RFR: 8276058: Some swing test fails on specific CI macos system [v2]

[Phil Race]

On Fri, 29 Oct 2021 04:43:10 GMT, Prasanta Sadhukhan <psadhukhan at openjdk.org> wrote:

> > Regardless, there's value in the test updates, although my existing comments on the specifics there still stand.
> 
> Should I pursue with the test updates with this PR as it seems it can be addressed by product change? 

Yes.

I am not sure what specifics needs to be addressed...Are the below ones the one?

Whatever I wrote !
> 
> > And is an offset of 10 enough ? Its a bit arbitrary and cursors could be a larger shape or different orientation
> 
> SInce I am going in opposite direction of cursor by using (-10,-10) I guess it should be ok since normally cursor as shown in the screencapture, even if enlarged, willnot encroach on (-10, -10) area unless it's a inverted cursor wihch I am not sure exist or not.

So your assumption is the cursor is an arrow. It could be an X .. in which case your fix won't work of ots
anywhere near a 45 degree slope.
There's no 100% answer but a larger offset is better and parametrised like
// move away from cursor
OFFSET-X = -10
OFFSET_Y = -20
or  adding code that moves the cursor to window (0,0) before screen captutre .. 
> 
> > Shouldn't this use TESTCLASSES or something like that ?
> 
> We have used the same way of ImageIO.write() in jtreg tests

OK ..doesn't mean its right  please check with jtreg devs etc ..

-------------

PR: https://git.openjdk.java.net/jdk/pull/6140