In the java app I'm testing there are lists whose list-items change colour to reflect context. I have to track and verify these changes in my testcases.
The methods for JavaJFCListBox are fairly limited and so I printed out the properties using the GetPropertyList method:
Not much help there.
How do you verify the colour of list items?
JavaJFCTextField has properties
although curiously no $TextColor.
More disturbingly, I've been asked to track syntax highlighting in a JEditorPane which Silktest recognizes as a JavaJFCTextField.
Is Silktest selectively colour blind?
Not my thought at all actually, although I definitely considered that of Segue (NOT as a friendly jab at the time, either) when they first implemented HTML and Java support.
Segue's support was and is expedient; they cover what most likely will be used. Under previous systems the use of custom classes was pretty rare. Now it's virtually a requirement for any Silk programmer. Unfortunately this aspect can be a rather difficult and complex work item which serves to compound the already steep learning curve and increases the potential for failure.
You're now on the bleeding edge where you must work directly with engineering to convince/cajole them into supplying you with information and possibly even custom commands that'll support your automation efforts. It'll keep you challenged.
I once had a similar problem and had a rather inelegant solution that proved to be fairly robust.
I used the SilkTest method GetBitmapCRC() on tiny (6x6 pixel) rectangles in the corner of the areas that I wanted to color test.
Note that the value returned by GetBitmapCRC depends on the color depth setting of your video card, so it is best if you COMPARE colors appearing in different areas or different times to see if the CRCs match.
This will tell you when the color changes.
[This message has been edited by RobKapteyn (edited 07-15-2002).]
John, what were your experiences with 1x1 bmps?
Did Bitmap_compare work well?
I'm trying to compare 1x1 bmps since hours and
hours, but Silk allways returns hits, regardless
of whitch color the baseline-bmp is. I'm getting