Version 1.5.3vdr-1.5.3

- Fixed some spelling errors in 'newplugin' (thanks to Ville Skyttä).
- Fixed a busy loop in fast forward if the next video data file is missing
  (thanks to Reinhard Nissl).
- Fixed handling frequencies in NitFilter::Process() (thanks to Anssi Hannula).
- Fixed a race condition with signal handlers at program exit (thanks to Udo
  Richter).
- Non-primary devices in Transfer mode are now also used for recording (thanks
  to Anssi Hannula).
- Fixed handling ChannelUp/Down keys if there is currently a replay running
  (thanks to Marco Schlüßler).
- The new SVDRP command REMO can be used to turn VDR's remote control off and
  on in case other programs need to be controlled (based on patches from Krzysztof
  Parma and Helmut Auer).
- Increased the maximum number of CA system ids to cope with the AlphaCrypt
  CAM's version 3.11 firmware.
- Fixed getting the code setting from the locale (thanks to Matthias Schwarzott).
- Implemented support for Freetype fonts (based on a patch from Alexander Riedel).
  The font names and sizes can be adjusted in the "Setup/OSD" menu.
  Note that VDR now requires freetype fonts to be installed in
  /usr/share/fonts/truetype.
- If the OSD device in use has at least 8bpp bitmap depth and this is also
  used by the current skin, Freetype fonts are displayed "anti-aliased".
  The new setup parameter "OSD/Anti-alias" can be used to turn this off.
- The new function cOsd::SetAntiAliasGranularity() can be used to help the OSD
  in managing the available color palette entries when doing anti-aliasing.
  Skins that use 8bpp bitmaps can call this function with the maximum number
  of colors used, and the maximum number of color combinations. The OSD will
  then evenly split the available palette entries between the various colors
  combinations, so that fonts can be "anti-aliased". By default a total of
  10 colors and 10 combinations is assumed.
- The pixel fonts have been completely removed from the VDR source.
- VDR is now "UTF-8 aware". It handles strings according to the character
  encoding used on the user's system. All internationalization strings and
  incoming SI data are converted to the system encoding.
- Plugins that handle strings need to be aware that on systems with UTF-8
  encoding a "character symbol" may consist of more than a single byte in
  memory. The functions and macros named Utf8...() can be used to handle
  strings without needing to care about the underlying character encoding
  (see tools.h for details).
- Even though the weekdays of repeating timers are presented to the user as UTF-8
  characters in the OSD, the timers.conf file and the SVDRP timer commands still
  use single byte characters ("MTWTFSS") to make sure this information is handled
  correctly between systems with different character encodings.
- Added a missing i18n string for "CAM" in the Turkish OSD texts.
- Improved editing strings that are too long to fit into the editable area.
- Changes to the OSD settings in the "Setup/OSD" menu now immediately take effect
  when the "Ok" key is pressed.

1 files changed, 11 insertions, 3 deletions

diff --git a/libsi/si.c b/libsi/si.c
index a1d2f34..2eb3dfd 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 1.18 2007/04/22 13:56:39 kls Exp $
+ *   $Id: si.c 1.20 2007/06/10 09:31:34 kls Exp $
  *                                                                         *
  ***************************************************************************/
 
@@ -219,7 +219,9 @@ char *String::getText() {
    int len=getLength();
    if (len < 0 || len > 4095)
       return strdup("text error"); // caller will delete it!
-   char *data=new char(len+1);
+   char *data=new char(len+1); // FIXME If this function is ever actually used, this buffer might
+                               // need to be bigger in order to hold the string as UTF-8.
+                               // Maybe decodeText should dynamically handle this? kls 2007-06-10
    decodeText(data, len+1);
    return data;
 }
@@ -370,7 +372,7 @@ static bool convertCharacterTable(const char *from, size_t fromLength, char *to,
 {
   if (SystemCharacterTable) {
      iconv_t cd = iconv_open(SystemCharacterTable, fromCode);
-     if (cd >= 0) {
+     if (cd != (iconv_t)-1) {
         char *fromPtr = (char *)from;
         while (fromLength > 0 && toLength > 1) {
            if (iconv(cd, &fromPtr, &fromLength, &to, &toLength) == size_t(-1)) {
@@ -404,6 +406,9 @@ void String::decodeText(char *buffer, int size) {
       }
    bool singleByte;
    const char *cs = getCharacterTable(from, len, &singleByte);
+   // FIXME Need to make this UTF-8 aware (different control codes).
+   // However, there's yet to be found a broadcaster that actually
+   // uses UTF-8 for the SI data... (kls 2007-06-10)
    for (int i = 0; i < len; i++) {
       if (*from == 0)
          break;
@@ -440,6 +445,9 @@ void String::decodeText(char *buffer, char *shortVersion, int sizeBuffer, int si
       }
    bool singleByte;
    const char *cs = getCharacterTable(from, len, &singleByte);
+   // FIXME Need to make this UTF-8 aware (different control codes).
+   // However, there's yet to be found a broadcaster that actually
+   // uses UTF-8 for the SI data... (kls 2007-06-10)
    for (int i = 0; i < len; i++) {
       if (    ((' ' <= *from) && (*from <= '~'))
            || (*from == '\n')