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

Re: Bad matching with wide-byte characters (ITS#1771)



It appears that the string library you are using is not
8-bit clean.  If an update is not available, I suggest
you use an alternative string library.

Kurt

At 09:21 PM 2002-04-23, hirofumi@momose.com wrote:
>Full_Name: Hirofumi Momose
>Version: 2.0.23
>OS: Solaris 2.6
>URL: 
>Submission from: (NULL) (163.135.10.34)
>
>
>Hello
>
>I found a problem of search operation with 2-byte characters.
>
>a) The directory has following entry.
>
> dn: cn=hoashi,o=foo.com
> cn: hoashi
> sn:: 44G744GC44GX
> objectClass: top
> objectClass: person
>
>b) Search by ldap client with filter byte sequence of
> 0x73,0x6e,0x3d,0xe3,0x81,0xbb,0xe3,0x81,0x92,0x2a
> which describe "sn=WIDE_BYTE_CHARACTERS*".
> Above sn does not match of entry.
>
>c) Server respond the entry matched.
>
> I found the problem caused by strncasecmp() in
> caseIgnoreSubstringsFilter(). 8-bit width data flows into
> strncasecmp() of this place. This result generates a problem.
>
>Best regards.