Remix.run Logo
nromiun 3 days ago

TIL that you can run frontend Javascript with a package like Deno. I thought you need a proper headless browser for it.

bob1029 3 days ago | parent | next [-]

I was thinking the same walking into this thread. I figured DOM/CSS/HTML would be part of the black box magic, but I suppose from the perspective of JS all of that can be faked appropriately.

skydhash 3 days ago | parent | prev [-]

I think you only need something like `jsdom` to have the core API available. The DOM itself is just a tree structure with special nodes. Most APIs are optional and you can provide stubs if you're targeting a specific websites. It's not POSIX level.

johnisgood 3 days ago | parent [-]

I would like to know more about this. I had some web scrapers in Perl but they no longer work. :(

immibis 2 days ago | parent [-]

The state of the art is to remote-control a real browser now. Defeats all not-a-real-browser checks. You can even click on the cloudflare captchas.