Eric Bidelman

Yea, apologies for this. It works in Safari and Chrome, but not FF and Edge. FF is tricky b/c they now support ES modules but not dynamic `import()`. So the...

So it appears Edge doesn't support URLSearchParams on URL. [This page](https://developer.mozilla.org/en-US/docs/Web/API/URL) suggests it does. [This page](https://developer.mozilla.org/en-US/docs/Web/API/URL/searchParams) suggests it doesn't. :\

For example, it's possible to repost https://paul.kinlan.me/the-web-is-my-api/

If you're interested in a PR. Here are the places you'd need to change: https://github.com/ebidel/devwebfeed/search?utf8=%E2%9C%93&q=chromiumdev&type=

Good idea. Better to play it safe. If the crx makes a fetch without cookies and gets a 200 back, that may work. Although if someone is on a corp...

Are you seeing this mainly from content from developers.google.com/web? Many articles there are bulk published at once, but are often written several days/weeks in advance. When this site picks them...

Generally it keeps things more ephemeral starting a new instance. Others have found the same: https://docs.browserless.io/blog/2018/06/04/puppeteer-best-practices.html Puppeteer just introduce browser contexts for cases like this. The feature basically open Chrome...

I think a `waitFor=x` param would be good. That maps well to pptr's API's which have a similar param that wait to resolve `page.goto` (and similar methods) by x milliseconds.

Launching a new instance is more costly, but it keeps things cleaner. Eventually, I've seen Chrome run out of memory and crash the server. In other cases you really want...

Can't assign issues, but I'll take this.