more standard conform handling for dvb-text encodings

with this fix no more special handling for UPC Direkt transponders
on 19.2°E and CS Link on 23.5°E is needed
but some polish providers still needs special handling... i think they
interpreted the spec in the wrong way

diff --git a/data/encoding.conf b/data/encoding.conf
index 29816ce3126a67784cf2577a616f4a917cd15a17..691f00276a0af074531b79b27a63b7c33c71fa4a 100644 (file)
--- a/data/encoding.conf
+++ b/data/encoding.conf
@@ -1,24 +1,25 @@
-#Fallback encoding when in dvb-text no encoding table is given
+# Fallback encoding when in dvb-text no encoding table is given. In normal
+# case this then is Latin1/ISO6397 with enabled two char byte mapping.
+# But here we override this for all EIT entries with given country code
+
 #Countycode ISO8859-X
 tur ISO8859-9
 gre ISO8859-7
 pol ISO8859-2
 rus ISO8859-5
 bul ISO8859-5
-#Table of transponder using two byte char mappings (VideoTexSuppl)
+
+# Table of transponder using two byte char mappings
+# in normal case two char byte mapping is only used when in dvb-text no
+# encoding is given... but some providers use two-char byte mapping in 
+# combination with other ISO8859-X tables.. so here it can be enabled
+# for some TSID/ONID combinations
+# a special case is when in dvb-text ISO8859-5 is selected... then
+# we also change the char table to ISO6397 this is used by 
+# Cyfra+ and Cyfrowy Polsat and other polish providers
+# please report when you have an idea how to auto detect this!
+
 #TSID ONID
-0x447 0x1 #ASTRA 19.2° UPC Direct
-0x427 0x1
-0x44b 0x1
-0x4ff 0x1
-0x407 0x1
-0x436 0x1 #ASTRA 19.2° MTV Euro - MTV Networks
-0x42a 0x1 #ASTRA 19.2° VH1 Classic - MTV Networks
-0xbc6 0x3 #ASTRA 23.5° Cslink, Skylink
-0xc85 0x3
-0xc89 0x3
-0xc8f 0x3
-0xbc7 0x3
 400 318 #Hotbird 13.0° Cyfra+
 1500 318
 11000 318
@@ -33,9 +34,13 @@ bul ISO8859-5
 12200 318 #Disney Channel Polska and other
 12800 318 #Viacom ... MTV / VH1 Polska
 13000 318 #BBC Polska and other
-#Fallback encoding table for following transponders
+
+# Fallback encoding when in dvb-text no encoding table is given. In normal
+# case this then is Latin1/ISO6397 with enabled two char byte mapping.
+# But here we override this for all EIT entries on transponders with given
+# TSID/ONID combination
+
 #TSID ONID ISO8859-X
-#0x447 0x1 ISO8859-9
 50200 126 ISO8859-9 # Digiturk 7°E 11.492 V 30.000 3/4
 50300 126 ISO8859-9 # Digiturk 7°E 11.639 H 30.000 3/4
 50400 126 ISO8859-9 # Digiturk 7°E 11.534 V 30.000 3/4
@@ -44,4 +49,3 @@ bul ISO8859-5
 50800 126 ISO8859-9 # Digiturk 7°E 11.678 H 30.000 3/4
 50900 126 ISO8859-9 # Skyturk  7°E 11.513 H 27.500 3/4
 51000 126 ISO8859-9 # Digiturk 7°E 11.617 V 30.000 3/4
-12800 318 ISO6397 #Viacom ... MTV / VH1 Polska

