fix(nextjs): Add edge polyfills for nextjs-13 in dev mode

[andreiborza]

This issue is only observed for Nextjs 13 and specifically for apps that don't use the app router. Similarly to what we do in vercel-edge, we polyfill performance. This is only done for dev mode as it's not an issue otherwise.

Note: I tried to reproduce this via our existing e2e test apps but they're all using the app router so it's not manifesting there.

Closes: #17343

[RulaKhaled]

Nice work, just quick comments/questions

[RulaKhaled]

Nit: Can we use a more precise type? or type this with unknown instead of any please?

[andreiborza]

These are just test fixtures, I don't think it's necessary to be more precise here.

[RulaKhaled]

This will run for all next versions with dev mode correct? is this intended?

[andreiborza]

Right, I'll have a look if I can scope this to 13 only.

[andreiborza]

Updated to only take 13 into account.

[RulaKhaled]

Maybe do:

 const timeOrigin = Date.now();
  return {
    timeOrigin,
    now: () => Date.now() - timeOrigin,
  };

I think anyone doing timing assertions in dev might see different values

[andreiborza]

Right, updated.

[github-actions]

node-overhead report 🧳

Note: This is a synthetic benchmark with a minimal express app and does not necessarily reflect the real-world performance impact in an application.

Scenario	Requests/s	% of Baseline	Prev. Requests/s	Change %
GET Baseline	8,919	-	9,388	-5%
GET With Sentry	1,251	14%	1,300	-4%
GET With Sentry (error only)	6,021	68%	6,248	-4%
POST Baseline	1,205	-	1,215	-1%
POST With Sentry	480	40%	535	-10%
POST With Sentry (error only)	1,061	88%	1,072	-1%
MYSQL Baseline	3,357	-	3,421	-2%
MYSQL With Sentry	428	13%	431	-1%
MYSQL With Sentry (error only)	2,712	81%	2,786	-3%

View base workflow run

[RulaKhaled]

thanks!