swedish by perl (bibliography program biber, a unicode aware replacement for BibTeX). It consequently puts ä (small and capital a with diaeresis above) before å (small and capital a with ring above), which are separate letters in swedish and comes after z and before ö (small and capital o with diaeresis). The order should be the reverse, å (a-ring) before ä (a-diaeresis).

The problem is clearly in Mac OS X since other systems (Linux) do get the correct output. I've been told that the problem seems to be that sv_SE.UTF-8 locale's collation table is a symlink to "la LN.US-ASCII/LCCOLLATE".

I cannot solve this myself. What is the problem and how can it be fixed?

Mac Book Pro, Mac OS X (10.6.2)

Posted on Feb 23, 2010 12:59 AM

Reply

Answer 1

etresoft

Level 8

45,898 points

Feb 23, 2010 6:41 AM in response to tomasj

In the terminal, type:


cd /usr/share/locale/sv_SE.UTF-8
sudo rm LC_COLLATE
sudo ln -s ../svSE.ISO8859-1/LCCOLLATE .

I strongly advise using copy and paste with the above commands.

Reply

Answer 2

tomasj Author

Level 1

0 points

Feb 23, 2010 7:20 AM in response to etresoft

Thanks, but it didn't work.

Reply

Answer 3

etresoft

Level 8

45,898 points

Feb 23, 2010 8:13 AM in response to tomasj

It does work, it is just that the other file has the same sort order. If you could point me to any official documentation about Swedish sort order then you (or I) could file a bug report.

Reply

Answer 4

tomasj Author

Level 1

0 points

Feb 23, 2010 12:54 PM in response to etresoft

But the two files are not the same as far as I can see, but I don't know how the LC_COLLATE files work. I find it strange that all the utf-8 LC_COLLATE files for all languages in /usr/share/locale are symlinks to "/usr/share/locale/la LN.US-ASCII/LCCOLLATE".

There is some documentation concerning the swedish sort order here: http://std.dkuug.dk/cultreg/registrations/narrative/svSE,1.0

Reply

Answer 5

etresoft

Level 8

45,898 points

Feb 23, 2010 1:16 PM in response to tomasj

tomasj wrote:
But the two files are not the same as far as I can see, but I don't know how the LC_COLLATE files work. I find it strange that all the utf-8 LC_COLLATE files for all languages in /usr/share/locale are symlinks to "/usr/share/locale/la LN.US-ASCII/LCCOLLATE".

They aren't all links to that file, but most of them are. That is the collation file for "Latin".

Come to think of it, there might be problems using that ISO 8859 file in there. I suggest you undo the change from above with:


cd /usr/share/locale/sv_SE.UTF-8
sudo rm LC_COLLATE
sudo ln -s ../laLN.US-ASCII/LCCOLLATE .

There is some documentation concerning the swedish sort order here: http://std.dkuug.dk/cultreg/registrations/narrative/svSE,1.0

I suggest you file a bug report with Apple.

Reply