There is definitely a vicious circle here. For many web sites/apps, the browsers that matter are Chrome on desktop and Chrome and Safari on mobile. There is now a huge gap between the number of people using those and the second tier browsers, which in most contexts are IE, Firefox, Edge and (desktop) Safari. That inevitably leads to development and testing resources concentrating on the 80% with the other 20% often being an afterthought if they are considered at all. But then there are more bugs in the second tier browsers, which just opens the gap wider.
But not testing Safari I can understand since it's only officially on Mac. Sure you can kind use it via Epiphany and Midori but a lot of features, like Service Workers for instance, don't exist and therefore can be tested with a download. For me, even after wasting time trying to build from source, it didn't seem like the feature flags worked.
