Comment 6 for bug 1589473

Yeah, this is where knowledge of the code base is useful. Knowing all the points that call the code where your patch is, when, and why, you can often intuit impacts. Of course not always, and this is why so much focus in commercial software is around testing.

My intuited regression test for this would be:
1. Create a layout containing multiple tabs where all tabs have multiple splits.
2. Pick a random terminal on each tab, and make a note of which one was active in each tab.
3. Save the layout as say "ActiveTermTabTest"
4. Using another terminal program (i.e. gnome-terminal) start with "terminator -u -l ActiveTermTabTest"
5. Check which terminal is active in each tab and compare to which was active when you saved.

I think you'll find that the inactive tabs are not all remembering the last active tab correctly (random luck says some might be).