Issue 45171 - Not using ICU sorting in the alphabetical index (OOo uses worng language setting?)
Summary: Not using ICU sorting in the alphabetical index (OOo uses worng language sett...
Status: CLOSED FIXED
Alias: None
Product: Writer
Classification: Application
Component: code (show other issues)
Version: 680m85
Hardware: PC Linux, all
: P2 Trivial (vote)
Target Milestone: ---
Assignee: eric.savary
QA Contact: issues@sw
URL:
Keywords: oooqa
Depends on:
Blocks:
 
Reported: 2005-03-16 11:22 UTC by ooo
Modified: 2013-08-07 14:42 UTC (History)
2 users (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
Example document showing the wrong sorting order in the alphabetical index (8.07 KB, application/vnd.oasis.opendocument.text)
2005-03-16 11:24 UTC, ooo
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description ooo 2005-03-16 11:22:48 UTC
When I create an alphabetical index it is not sort correctly in Danish: 
 
OOo 680m85    Correct sorting 
----------------------------- 
dag           dag 
måske         måske 
ændre         ändra 
øste          ændre 
året          øste 
ändra         öste 
öste          året 
----------------------------- 
 
This issue is based on issue 
http://qa.openoffice.org/issues/show_bug.cgi?id=25323
Comment 1 ooo 2005-03-16 11:24:06 UTC
Created attachment 23894 [details]
Example document showing the wrong sorting order in the alphabetical index
Comment 2 lohmaier 2005-03-16 12:45:24 UTC
confirming issue.

OOo does not respect the language settings.

Just compare the result of "Tools|Sort" (with language set to danish) and then
the sorting in the index (with language set to danish) -> different results.

The language set in the properties of the index does make a difference (compare
the result with language set to german vs the result with language set to french
vs the result with language set to danish) but it doesn't seem to use the same
sorting as e.g. Tools|Sort

this was already reported as issue 22921 which was marked as duplicate of issue
25323. As it turns out, it is not a duplicate. Please have a look.

Justification of P2: An alphabetical index is useless without the correct
alphabetical sorting -> leaving as P2 (basic functionality not working
correctly/functional area lost)
Comment 3 michael.ruess 2005-03-16 13:39:05 UTC
Reassigned to ES.
Comment 4 eric.savary 2005-03-16 13:58:43 UTC
ES->KHONG: please have a look.
Comment 5 karl.hong 2005-03-16 20:28:42 UTC
The problem is in IndexKey defined in Danish locale data,

http://l10n.openoffice.org/source/browse/l10n/i18npool/source/localedata/data/da_DK.xml

I remember I took the list from IBM web page, 

http://oss.software.ibm.com/icu/charts/collation/

but they have re-arranged their web site, the link is not valid now. I may made
mistake.

According icu collation data for Danish, the list in OOo locale data is not
right, it should be

<æ<<<Æ<<ä<<<Ä<<Ä™<<<Ę<ø<<<Ø<<ö<<<Ö<<Å‘<<<Å<<Å“<<<Å’<Ã¥<<<Ã…<<<

I will update it.
Comment 6 lohmaier 2005-03-16 20:51:14 UTC
Please note that Tools|Sort already uses the correct sorting. So if you write
the index uses the algorithm for danish, this means that OOo applies two
different sorting-algorithms.

So please do not "fix" the data for danish, but use the same sorting-mechanism
for the index as is used for the rest of OOo (tables, Tools|Sort).

The new location can be found here:
http://icu.sourceforge.net/userguide/Collate_Intro.html (use the left navbar to
jump to the specific pages)

The online-demo-sorting can be found here:
http://www-950.ibm.com/software/globalization/icu/demo/locales
Comment 7 ooo 2005-05-24 17:31:08 UTC
Cloph,

Please note that index title sorting is not necessarily identical to general
sorting (collation), it depends on the locale. That's why we introduced the
element in locale data.

Eike
Comment 8 karl.hong 2005-08-26 23:08:09 UTC
The index key defined in da_DK.xml is correct. I was wrong in my previous comment.

From the icu tailoring for Danish,

<æ<<<Æ<<ä<<<Ä<<Ä™<<<Ę<ø<<<Ø<<ö<<<Ö<<Å‘<<<Å<<Å“<<<Å’<Ã¥<<<Ã…<<<

only 3 primary different letter groups, and they are correctly defined in
da_DK.xml as Æ Ø Å  in right order.

The problem of this issue is index grouping is not using primary collation, but
secondary, which causes Ä is not grouped under Æ. The root cause is we don't
have primary collation in OO.o, we only have ignore case, which is mapped to
secondary. 

I will add new  collation option called ignore accent, which will map to primary
collation in icu to solve the problem.
Comment 9 karl.hong 2005-08-31 21:29:06 UTC
ready for QA.

re-open issue and reassign to oc@openoffice.org
Comment 10 karl.hong 2005-08-31 21:29:14 UTC
reassign to oc@openoffice.org
Comment 11 karl.hong 2005-08-31 21:29:19 UTC
reset resolution to FIXED
Comment 12 karl.hong 2005-09-13 02:27:47 UTC
Ready for QA.

re-open issue and reassign to oc@openoffice.org
Comment 13 karl.hong 2005-09-13 02:28:20 UTC
reassign to oc@openoffice.org
Comment 14 karl.hong 2005-09-13 02:28:47 UTC
reset resolution to FIXED
Comment 15 oc 2005-09-26 16:00:48 UTC
Hi Eric, please take over

re-open issue and reassign to es@openoffice.org
Comment 16 oc 2005-09-26 16:01:40 UTC
reassign to es@openoffice.org
Comment 17 oc 2005-09-26 16:01:59 UTC
reset resolution to FIXED
Comment 18 eric.savary 2005-10-14 12:49:42 UTC
Verified in CWS i18n20
Comment 19 eric.savary 2005-10-20 13:24:47 UTC
Ok in src680m135