[vlmc-devel] [PATCH] TracksRuler: Use the optimized update()

[Ludovic Fauvet]

On Wed, Jun 29, 2016, at 14:12, Yikai Lu wrote:
> I have no problem with this. Could someone tell me how to reproduce the
> bug?
> 
> ---
>  src/Gui/timeline/TracksRuler.cpp | 9 ++-------
>  1 file changed, 2 insertions(+), 7 deletions(-)
> 
> diff --git a/src/Gui/timeline/TracksRuler.cpp
> b/src/Gui/timeline/TracksRuler.cpp
> index 3ed010a..216f4ed 100644
> --- a/src/Gui/timeline/TracksRuler.cpp
> +++ b/src/Gui/timeline/TracksRuler.cpp
> @@ -102,13 +102,8 @@ TracksRuler::setDuration( int duration )
>      int oldDuration = m_duration;
>      m_duration = duration;
>  
> -    Q_UNUSED( oldDuration );
> -    //FIXME The optimized update() version cause wrong values to be
> shown in
> -    //the ruler. I don't understand what's happening here.
> -
> -    /*update( ( int ) ( qMin( oldDuration, m_duration ) * m_factor - 1 -
> m_offset ), 0,
> -            ( int ) ( qAbs( oldDuration - m_duration ) * m_factor + 2 ),
> height() );*/
> -    update();
> +    update( ( int ) ( qMin( oldDuration, m_duration ) * m_factor - 1 -
> m_offset ), 0,
> +            ( int ) ( qAbs( oldDuration - m_duration ) * m_factor + 2 ),
> height() );
>  }
>  
>  void
> -- 
> 1.9.1

After a quick check to the file history (from 7 year ago!), Hugo seems
right. My best guess is that there was a regression in the Qt version we
were using at  the time causing artefacts because the updated region was
wrong. If it works now, you can safely restore the code and push.

Best,
-- 
Ludovic Fauvet
www.videolan.org