From ab308bea3175dd21139b7b842b6de56bf88d8b1c Mon Sep 17 00:00:00 2001 From: Klaus Schmidinger Date: Thu, 14 May 2020 21:21:03 +0200 Subject: Fixed handling multi part ExtendedEventDescriptors where only the first part contains information about the character table --- libsi/si.c | 112 +++++++++++++++++++++++++++++++++---------------------------- 1 file changed, 61 insertions(+), 51 deletions(-) (limited to 'libsi/si.c') diff --git a/libsi/si.c b/libsi/si.c index e51770ab..c294c2fc 100644 --- a/libsi/si.c +++ b/libsi/si.c @@ -6,7 +6,7 @@ * the Free Software Foundation; either version 2 of the License, or * * (at your option) any later version. * * * - * $Id: si.c 3.3 2015/02/10 13:42:41 kls Exp $ + * $Id: si.c 4.1 2020/05/14 21:21:03 kls Exp $ * * ***************************************************************************/ @@ -230,14 +230,14 @@ char *String::getText() { return data; } -char *String::getText(char *buffer, int size) { +char *String::getText(char *buffer, int size, const char **fromCode) { int len=getLength(); if (len < 0 || len >= size) { strncpy(buffer, "text error", size); buffer[size-1] = 0; return buffer; } - decodeText(buffer, size); + decodeText(buffer, size, fromCode); return buffer; } @@ -386,9 +386,25 @@ const char *getCharacterTable(const unsigned char *&buffer, int &length, bool *i return cs; } -bool convertCharacterTable(const char *from, size_t fromLength, char *to, size_t toLength, const char *fromCode) +// A similar version is used in VDR/tools.c: +static int Utf8CharLen(const char *s) +{ + if (SystemCharacterTableIsSingleByte) + return 1; +#define MT(s, m, v) ((*(s) & (m)) == (v)) // Mask Test + if (MT(s, 0xE0, 0xC0) && MT(s + 1, 0xC0, 0x80)) + return 2; + if (MT(s, 0xF0, 0xE0) && MT(s + 1, 0xC0, 0x80) && MT(s + 2, 0xC0, 0x80)) + return 3; + if (MT(s, 0xF8, 0xF0) && MT(s + 1, 0xC0, 0x80) && MT(s + 2, 0xC0, 0x80) && MT(s + 3, 0xC0, 0x80)) + return 4; + return 1; +} + +size_t convertCharacterTable(const char *from, size_t fromLength, char *to, size_t toLength, const char *fromCode) { - if (SystemCharacterTable) { + char *result = to; + if (SystemCharacterTable && fromCode) { iconv_t cd = iconv_open(SystemCharacterTable, fromCode); if (cd != (iconv_t)-1) { char *fromPtr = (char *)from; @@ -407,29 +423,44 @@ bool convertCharacterTable(const char *from, size_t fromLength, char *to, size_t } *to = 0; iconv_close(cd); - return true; } } - return false; -} - -// A similar version is used in VDR/tools.c: -static int Utf8CharLen(const char *s) -{ - if (SystemCharacterTableIsSingleByte) - return 1; -#define MT(s, m, v) ((*(s) & (m)) == (v)) // Mask Test - if (MT(s, 0xE0, 0xC0) && MT(s + 1, 0xC0, 0x80)) - return 2; - if (MT(s, 0xF0, 0xE0) && MT(s + 1, 0xC0, 0x80) && MT(s + 2, 0xC0, 0x80)) - return 3; - if (MT(s, 0xF8, 0xF0) && MT(s + 1, 0xC0, 0x80) && MT(s + 2, 0xC0, 0x80) && MT(s + 3, 0xC0, 0x80)) - return 4; - return 1; + else { + size_t len = fromLength; + if (len >= toLength) + len = toLength - 1; + strncpy(to, from, len); + to[len] = 0; + } + // Handle control codes: + to = result; + size_t len = strlen(to); + while (len > 0) { + int l = Utf8CharLen(to); + if (l <= 2) { + unsigned char *p = (unsigned char *)to; + if (l == 2 && *p == 0xC2) // UTF-8 sequence + p++; + bool Move = true; + switch (*p) { + case 0x8A: *to = '\n'; break; + case 0xA0: *to = ' '; break; + default: Move = false; + } + if (l == 2 && Move) { + memmove(p, p + 1, len - 1); // we also copy the terminating 0! + len -= 1; + l = 1; + } + } + to += l; + len -= l; + } + return strlen(result); } // originally from libdtv, Copyright Rolf Hakenes -void String::decodeText(char *buffer, int size) { +void String::decodeText(char *buffer, int size, const char **fromCode) { const unsigned char *from=data.getData(0); char *to=buffer; int len=getLength(); @@ -437,38 +468,17 @@ void String::decodeText(char *buffer, int size) { *to = '\0'; return; } - bool singleByte; - const char *cs = getCharacterTable(from, len, &singleByte); - if (singleByte && SystemCharacterTableIsSingleByte || !convertCharacterTable((const char *)from, len, to, size, cs)) { + const char *cs = getCharacterTable(from, len); + if (fromCode) { if (len >= size) len = size - 1; - strncpy(to, (const char *)from, len); - to[len] = 0; + strncpy(buffer, (const char *)from, len); + buffer[len] = 0; + if (!*fromCode) + *fromCode = cs; } else - len = strlen(to); // might have changed - // Handle control codes: - while (len > 0) { - int l = Utf8CharLen(to); - if (l <= 2) { - unsigned char *p = (unsigned char *)to; - if (l == 2 && *p == 0xC2) // UTF-8 sequence - p++; - bool Move = true; - switch (*p) { - case 0x8A: *to = '\n'; break; - case 0xA0: *to = ' '; break; - default: Move = false; - } - if (l == 2 && Move) { - memmove(p, p + 1, len - 1); // we also copy the terminating 0! - len -= 1; - l = 1; - } - } - to += l; - len -= l; - } + convertCharacterTable((const char *)from, len, to, size, cs); } void String::decodeText(char *buffer, char *shortVersion, int sizeBuffer, int sizeShortVersion) { -- cgit v1.2.3