It seems that my modem (HUAWEI 1762) is only able to use the +CUSD command in PDU mode (eg. to check balance: AT+CUSD=1,"AAD80C3602",15 instead of AT+CUSD=1,"*130#",15..

also the answer is in PDU format.

I wrote a quick hack to decode the PDU with a simple C-Routine I found somewhere (if anyone interested)..

However, any chance of a "standard" PDU encoding and decoding for +CUSD messages within the original smsools anytime soon?

regards

STefan

PS: original and changed loglines:
2010-03-22 16:07:16,5, GSM1: CMD: AT+CUSD=1,"AAD80C3602",15: OK +CUSD: 0,"49B41C242DCFE9E73A1D1D1697DDA0F49C0E1A228DA0188E85BB01",15

2010-03-22 16:07:16,5, GSM1: CMD: AT+CUSD=1,"AAD80C3602",15: OK +CUSD: 0,"49B41C242DCFE9E73A1D1D1697DDA0F49C0E1A228DA0188E85BB01",15 // Ihr Restguthaben ist CHF 18,87

clumsy wrote

I wrote a quick hack to decode the PDU with a simple C-Routine I found somewhere (if anyone interested)..

It might be interesting. I have not received CUSD in packed format, but probably someone else has. You could publish your hack here.

clumsy wrote

However, any chance of a "standard" PDU encoding and decoding for +CUSD messages within the original smstools anytime soon?

There has been no plans for PDU encoding and decoding (= USSD Packing).

Next major version has been delayed strongly . Into that version there is a CUSD related code published in this post:

yjh wrote

3. NOTE (unexpected addonce ).
If the USSD command is defined in regular_run_cmd
and if this command receive replay in HEX-UCS2,
this string is automatically converted to UTF8.

It's not the same as in your case, but should be handled with some new flag. In your case, probably a new flag is needed to define that USSD Packing is used (GSM 03.38 6.1.2.3.1).

No worries, the hack is simple, first in smsd.c
--- smsd.c.orig 2009-11-30 18:53:22.000000000 +0100 +++ smsd.c 2010-03-23 07:43:58.000000000 +0100 @@ -4847,6 +4847,8 @@ if (*answer) { char timestamp[81]; + char answer_a[256] = ""; + char tmpanswer[256] = ""; time_t now; time(&now); @@ -4861,6 +4863,21 @@ if (*answer == ' ') strcpy(answer, answer +1); + if ((p = strstr(answer, "+CUSD: 0"))) { + strncpy(tmpanswer, p+10, strlen(p)-15); + tmpanswer[strlen(p)-5] = '\0'; +#ifdef DEBUGMSG + printf("convert to text: %s\n", tmpanswer); +#endif + if (toASCII(tmpanswer, answer_a) < 0) { +// todo + } +#ifdef DEBUGMSG + printf("text converted: %s\n", answer_a); +#endif + + } + if (DEVICE.dev_rr_logfile[0]) { while (log_retry-- > 0) @@ -4868,8 +4885,8 @@ // NOTE: log files use mode 640 as a fixed value. if ((fd = open(DEVICE.dev_rr_logfile, O_APPEND|O_WRONLY|O_CREAT, 0640)) >= 0) { - snprintf(buffer, sizeof(buffer), "%s,%i, %s: CMD: %s: %s\n", - timestamp, DEVICE.dev_rr_loglevel, DEVICE.name, command, answer); + snprintf(buffer, sizeof(buffer), "%s,%i, %s: CMD: %s: %s // %s\n", + timestamp, DEVICE.dev_rr_loglevel, DEVICE.name, command, answer, answer_a); write(fd, buffer, strlen(buffer)); close(fd); break; @@ -4885,7 +4902,7 @@ } } else - writelogfile(DEVICE.dev_rr_loglevel, 0, "CMD: %s: %s", command, answer); + writelogfile(DEVICE.dev_rr_loglevel, 0, "CMD: %s: %s // %s", command, answer, answer_a); // 3.1.5: If signal quality was checked, explain it to the log: if (!strcasecmp(cmd, "AT+CSQ\r")) @@ -4897,10 +4914,10 @@ //if ((fd = open(DEVICE.dev_rr_statfile, O_APPEND | O_WRONLY | O_CREAT, 0640)) >= 0) if ((fp = fopen(DEVICE.dev_rr_statfile, "a"))) { - //snprintf(buffer, sizeof(buffer), "%s,%i, %s: CMD: %s: %s\n", timestamp, LOG_NOTICE, DEVICE.name, command, answer); + //snprintf(buffer, sizeof(buffer), "%s,%i, %s: CMD: %s: %s // %s\n", timestamp, LOG_NOTICE, DEVICE.name, command, answer, answer_a); //write(fd, buffer, strlen(buffer)); //close(fd); - fprintf(fp, "%s,%i, %s: CMD: %s: %s\n", timestamp, LOG_NOTICE, DEVICE.name, command, answer); + fprintf(fp, "%s,%i, %s: CMD: %s: %s // %s\n", timestamp, LOG_NOTICE, DEVICE.name, command, answer, answer_a); fclose(fp); } else

