> yeah - I'm curious about this - where did you find out
sun.lwawt.macosx.LWCToolkit would solve it ? is this swtbot being weird ?
SWT bot is using awt under the hood for something - simulating keyboard pressing I think.
Lucia found out about the fix so I don't know how. And she's on maternity leave,
so she won't tell us :)
:)
> note I google this and found we apparently have code that messes
with this in our tests ?!
>
>
http://lists.jboss.org/pipermail/jbosstools-commits/2012-September/041589...
>
> so this is not a new issue?
Yes, it is not a new issue. But I didn't know about the hack in bot ext that is
supposed to deal with this. And our dummy test where I was debugging this doesn't use
bot ext (Jirka just submitted a pull request changing this test to use ReqAwareSuite from
bot ext so that should fix it).
okey - so no need to change parent pom for now ?
> What is the full stacktrace of that error about "headless
environment" ?
Strangely enough when I tried this on my own MacBook the property was set correctly and
the test ran fine.
For AppleJDK the property is: awt.toolkit=apple.awt.CToolkit
For OracleJDK it is: awt.toolkit=sun.lwawt.macosx.LWCToolkit
So there must be something wrong with the jenkins slaves that they set the property
incorrectly for OracleJDK1.7 on Mac.
okey.
But anyway, it seems for now we will try to solve this with the hack
in bot ext (or fixing the slaves if we can determine how it happens).
cool.
The exception when this happens is this:
Error Message
java.awt.AWTException: headless environment
Stacktrace
java.lang.RuntimeException: java.awt.AWTException: headless environment
at
org.eclipse.swtbot.swt.finder.keyboard.AWTKeyboardStrategy.<init>(AWTKeyboardStrategy.java:39)
at
org.eclipse.swtbot.swt.finder.keyboard.KeyboardFactory.getAWTKeyboard(KeyboardFactory.java:103)
weird it thinks you are in headless mode - any chance this is called "too soon"
and AWT is not fully started yet ?
/max