> I am writing an app that needs to search through the Palm phone book.
> It uses a binary search and uses TxtCaselessCompare for the comparison
> operator.
>
> I am getting incorrect results from TxtCaselessCompare when using a
> mixture or ASCII string and MBCS strings.
>
> My phonebook order is
>
> aaa
> bbb
> ccc
> <Hirigana a>
> <Hirigana i>
> <Hirigana u>
> <Hirigana e>
> <Hirigana o>
>
>
> When I call
> TxtCaselessCompare ("b", <Hirigana i>,...);
>
> it returns 1 instead of -1.
Here's the declaration for TxtCaselessCompare:
Int16 TxtCaselessCompare(const Char* s1, UInt16 s1Len, UInt16 * s1MatchLen,
const Char* s2, UInt16 s2Len, UInt16 * s2MatchLen)
Based on your code snippet, it looks like you might be passing the pointer to the
Hiragana i as the s1MatchLen parameter, versus the s2 parameter. For the code snippet
to really match up with the real API, it would have to be:
Int16 result = TxtCaselessCompare("b", 1, NULL, "Ǣ", 2, NULL);
I also checked the sort order in my test app, and it sorts single-byte before
double-byte by default.
-- Ken
Ken Krugler
TransPac Software, Inc.
<http://www.transpac.com>
+1 530-470-9200 (direct) +1 408-261-7550 (main)
