Vincent Weevers
Vincent Weevers
That's expected because `buffer@6` dropped support of IE11: https://github.com/feross/buffer/commit/840160f352ff0d8e88d4ac308a1640fd278f50fc
I find it difficult to write guidelines for this because it depends on several factors (e.g. was everything planned out before the PR started or was the PR opened early...
> possibly using a reference link like the version headings I tend to avoid this so that content can be copied in full to e.g. a GitHub Release.
> Does it actually do what I hope it is doing? I had the same question and the answer is yes: https://github.com/delfrrr/npm-consider/blob/8c6b7175bf957513abe494bb694bb2110cac73d7/lib/install.js#L163-L168
> mentioned not adding features "like BrowserStack support". Is this still the current opinion? Short answer: no. Limiting the scope at first, allowed us to refactor airtap into a modular...
Airtap 4 [prefers `localhost` over `airtap.local`](https://github.com/airtap/browser-manifest#wants) when possible ([per browser](https://github.com/airtap/airtap/blob/21a0dd9aa25623e1fe82e2c7abf9491f31938819/lib/proxy-server.js#L25)). Unfortunately I noticed that Chrome and Firefox now also need `airtap.local` to correctly route traffic through the Sauce Connect tunnel,...
Another option is to use engine.io's http polling transport instead of websockets. The logic would be: if you enable SSL Certificate Bumping (through a new airtap option, perhaps combined with...
A browserify transform or plugin, like babel or esmify, should do the trick. Unless you want to test real ESM in browsers, then I'm not sure.
I haven't used real ESM before, some guidance is welcome. First thoughts: I suppose airtap would have to serve the user's input files as-is without bundling them, acting as a...