Issues with unsupported toolkit functionality

Roman Kennke roman at kennke.org
Fri Jun 1 06:40:30 PDT 2012


Am Donnerstag, den 31.05.2012, 14:06 +0200 schrieb Roman Kennke:
> Hi Martin,
> 
> > I'm currently porting our FEST test suite from using Xvnc to Cacio.
> > In general it seems to work quite well but there are a couple of
> > API-Methods of CTCToolkit which are not (really)
> >  implemented yet, although I think it should be quite simple to do so:
> > 
> > * currently the getScreenResolution() method returns 0, (which is
> > pretty much the worst number in this position)
> >   a constant 72 or 96 would help a lot.
> >   A possibility for dynamic configuration in tests would be great.
> >   We have a lot of Screenshot/Save Image actions which use this number
> > for a division and currently crash.
> 
> Right. This seems to be fairly easy to implement.
> 
> > * getSystemClipboard() currently returns null. Although I understand
> > that a fully-fledged implementation would require
> >   some intelligence but I believe in tests it would be sufficient to
> > create a singleton instance of Clipboard in the toolkit, that can
> >   be filled and read by java test and client code.
> 
> Yeah, I think that should be good enough especially for testing. For
> real platform ports, this would need to be integrated with the platform
> clipboard, but this is not part of core Cacio anyway.
> 
> 
> > * We are supporting only Oracle JDK. This works in general quite well
> > with Cacio, but there is an Issue in the following method:
> > 
> > sun.awt.DefaultMouseInfoPeer.fillPointWithCoords(Ljava/awt/Point;)Ijava.lang.UnsatisfiedLinkError:
> > sun.awt.DefaultMouseInfoPeer.fillPointWithCoords(Ljava/awt/Point;)I
> > 	at sun.awt.DefaultMouseInfoPeer.fillPointWithCoords(Native Method)
> 
> Oh yeah, we don't provide an implementation of that one yet, and it
> simply falls through to the JDK SunToolkit implementation, which of
> course does not work. Should be fairly easy to implement though.
> 
> 
> > These are my observations so far, anyway: Great Work, will save us a
> > lot of trouble in the future.
> 
> Thanks a lot! I will fix the above mentioned issues as soon as I find
> some time and then probably push out a release.

I just committed fixes for all 3 issues. Would you be able to check out
Cacio from Mercurial and try if it fixes your problems?

Cheers,
Roman





More information about the caciocavallo-dev mailing list