Just curious what everyone else is doing. I typically never bothered but recently I've been thinking I might as well use it. The only downside I see is that it's annoying to maintain the list of dependancies to extract in your webpack.mix.js file.
It would only be useful if you had a lot of “sections” to your app (say the front-end, back-end, etc) where you wanted to use say, Vue in both of them. You could extract vue which would theoretically reduce the size of your frontend.js and backend.js files (because Vue would be extracted, instead of being included in both).
I just learned that you can leave extract() empty, and it will automatically extract ALL of your vendor libraries. There's no need to manually maintain the list of vendors to extract. Awesome!
@Swaz But do you have lots of individual entry points using the same libraries? Because if not, all you’ve done is split your JS over three files that require three separate HTTP requests to load now, instead of just one.
@martinbean What do you mean by individual entry points?
My vendor libraries hardly ever change, but my appjs changes daily. I like that the vendor file can stay cached and the client only needs to re-download the file that has actually changed. Am I missing something here?
I thought separate requests was no longer an issue with Http2 multiplexing.