Re: BUG REPORT: failed with Chinese charaters (I've fixed it)

[Pawel Kot]

Hi,

On 10/12/06, waterser forrest <address@hidden> wrote:
>  Message received: 0x14 / 0x006e
> 01 55 00 03 00 01 01 02 00 43 00 00 00 01 00 60 |  U       C     `
> 44 00 08 60 01 80 02 62 85 23 00 00 00 03 82 10 | D  `   b #
>  01 09 0d 91 68 31 67 31 90 79 f6 00 00 00 82 10 |     h1g1 y
>  02 09 08 91 68 31 08 20 01 05 f0 00 00 00 80 30 |     h1         0
> 2a 15 05 00 03 0f 02 02 6c e8 61 0f 90 fd 67 65 | *       l a   ge
> 4e 0d 60 25 4e 86 30 02 53 17 4e ac 73 b0 57 28 | N `%N 0 S N s W(
>  7a 7f 4e c0 4e 48 88 63 67 0d ff 1f 00 00       | z N NH cg
>  Received message type 14
> Trying to get message # 67 in folder # 2
>  Type: Deliver
> Mobile Terminated message:
>      Date: 2006-10-08-20-26-58-+0800
>     Remote number (recipient or sender): +8613761309976
>     SMS center number: +8613800210500
>  UDH found
> Concatenated messages
>  Unicode message
> 注意都来不急了。北京
>  !!^^^^^^^^^^^^^^^This message is truncated.
> !!Let's have a look at the 5th line:
>  !!2a 15 05 00 03 0f 02 02 6c e8 61 0f 90 fd 67 65
> !!2a=2*15 (HEX)
> !!Obviously `2a' is the correst length, whereas `15' is just a half.
>  !!`2a' is block[2] and `15' is block[3]
> !!So I'm sure block[2] indicates the correst length.

It turns out that:
block[2] is the number of bytes used to encode the message
block[3] is number of characters stored in the message
It seems to be a bit different in unicode messages. But still it
doesn't explain your case. I suspect it may have something to do with
message being concatenated -- can you find such dependence? How was
this message sent? How many characters does it have? Could you please
print our result (ie size of each character) of char_wctomb()?

The summary is that I need to back out the patch since it breaks
completly non unicode messages. But I'm working on it.

take care,
pkot
--
Pawel Kot