Re: [Xnee-devel] [Fwd: AW: [sr #103871] Verifying GUI-Testresults with xnee using the Tool "xgrabsc"]

[Henrik Sandklef]

> > -------- Forwarded Message --------
> > From: Kaplick, Dirk <address@hidden>
> > To: Veijo Ryhänen <address@hidden>, Henrik Sandklef
> > <address@hidden>
> > Cc: Henrik Sandklef <address@hidden>
> > Subject: AW: [sr #103871] Verifying GUI-Testresults with xnee using
> > the Tool "xgrabsc"
> > Date: Thu, 19 Jan 2006 06:34:54 +0100

> >  
> > I think, your are right with your idea, it could be enough saving
> > the checksum of the picturefiles to the xnee-session-file.

Checksums are good if the you expect the screenshots to be identical.

 
> > If your are saving the screenshots, so you can verify a defect of
> > your grafical interface by comparing the reference-screenshot
> > against
> > the actual taken screenshot while replaying the session.
> > So you can see the differences between the pictures and verify, what
> > your grafical interface has done.

I think you should (during record/replay of session):

 * Take a screenshot (using exec functionality)


And after the session:

 1) Calculate the checksum

 2) If checksums differ, compare the pics (using ImageMagick compare +
sed) and present 

   2.1) a picture with the diffs

   2.2) some kind of percentage of how big the diff is
 
 

 
> > The reference-screenshot is my requirement for testing.
> > Ok. Picturefiles can be very large on filesystem, so there could be
> > a problem after any time.
> > Hm. 
> > We could compress them while saving and uncompress them, if we need
> > to see them.

or, Get bigger disks or throw old stuff ;)

>