So start, of course, by reading the book! Then practice, practice, practice, practice, and practice some more. Look at other peoples code. If possible, have other people look at your code. Nothing is going to teach you better than making all the same mistakes we all make on the way to perfecting our craft, and learning from them.
You can't. At least I can't. I'm not a spring chicken anymore, and I've got actual work to get done, and a life to live while I'm doing it. There's just too much out there, in my opinion, to be able to master it all. My approach is...
One of the advices I heard is to wait some time to see if a framework can survive the hype cycle and be able to survive in the long run.
You may have heard this from me, or you may have heard it from someone like me. I try to learn enough about what's going on to track the "hype cycle" and see what's going to "stick" and what's going to be yesterday's quaint project or language.
I also keep tab of job descriptions. For example, when AngularJS started appearing in most front-end job descriptions, I knew it was more than just a fad.
What else should one do
Code. And be critical of your code. I look upon it as a craft. A craft is never perfected; there's always more to learn and more to grow. And to me, it's not always a matter of learning new things, but of making sure that the things you already know are being used well.