Another related issue; How is possible that while connecting to the same SMSC via SMPP from another SMPP gateway (test purpose) I`m able to send 7-bit alphabet characters according to the specification here http://www.cardboardfish.com/support/bin/view/Main/GSMEncoding I dont set any coding schema and the accents characters are sent properly. Is it Kannel issue then to involve 8-bit in order to support this?
Regards Elton 2009/5/5 Nikos Balkanas <[email protected]> > Hi Elton, > > If you consult any ASCII table, you can verify for yourself that any such > characters are above '~', which is 126. The only way to transmit such > characters within 7 bits, is to throw away some other character and take its > place. I presume that the aforementioned mobiles, agree on some special > non-ASCII encoding and can use that. I would imagine that: > > 1) You need to have same regional settings in each mobile, and possible > same brand? > 2) You need to have the same extended encoding activated. Which means you > are missing on regular characters. > > I wish I could be more specific, but I don't have a mobile. Maybe a few > other guys from the list can help out. > > BR, > Nikos > ----- Original Message ----- From: "elthox" <[email protected]> > To: <[email protected]> > Sent: Tuesday, May 05, 2009 9:46 AM > Subject: Re: kannel and greek characters > > > > Hi Nikos, > > You mentioned that umlauts and accents are extended characters that are out > of the range of 7-bit alphabet. How come, when sending mobile to mobile > SMS, > with a reduced character set (7-bit) the message is delivered properly > writing é, ñ etc. But while trying it from SMPP channel doesnt work, for > example giving the encoded hexa value %05 for é, it will translate it as > > "?". > > Do you have anyidea? > > Regards > Elton > > > > Nikos Balkanas wrote: > >> >> Hi Nick, >> >> Almost, not quite. GSM encoding is not doing any recoding at the character >> level, just at the byte level. I.e. all ASCII characters below 128 can be >> represented as 7 bits and these include all the major Latin-1 or >> ISO-8859-1. >> There are some extended characters in ISO-8859-1 (umlauts, accents etc) >> that >> if you include even a single one it will fall back to 8bit. So sending >> capitalized english chars as Greek to save SMS space, might work as long >> as >> you stick to "Greeklish". >> >> BR, >> Nikos >> ----- Original Message ----- From: "Nick" <[email protected]> >> To: <[email protected]> >> Sent: Friday, March 27, 2009 2:38 PM >> Subject: Re: kannel and greek characters >> >> >> Thanks Nikos >> >> When i set the coding to 1 or to 2 , i cant send 160 char messages anymore >> >> There must be a way to encode the message to GSM encoding , and send 160 >> char messages with capital greek characters >> >> From this URL >> >> http://www.cardboardfish.com/support/bin/view/Main/GSMEncoding >> >> According to GSM specification, a standard SMS message can contain up to >> 140 bytes of data (payload). Standard latin (ISO-8859-1) character >> encoding >> represents a single character using 1 byte, which is 8 bits. Therefore, >> the >> maximum number of latin 1 characters that could be included in an sms is >> 140. >> >> GSM encoding represents characters using 7 bits instead of 8. This >> therefore provides a maximum of 160 characters per SMS. (140 * 8 bits) / 7 >> bits = 160 >> >> This effectively halves the number of characters that the GSM character >> set >> can support, compared to ISO-8859-1. In order to include common characters >> that are usually represented using the 8th bit, these characters as well >> as >> other symbol characters must be re-mapped to a combination of lower bits. >> These re-mapped characters are often referred to as special characters. >> This re-mapping, in combination with packing 7-bit characters into 8-bit >> bytes is called GSM Encoding. >> >> >> >> Hi Nick, >>> >>> SMS can hold 160 B of 7bit chars or 140 B of 8bit chars (+/- some UDF >>> headers). Half of that for Unicode. ASCII characters are 7bit, Greek are >>> 8bit. If you choose to send them as Unicode it is 70 B. >>> >>> In your URL coding=0 corresponds to 7bit coding. Wrong one for Greek. Try >>> with coding 2 (Unicode) or coding 1 (8bit). >>> >>> BR, >>> Nikos >>> ----- Original Message ----- From: "Nick" <[email protected]> >>> To: <[email protected]> >>> Sent: Thursday, March 26, 2009 9:14 PM >>> Subject: kannel and greek characters >>> >>> >>> Hello all >>> >>> I try to send greek sms 160 characters long with kannel with no luck >>> I can send 70 characters Unicode messages successfully, but cant send 160 >>> char long messages >>> >>> I have installed kannel on 2 linux servers, one with centos 4.6 and one >>> with centos 5.1 >>> I have installed the latest version 1.4.3 and the latest cvs version >>> I have compiled both version with and without the greek patch bellow >>> http://www.kannel.org/pipermail/users/2006-July/000040.html >>> I have read and followed all post with keyword "greek" in both users and >>> developers mailing list >>> >>> No luck >>> >>> According the posts in the mailling list, when i send >>> >>> http://192.168.0.1:13131/cgi-bin/sendsms?user=****&pass=****&to=306977000000&coding=0&text=KA%14HMEPA >>> (%14 = greek letter Λ) >>> it suppose to send the message ΠΑΛΗΠΕΡΑ >>> >>> It does not. >>> I get ?????? or sometimes _____ >>> >>> I have tried countless combinations of all available query string and >>> config variables (coding charset alt charset etc) >>> >>> Since i know that many users are using kannel and send greek characters >>> successfully, >>> i am sure am missing something, but i cant figure out what it is. >>> >>> Any help will be appreciated >>> >>> Thank you >>> >> >> >> >> >> >> > -- > View this message in context: > http://www.nabble.com/kannel-and-greek-characters-tp22729188p23381944.html > Sent from the Kannel - User mailing list archive at Nabble.com. > > >
