[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: (ITS#7072) caseIgnoreMatch equals regular capital I with dotted capital I



Frank.Meisschaert@UGent.be wrote:
> Full_Name: Frank Meisschaert
> Version: 2.4.26
> OS: linux
> URL: ftp://ftp.openldap.org/incoming/
> Submission from: (NULL) (157.193.44.200)
>
>
> When I try to add the names "İsmail" (original, with dotted capital I) and
> "Ismail" (transcribed, regular capital I) to an entry using ldapmodify:
>
> dn: ...
> changetype: modify
> replace: givenName
> givenName: Ismail
> givenName:: xLBzbWFpbA==
> -
>
> I get the error:
>
> ldap_modify: Type or value exists (20)
> 	additional info: givenName: value #0 provided more than once
>
> The lowercase version of dotted and regular capital I is regular i which
> explains this error. However according to unicode case folding the dotted
> capital I should be mapped to 0069 0307 (i, combining dot above) which would
> make Ismail andİsmail different when compared case-insensitively as I
> would expect.

Apparently our liblunicode implementation doesn't use the CaseFolding data. I 
believe that means our implementation corresponds to the "Simple" case folding 
rules. They're valid, just not doing what you want in this case. So it appears 
that this is an enhancement request.

I don't believe this request warrants a very high priority. Looking at the 
Unicode charts, I-with-dot-above is the only character where the two case 
folding alternatives actually makes a difference.

-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/