I am not a language specification lawyer. If I hit a difference in behavior, I don't open up the spec to try and figure out who is right and wrong. It's entirely possible that Safari is in the right. But when Chrome/Firefox/Latest webkit agree, and Safari doesn't, I'm fairly confident it's Safari's fault at that point.
But even if it was Safari's fault and I could test that on Linux/Android, that would be fine. The biggest annoyance is that you can't possibly know if what you're doing would work without buying Apple hardware.
This is weird to me. Say if I’m working with Swift, JavaScript, and Java. And only Swift built-in mergesort is the right implementation (say it’s the common case where it correctly implements the mid logic not shown in textbooks). That’s not Swift’s fault, that’s Javascript & Java’s fault for implementing a bad Mergesort.
Developing on Safari it feels like the opposite.
A few years ago I’d have agreed somewhat, but Safari has accelerated considerably and improved wrt standards. In fact better than Chrome as of late, if you follow the Interop project, and it feels as much.