Re: (ITS#3869) Unicode datatypes @ 64-bit host

[hyc@symas.com]

hyc@symas.com wrote:
> Sounds like we should change ldap_unicode_t ac_uint4. It doesn't appear 
> that this definition affects anything besides ucstr.c.
>   

Of course it's not as simple as that; ucstr.c also calls 
ldap_x_utf8_to_ucs4 which uses ldap_ucs4_t. One or the other of these 
types needs to go away. I think I'll replace ldap_unicode_t with 
ldap_ucs4_t.
> h.b.furuseth@usit.uio.no wrote:
>   
>> liblunicode/ucstr.c:ucstrncasecmp() and ucstrncmp() take ldap_unicode_t*
>> strings, but are passed ac_uint4* strings from UTF8bvnormcmp().
>>
>> include/ldap_pvt_uc.h says
>>    /* UCDATA uses UCS-2 passed in an unsigned long */
>>    typedef unsigned long ldap_unicode_t;
>> which is 8 bytes on a 64-bit host, but include/ac/bytes.h says
>>    typedef unsigned AC_INT4_TYPE ac_uint4;
>> where AC_INT4_TYPE is a 4-byte integer type.
>>
>> Yet it passes 'make tests', which maybe means some character
>> encoding tests are needed...
>>
>> BTW, ldap_pvt_uc.h also has a declaration
>>    typedef short ldap_ucs2_t;
>> Looks strange to me that this one is signed.  The type is not
>> used though, so maybe it should simply be deleted.
>>
>>
>>
>>
>>   
>>     
>
>
>   


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