On Wed, 28 Sep 2005 11:58:41 +0530, Sastry <[EMAIL PROTECTED]> wrote
> Hi Sadahiro
> I was preoccupied with the other stuff. So i could not
> respond you immediately!
> For the previous four tests that were failing, I had collected the
> verbose ouput and it looks like this. I had also made changes to
> NATIVE_TO_UTF from NATIVE_TO_ASCII
> -----------------------------------------------------------------------------------------
>
> SWASHNEW utf8 # comment
> 0100 0058
> 0069 XXXX
> 00c6 XXXX
> 00c7 XXXX
> 00c8 XXXX
> 00c9 XXXX
> 00ca XXXX
> 00cb XXXX
> 00cc XXXX
> 006a XXXX
To me, this output is expectable if it comes from /\x{0100}\x89-\x91/
(beginning with 256 in decimal) with NATIVE_TO_UTF() misused;
but not from \x{1000}\x89-\x91 (beginning with 4096 in decimal)
I think (0x69,0xc6,0xc7,0xc8,0xc9,0xca,0xcb,0xcc,0x6a) should come
from PL_e2utf[] in the range of 0x89 to 0x91.
-- see utfebcdic.h for PL_e2utf[], which is used in NATIVE_TO_UTF(),
#define NATIVE_TO_UTF(ch) PL_e2utf[(U8)(ch)]
> SWASHNEW utf8 # comment
> 1000 0058
> 0049 XXXX
> 00e8 XXXX
> 00e9 XXXX
> 00ea XXXX
> 00eb XXXX
> 00ec XXXX
> 00ed XXXX
> 007d XXXX
> 004a XXXX
To me, this output is expectable if it come from /\x{1000}\xc9-\xd1/
(beginning with 4096 in decimal) with NATIVE_TO_UTF() misused.
I think (0x49,0xe8,0xe9,0xea,0xeb,0xec,0xed,0x7d,0x4a) should come
from PL_e2utf[] in the range of 0xc9 to 0xd1.
Thus I think these outputs reflect misusing of NATIVE_TO_UTF().
Regards,
SADAHIRO Tomoyuki
