CI seems to be broken, btw. Would you be open to a PR, switching from Travis to GitHub Actions?

Thanks so much for this! Limiting the amount of updates makes this easier to review and merge than #315. As for the changes:

• In Engine.IO v3, the io cookie was sent by default. This cookie can be used to enable sticky-session, which is required when you have several servers (doesn't seem to be relevant for airtap);

Correct. Airtap does run multiple servers, but it's one per browser, so a browser only ever connects to one server.

• handlePreflightRequest removed from CORS (was not used in the airtap);
• perMessageDeflate is now disabled by default (shouldn't make much difference for airtap);
• Support for Node.js 8 was dropped (airtap is 10+);
• the syntax of the "wsEngine" option is updated (airtap does not use wsEngine option).

👍

CI seems to be broken, btw. Would you be open to a PR, switching from Travis to GitHub Actions?

That would be most welcome. Thanks!

I do have two remaining questions - same as #315 (comment):

1. Is latest engine.io-client using ES6+ features? Which is why we currently pin to 3.3.2.
2. Does it support the same browsers (versions)?

@vweevers I see classes and default arguments, I take it that means "yes" on the first one?
How can I check supported versions?

I see classes and default arguments, I take it that means "yes" on the first one?

In that case we will (unfortunately) need to transpile the client bundle to ES5 (#276). E.g. by adding a babel transform to this browserify instance.

How can I check supported versions?

If not documented by engine-io (?) then we can run tests in Sauce Labs. Which would be good anyway. To do so, clone airtap/sauce, run npm i airtap/airtap#kibertoad:chore/deps-2021-09-05 --save-dev and open a draft PR on that repo. Which won't immediately run tests for lack of access to secrets, but I will trigger them. This will cover a whole bunch of browsers.

I see that engine-io-client has a dist folder, perhaps it's already transpiled?

Hey folks is this stalled? Does anyone need help getting this landed?

Feel free to pick this up, the state is as written above (and in addition this PR needs a rebase). I'm not going to work on this in the foreseeable future, because it's only about removing a false positive alert. Which is not a priority for me especially if it results in breaking changes that necessitate adding a transpiler. Do you want to send a PR to tackle that?