and pduconv.c (found somewhere in the net):
unsigned char send[160]; //////////////////////////////////////////////////////////////////////////////// void toPDU(char *asci, char *temp) { unsigned short pos=0, get=1, count=0,h=0,len=0; len=strlen(asci); if(15>=len>= 8 ) len--; else len=len-(len/8 ); while (pos!=len) { switch (count) { case 0: get=1; break; case 1: get=3; break; case 2: get=7; break; case 3: get=15; break; case 4: get=31; break; case 5: get=63; break; case 6: get=127; break; } temp[pos]=( asci[pos+h]>>count) | (( asci[pos+(h+1)]& get)<<(7-count)); pos++; count++; if (count>6) count=0, h++; } } //////////////////////////////////////////////////////////////////////////////// //////////////////////////////////////////////////////////////////////////////// void toASCII(char *asci, char temp[]) { unsigned short pos=0, get1=127, get2=128, count=0,h=0,len=0, alpha=0,mem0=0,mem1=0; unsigned short convert[160]; len=strlen(asci); while(*asci) { alpha=*asci; switch (alpha ) { case 48: alpha=0; break; case 49: alpha=1; break; case 50: alpha=2; break; case 51: alpha=3; break; case 52: alpha=4; break; case 53: alpha=5; break; case 54: alpha=6; break; case 55: alpha=7; break; case 56: alpha=8; break; case 57: alpha=9; break; case 65: alpha=10; break; case 66: alpha=11; break; case 67: alpha=12; break; case 68: alpha=13; break; case 69: alpha=14; break; case 70: alpha=15; break; } convert[pos]=alpha; pos++; *asci++; } pos=0; while (pos!=len) { convert[pos]=(convert[pos+pos]*16)+ convert[pos+pos+1]; pos++; } pos=0; count=0; while (pos!=(len/2+h)) { switch (count) { case 0: get1=127; get2=128; break; case 1: get1=63; get2=128; break; case 2: get1=31; get2=192; break; case 3: get1=15; get2=224; break; case 4: get1=7; get2=240; break; case 5: get1=3; get2=248; break; case 6: get1=1; get2=252; break; } temp[pos]=((convert[pos-h]& get1)<<(count)) | ((convert[pos-h-1]& get2) >>(7-(count-1))); pos++; count++; if(count==7) { temp[pos]=convert[pos-h-1]>>1; convert[pos-h-1]=0; pos++; count=0; h++; } } } ////////////////////////////////////////////////////////////////////////////////

pduconv.c needs to be added in the makefile and compiled in...

I'm sorry for the coding style, but that's all I'm able to

Thanks, I have to think about how this functionality could be implemented.

Probably this could be an external program which can be used to create USSD packed string, and decode it. This kind of helper could be used when creating a setup, or when creating a regular_run_cmdfile content. External helper program could be usefull with older versions of smsd too.

ok. let me know if I can be of any further help (at least for testing then).

btw: great sowftware!

Have you seen and tested 3.1.7beta4 and the modem setting ussd_convert = 2?

It applies for received strings only. A tool PDU Converter can be used to create a packed string to send to the modem. Also that version of smsd has a command line option -Ex for encoding, and -Dx for decoding.

Hi

yes!! installed it last week (beta4) and it works like a charm! thanks a lot!! great work!

rgds

Thanks for this great software !

As we see more and more of those (crappy) Huawei sticks around, I guess an USSD packed string convertion is more and more needed. For incoming stuff ussd_convert = 2 works well. Outgoing is not a problem as long as the USSD message is static like a *100#. Things change if someone has to enter special stuff like a recharging code. Sure you can externally convert it to PDU, but smstools do have all needed on board already.

muellmaen wrote

Thanks for this great software !

Thanks for choosing this!

muellmaen wrote

Outgoing is not a problem as long as the USSD message is static like a *100#. Things change if someone has to enter special stuff like a recharging code. Sure you can externally convert it to PDU, but smstools do have all needed on board already.

Because PDU coded USSD messages are often static too, smsd has no internal conversion to PDU on the fly. The string is sent as it is, which also means that no additional definitions are needed. However, if different stuff is required, regular_run script can be used to create regular_run_cmdfile. The option smsd -E message can be used for encoding.