Marco Kremer wrote:

Hi Klaus,

I tried 1.3.32 (plain, only remote-Plugin) and it always crashed with a "Speicherzugriffsfehler". I tried to debug it (i'm no programmer at all) and it fails at

[Switching to Thread -1208472992 (LWP 18423)] 0x080cb1d6 in cRecording (this=0x840f170, FileName=0x840f114 "Chaos_City/1x01_-_001_-_New_York_New_York_(Pilot)/2007-01-01.00.12.10.99.re c") at recording.c:506 506 data[1] = (char *)realloc(data[1], len + 1 + strlen(data[2]) + 1);

I use future dates to reorder my recordings -> this doesn't happen with vdr-1.3.28.

Could this be changed ?

See the patch posted here by Achim Tuffentsammer a.tuffentsammer@web.de yesterday:

--- recording-org.c 2005-09-15 19:25:49.000000000 +0200 +++ recording.c 2005-09-15 19:26:15.000000000 +0200 @@ -500,7 +500,7 @@ // 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: - int len = strlen(data[1]); + int len = ( data[1] ) ? strlen(data[1]) : 0; if (len > 80) { data[1] = (char *)realloc(data[1], len + 1 + strlen(data[2]) + 1); strcat(data[1], "\n");

Klaus

hgm.bg wrote:

vdr-bounces@linuxtv.org 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

hgm.bg schrieb:

vdr-bounces@linuxtv.org wrote:

...

hgm.bg wrote:

...

vdr-bounces@linuxtv.org 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