From 2ac9030e653c55d81e311b90d68669d4f97c2b1c Mon Sep 17 00:00:00 2001 From: Klaus Schmidinger Date: Sun, 22 Apr 2007 14:49:26 +0200 Subject: Implemented character set conversion in 'libsi' --- HISTORY | 21 ++++++++++++++++++++- 1 file changed, 20 insertions(+), 1 deletion(-) (limited to 'HISTORY') diff --git a/HISTORY b/HISTORY index ecec6c02..3aa97f5f 100644 --- a/HISTORY +++ b/HISTORY @@ -5139,10 +5139,29 @@ Video Disk Recorder Revision History parameter to 0 turns off the automatic channel switching, and the user will have to confirm the entry by pressing the "Ok" key. -2007-03-10: Version 1.5.2 +2007-04-22: Version 1.5.2 - Updated the Finnish OSD texts (thanks to Rolf Ahrenberg). - Fixed handling user activity for shutdown, which I had messed when adopting Udo's original patch (thanks to Udo Richter). - Added Turkish language texts (thanks to Oktay Yolgeçen). - Added missing rules for generating iso8859-13 font to Makefile. +- 'libsi' now converts the incoming strings into the system's character set + according to the DVB standard. The system's character set is determined from + the LANG environment variable. If no recognizable setting can be found, no + conversion will take place. Note that currently only the strings received from the + SI data stream are converted, there have not been any changes regarding displaying + UTF-8 characters on the OSD, yet - this will follow in one of the next steps. + With this conversion, it should now be safe to run VDR on a UTF-8 file system, + because all incoming characters are converted to UTF-8. This will most likely + result in wrong characters being displayed on the OSD (because there UTF-8 is + not known, yet), but the file names should be ok (haven't tested this myself, + though, because I don't do UTF-8 - so please be very careful when testing!). + There's one piece of bad news here: the German pay-tv broadcaster Premiere + apparently encodes all EPG strings as ISO8859-1, but fails to correctly mark + these strings as such. Therefore 'libsi' (following the DVB standard) considers + the strings to be encoded in the default ISO6937 and converts them to whatever + the system's character set is. This, of course, results in wrong umlauts. + On its old transponder, the ProSieben/SAT.1 channels also had their EPG data + wrongly encoded, but apparently on the new transponder they started broadcasting + on this month, they got it right. -- cgit v1.2.3