Re: mo-recode for SMPP (possibly not working)

Citando David Chkhartishvili <[email protected]>:

> Here  is smsbox log. original message I've sent was: 99182418*t t t t
> and curillic letter on the end. You can see that '@' sign was added
> before all characters.
> 
> 
> 2003-03-28 18:16:02 [4] INFO: Starting to service <@9@9@1@8@2@4@1@8@*@t@
> @t@ @t@ @t@ è<> from <99599182418> to <5244>
> 2003-03-28 18:16:02 [9] DEBUG:   data: 50 4f 53 54 20 2f 72 65   POST /re
> 2003-03-28 18:16:02 [9] DEBUG:   data: 70 6c 79 5f 43 4c 49 52   ply_CLIR
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2e 70 68 70 3f 73 65 6e   .php?sen
> 2003-03-28 18:16:02 [9] DEBUG:   data: 64 65 72 3d 39 39 35 39   der=9959
> 2003-03-28 18:16:02 [9] DEBUG:   data: 39 31 38 32 34 31 38 26   9182418&
> 2003-03-28 18:16:02 [9] DEBUG:   data: 74 65 78 74 3d 25 34 30   text=%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 39 25 34 30 39 25 34 30   9%409%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 31 25 34 30 38 25 34 30   1%408%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 32 25 34 30 34 25 34 30   2%404%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 31 25 34 30 38 25 34 30   1%408%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2a 25 34 30 74 25 34 30   *%40t%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2b 25 34 30 74 25 34 30   +%40t%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2b 25 34 30 74 25 34 30   +%40t%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2b 25 34 30 74 25 34 30   +%40t%40
> 2003-03-28 18:16:02 [9] DEBUG:   data: 2b 25 45 38 25 33 43 20   +%E8%3C
> 2003-03-28 18:16:02 [9] DEBUG:   data: 48 54 54 50 2f 31 2e 31   HTTP/1.1

There's two errors in this.

First, something is decoding ucs2 (double bytes) chars from gsm alphabet, which
is wrong. ucs2 should be treated just like 8 bits and passed through without
conversion.
You should get a %009 (is %00 + '9') .... %00+ (is %00 + encoded space) %E8%3C

Could someone test if this smsc module, in cvs, still have this behaviour ?
Can you, David, use a newer version and see if you get the %00 insted of %40 ? 


Second, mo-recode will only recode to iso IF it can do it. You have a %E8%3C
cyrilic char, which probably won't be possible to convert to iso.

As soon as you have a working version that sends %00, you'll probably have to,
in you php, analise content-type header (charset=utf-16be) and convert it
yourself to utf-8 or try to convert to iso and ignore non-converted chars.


PS: This reminds me that we could mo-recode it to utf-8 instead of iso, with a
parameter ;)

-- 
<br/>
 00:47:56 up 132 days,  1:01,  2 users,  load average: 0.41, 0.79, 0.92
[It is] best to confuse only one issue at a time.
                -- K&R