[vlc-devel] commit: macosx: cleanup (Jean-Paul Saman )

[Jean-Paul Saman]

Pierre d'Herbemont wrote:
> 2009/3/16 git version control <git at videolan.org>:
>> vlc | branch: master | Jean-Paul Saman <jpsaman at videolan.org> | Tue Mar 10 10:51:36 2009 +0100| [98c893ec7f44e8962a6bfd3e41cc8be6c41ec468] | committer: Jean-Paul Saman
>>
>> macosx: cleanup
>>
>>> http://git.videolan.org/gitweb.cgi/vlc.git/?a=commit;h=98c893ec7f44e8962a6bfd3e41cc8be6c41ec468
>> ---
>>
>>  modules/gui/macosx/intf.m         |   26 +++++++++++++++-----------
>>  modules/gui/macosx/playlistinfo.m |   11 ++++++-----
>>  2 files changed, 21 insertions(+), 16 deletions(-)
>>
>> diff --git a/modules/gui/macosx/intf.m b/modules/gui/macosx/intf.m
>> index 8547121..318b94f 100644
>> --- a/modules/gui/macosx/intf.m
>> +++ b/modules/gui/macosx/intf.m
>> @@ -79,9 +79,7 @@ int OpenIntf ( vlc_object_t *p_this )
>>
>>     p_intf->p_sys = malloc( sizeof( intf_sys_t ) );
>>     if( p_intf->p_sys == NULL )
>> -    {
>> -        return( 1 );
>> -    }
>> +        return VLC_ENOMEM;
>>
>>     memset( p_intf->p_sys, 0, sizeof( *p_intf->p_sys ) );
>>
>> @@ -92,7 +90,7 @@ int OpenIntf ( vlc_object_t *p_this )
>>     p_intf->pf_run = Run;
>>     p_intf->b_should_run_on_first_thread = true;
>>
>> -    return( 0 );
>> +    return VLC_SUCCESS;
>>  }
>>
>>  /*****************************************************************************
>> @@ -184,10 +182,13 @@ static int PlaylistChanged( vlc_object_t *p_this, const char *psz_variable,
>>                      vlc_value_t old_val, vlc_value_t new_val, void *param )
>>  {
>>     intf_thread_t * p_intf = VLCIntf;
>> -    p_intf->p_sys->b_intf_update = true;
>> -    p_intf->p_sys->b_playlist_update = true;
>> -    p_intf->p_sys->b_playmode_update = true;
>> -    p_intf->p_sys->b_current_title_update = true;
>> +    if( p_intf && p_intf->p_sys )
>> +    {
>> +        p_intf->p_sys->b_intf_update = true;
>> +        p_intf->p_sys->b_playlist_update = true;
>> +        p_intf->p_sys->b_playmode_update = true;
>> +        p_intf->p_sys->b_current_title_update = true;
>> +    }
> 
> This should not be needed. Or else we are in a very bad situation. I
> would prefer a crash or an assert then :) What's the rationale behind
> this?

hmm, shouldn't we do assert() on p_intf and p_intf->p_sys then?