Bug: names with accented characters are sorted incorrectly

asked 2014-01-04 00:06:45 +0300

Venemo gravatar image

updated 2014-01-27 20:06:23 +0300

chemist gravatar image

In the people app, the names of contacts that have accented charactes in their names are sorted incorrectly.

For example, á should come after a and not after z. Currently all accented characters are sorted after z.

NOTE: since the only one of the offered languages I speak is English, I'm using the device in the English language. It seems that the choice of language also affects every other locale setting as well. (Such as collation, etc.)

edit retag flag offensive close delete

Comments

I have not checked if locale is taken in consideration in the People application, but the alphabetic sorting should be locale dependent. Maybe it is performed in the "C" locale now (meaning, ASCII/UNICODE codepoint order), or maybe you are using the wrong locale? If you are using English, it is possible the collation order is not what you would expect.

mikelima ( 2014-01-07 19:33:33 +0300 )edit

@mikelima I didn't find an option to set locale in settings. Where is it?

Venemo ( 2014-01-07 20:13:30 +0300 )edit

There is a "Language" setting page, that let you chose the language, and thus the locale... It should be more fine grained than that, e.g, let people use English messages but metric system... and collation system. Maybe you can change the LC_COLLATE variable in some system script... I think it was described on the JollaTides site.

mikelima ( 2014-01-07 21:53:29 +0300 )edit

@mikelima - The only language I understand from the list of available languages is English, so that's my setting now. Although I'd prefer to see everything else in Central European format.

Venemo ( 2014-01-08 00:08:02 +0300 )edit

@Venemo - I just tried adding a contact named "A-grave_accent" (sorry, us keyboard here), with language set to Italian, and it is sorted in the "A" bucket. It ends up at the end of the "A" contact list. So, collating seems to work correctly, at least for the Italian language. What I think might happen is that if you select English as the language, you end up with the "C" locale, which means no advanced collation is done, it just sorts based on the numeric character value. I think this is not what users, even English speaking users, expect in this case. Maybe not many will notice the problem, English not using accents normally, but as soon as you have a few international contacts, you would hit the problem as well.

mikelima ( 2014-01-09 11:37:16 +0300 )edit