diff --git a/lib/base/encoding.cpp b/lib/base/encoding.cpp
index 831de8474583c581ff1596de4ce8f58f4317e4fc..e4e1531c921d4ed4483fc1d8608703b8c1ba1501 100644 (file)
--- a/lib/base/encoding.cpp
+++ b/lib/base/encoding.cpp
@@ -42,14 +42,14 @@ eDVBTextEncodingHandler::eDVBTextEncodingHandler()
                        }
                        else if ( (sscanf( line, "0x%x 0x%x ISO%d", &tsid, &onid, &encoding ) == 3 && encoding == 6397 )
                                        ||(sscanf( line, "%d %d ISO%d", &tsid, &onid, &encoding ) == 3 && encoding == 6397 ) )
-                               m_TransponderDefaultMapping[(tsid<<16)|onid]=64;
+                               m_TransponderDefaultMapping[(tsid<<16)|onid]=0;
                        else if ( sscanf( line, "%s ISO%d", countrycode, &encoding ) == 2 && encoding == 6397 )
                        {
-                               m_CountryCodeDefaultMapping[countrycode]=64;
+                               m_CountryCodeDefaultMapping[countrycode]=0;
                                countrycode[0]=toupper(countrycode[0]);
                                countrycode[1]=toupper(countrycode[1]);
                                countrycode[2]=toupper(countrycode[2]);
-                               m_CountryCodeDefaultMapping[countrycode]=64;
+                               m_CountryCodeDefaultMapping[countrycode]=0;
                        }
                        else if ( (sscanf( line, "0x%x 0x%x", &tsid, &onid ) == 2 )
                                        ||(sscanf( line, "%d %d", &tsid, &onid ) == 2 ) )
@@ -83,5 +83,5 @@ int eDVBTextEncodingHandler::getCountryCodeDefaultMapping( const std::string &co
                m_CountryCodeDefaultMapping.find(country_code);
        if ( it != m_CountryCodeDefaultMapping.end() )
                return it->second;
-       return 0;  // ISO8859-1 / Latin1
+       return 0;  // Latin1/ISO6397
 }

diff --git a/lib/base/estring.cpp b/lib/base/estring.cpp
index be071454923ff6be1bc1f32fc61dbc96fbd6fa28..728a4c4fd26b55512e1df75ff812201bd67dce10 100644 (file)
--- a/lib/base/estring.cpp
+++ b/lib/base/estring.cpp
@@ -152,11 +152,11 @@ static unsigned long c885916[96]={
 0x0111, 0x0144, 0x00F2, 0x00F3, 0x00F4, 0x0151, 0x00F6, 0x015B, 0x0171, 0x00F9, 0x00FA, 0x00FB, 0x00FC, 0x0119, 0x021B, 0x00FF};
 
 static unsigned long iso6397[96]={
-0x00A0, 0x00A1, 0x00A2, 0x00A3, 0x0000, 0x00A5, 0x0000, 0x00A7, 0x00A4, 0x2018, 0x201C, 0x00AB, 0x2190, 0x2191, 0x2192, 0x2193,
+0x00A0, 0x00A1, 0x00A2, 0x00A3, 0x20AC, 0x00A5, 0x0000, 0x00A7, 0x00A4, 0x2018, 0x201C, 0x00AB, 0x2190, 0x2191, 0x2192, 0x2193,
 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00D7, 0x00B5, 0x00B6, 0x00B7, 0x00F7, 0x2019, 0x201D, 0x00BB, 0x00BC, 0x00BD, 0x00BE, 0x00BF,
 0x0000, 0xE002, 0xE003, 0xE004, 0xE005, 0xE006, 0xE007, 0xE008, 0xE009, 0xE00C, 0xE00A, 0xE00B, 0x0000, 0xE00D, 0xE00E, 0xE00F,
-0x2014, 0x00B9, 0x00AE, 0x00A9, 0x2122, 0x266A, 0x00AC, 0x00A6, 0x0000, 0x0000, 0x0000, 0x0000, 0x215B, 0x215C, 0x215D, 0x215E,
-0x2126, 0x00C6, 0x00D0, 0x00AA, 0x0126, 0x0000, 0x0132, 0x013F, 0x0141, 0x00D8, 0x0152, 0x00BA, 0x00DE, 0x0166, 0x014A, 0x0149,
+0x2015, 0x00B9, 0x00AE, 0x00A9, 0x2122, 0x266A, 0x00AC, 0x00A6, 0x0000, 0x0000, 0x0000, 0x0000, 0x215B, 0x215C, 0x215D, 0x215E,
+0x2126, 0x00C6, 0x0110, 0x00AA, 0x0126, 0x0000, 0x0132, 0x013F, 0x0141, 0x00D8, 0x0152, 0x00BA, 0x00DE, 0x0166, 0x014A, 0x0149,
 0x0138, 0x00E6, 0x0111, 0x00F0, 0x0127, 0x0131, 0x0133, 0x0140, 0x0142, 0x00F8, 0x0153, 0x00DF, 0x00FE, 0x0167, 0x014B, 0x00AD};
 
