hgm.bg schrieb: > vdr-bounces@xxxxxxxxxxx wrote: > > >>hgm.bg wrote: >> >>>vdr-bounces@xxxxxxxxxxx wrote: >>> >>>It doesn't help: >>>Program received signal SIGSEGV, Segmentation fault. >>>[Switching to Thread -1208874400 (LWP 2948)] >>>0x080cb1e3 in cRecording (this=0x8409ba8, >>> FileName=0x840efc4 >>>"Chaos_City/1x01_-_001_-_New_York_New_York_(Pilot)/2007-01-01.00.12.1 >>>0.99.re c") at recording.c:506 >>>506 data[1] = (char *)realloc(data[1], len + 1 >>>+ strlen(data[2]) + 1); Current language: auto; currently c++ >> >> >>Try with this patch: >> >>--- vdr-1.3.32/recording.c 2005-09-11 19:02:51.000000000 +0200 >>+++ vdr-1.3.32/recording.c 2005-09-16 10:38:03.000000000 +0200 >>@@ -496,7 +496,7 @@ >> data[2] = data[1]; >> data[1] = NULL; >> } >>- else if (line == 2) { >>+ else if (line == 2 && data[1] && data[2]) { >> // if line 1 is too long, it can't be the short text, >> // so assume the short text is missing and concatenate >> // line 1 and line 2 to be the long text: >> >>Tom > > > Thanks Tom but i converted all summary.vdr's to info.vdr (was a little bit > difficult with nested directories) and now it starts cleanly, i thought that > it is not necessary to convert the old ones (it worked fine with 1.3.28) What did the summary.vdr looks like? I reproduced your crash with a summary like that: Titel <blank line> Short text longer than 80 characters ... <blank line> My first patch solved the crash but didn't show any text. Here is now the patch that should solve all problems with disfigured summary files. --- vdr-1.3.32/recording.c 2005-09-11 19:02:51.000000000 +0200 +++ vdr-1.3.32/recording.c 2005-09-16 13:52:34.000000000 +0200 @@ -492,11 +492,11 @@ line++; } fclose(f); - if (line == 1) { + if (!data[2]) { data[2] = data[1]; data[1] = NULL; } - else if (line == 2) { + else if (data[1] && data[2]) { // if line 1 is too long, it can't be the short text, // so assume the short text is missing and concatenate // line 1 and line 2 to be the long text: Tom