Can you provide me with a brief synopsis on updating the translation tables. I'm assuming (correct me if I'm wrong) the updates are applied into charset.c and then you recompile. Is this correct?

I've got a ♥ symbol that sometimes comes in on messages and it doesn't do to well.

Also, if you know, where can I find current a ISO character set to plug in, that would be great.

Log info:2010-07-07 15:08:00,5, PHS1: Cannot convert 83. character  0x 3 to GSM, you might need to update the translation tables.


I also just noticed that cygwin doesn't like the character either so that may be problematic.

Characters outside the GSM alphabet are ignored. That message in the log is somewhat misleading, in general there is no need nor the opportunity to update translation tables.

If you need to send those kind of symbols, you must convert message body to Unicode and use Alphabet: UCS header.

In the file charset.c, there is "optical near" table called iso_8859_15_chars[]. If you want to, you can add more replacements to this table, before the ending bytes (0, 0).

The GSM alphabet is shown here.

I was a bit off on the original problem. The messages I was receiving were ucs2 and not ISO. Oddly enough, they get delivered ok but the storage in the mysql database shows the msg a bit garbled (0�O��.�.�r�e� �n�a� �k�) for example.

You could try a modem setting decode_unicode_text = yes. If it does not work properly with your messages, there is ICONV style also available. In Cygwin you need to install iconv sources and libraries. In the src/Makefile there are two lines which should be edited before recompiling the smsd.