 // Two Char Mapping (aka ISO6397) ( many polish services and UPC Direct/HBO services)
@@ -325,7 +325,8 @@ static inline unsigned int recode(unsigned char d, int cp)
                return d;
        switch (cp)
        {
-       case 0:         // Latin1 <-> unicode mapping
+       case 0:         // Latin1/ISO6397 <-> unicode mapping
+               return iso6397[d-0xA0];
        case 1:         // 8859-1 <-> unicode mapping
                return d;
        case 2:         // 8859-2 -> unicode mapping
@@ -358,8 +359,6 @@ static inline unsigned int recode(unsigned char d, int cp)
                return c885915[d-0xA0];
        case 16:// 8859-16 -> unicode mapping
                return c885916[d-0xA0];
-       case 64:// iso6397
-               return iso6397[d-0xA0];
        default:
                return d;
        }
@@ -372,6 +371,10 @@ std::string convertDVBUTF8(const unsigned char *data, int len, int table, int ts
 
        int i=0, t=0;
 
+       // table given two this function is default 0... but when a default table for a country code is set in encoding.conf
+       // then here this table is given
+
+       // when a tsid/onid is avail the table can be overriden in encoding.conf based on tsid/onid combinations
        if ( tsidonid )
                encodingHandler.getTransponderDefaultMapping(tsidonid, table);
 
@@ -400,7 +403,7 @@ std::string convertDVBUTF8(const unsigned char *data, int len, int table, int ts
                }
                case 0x11: //  Basic Multilingual Plane of ISO/IEC 10646-1 enc  (UTF-16... Unicode)
                        table = 65;
-                       tsidonid = 0;
+//                     eDebug("(0x11)text encoded in ISO-10646-1 (UTF-16)");
                        ++i;
                        break;
                case 0x12:
@@ -423,16 +426,22 @@ std::string convertDVBUTF8(const unsigned char *data, int len, int table, int ts
                        eDebug("reserved %d", data[0]);
                        ++i;
                        break;
+               default:
+//                     eDebug("no encoding in dvb string given.. use default table");
+                       break;
        }
 
-       bool useTwoCharMapping =
-               tsidonid && encodingHandler.getTransponderUseTwoCharMapping(tsidonid);
+       bool useTwoCharMapping = (!table) || (tsidonid && encodingHandler.getTransponderUseTwoCharMapping(tsidonid));
+       // two char byte mapping is default enabled for table 0 (ISO6397) (described in ETSI EN300468)
+       // or it can be enabled for tsid/onid combinations in our encoding.conf
 
-       if (useTwoCharMapping) {
-               if (table == 5)
-                       table = 64;
+       if (useTwoCharMapping && table == 5) { // i hope this dont break other transponders which realy use ISO8859-5 and two char byte mapping...
+//             eDebug("Cyfra / Cyfrowy Polsat HACK... override given ISO8859-5 with ISO6397");
+               table = 0;
        }
 
+       eDebug("used table is now %d, twoChar %d", table, useTwoCharMapping);
+
        unsigned char res[2048];
        while (i < len)
        {
@@ -536,8 +545,8 @@ std::string convertUTF8DVB(const std::string &string, int table)
                                                case 11:
                                                        coding_table = c885911;
                                                        break;
-/*                             case 12:   // reserved.. for indian use
-                                               coding_table = c885912;
+/*                                             case 12:   // reserved.. for indian use
+                                                       coding_table = c885912;
                                                break;*/
                                                case 13:
                                                        coding_table = c885913;