Using a workaround to be able to check the contents of a new tab, I get unexpected results. If I play the test case, the following line fails:
storeEval | javascript{"selenium.browserbot.getCurrentWindow().open('', 'my_window')"} | my_window
with the following error message:
[error] Threw an exception: result is null
The following alternative statements show the exact same symptoms:
getEval | window.open("", "my_window")
getEval | selenium.browserbot.openWindow("", "my_window")
Now, if I double-click it, it doesn't fail, but it opens up a new window, which I think the original solution didn't intend (the next step is to click on a link in the parent window, after all). At this point I can click on the parent window, manually run the next step to launch the link, manually change back to the child window (selectWindow | my_window says it can't find that window), and manually run each of the following steps to check the contents of the new window.
How do I proceed from here to make the test properly automatic?
Update: There is a built-in action for opening windows, but it also fails. Maybe this verbatim log output can shed some light on what's going on:
[info] Executing: |store | license_window | windowName |
...
[info] Executing: |openWindow | | ${windowName} |
[error] Unexpected Exception: fileName -> chrome://selenium-ide/content/recorder.js, lineNumber -> 74
The relevant lines from recorder.js:
66: window.open = function(url, windowName, windowFeatures, replaceFlag) {
...
74: if (result.wrappedJSObject) {
75: result = result.wrappedJSObject;
76: }
