[Hydration] Bail out from hydrating suspense boundary if component suspends

[salazarm]

Summary

This fixes the issue pointed out by @gaearon in this PR: https://github.com/facebook/react/pull/21630/files.

The issue is that when a component in a boundary suspends while hydrating we attempt to hydrate its sibling. However we don't know how many nodes the suspended component would have rendered (because we don't add comment nodes to signal this since it would bloat the HTML) so we can't tell where the sibling should hydrate. Rather than attempt to hydrate it the solution in this PR bails on hydrating the boundary from within the current parent fiber. It will continue with the sibling to the parent fiber which is fine since the hydration pointer will be correct in this case.

This means we won't call render on any siblings or children but I think it should be fine since we don't reuse the work anyways. Since we don't call render we also don't "warm up" the codepath but that might be okay if there's other CPU work to do anyways we might actually see an improvement from allowing other higher priority work to happen.

How did you test this change?

Jest

[sizebot]

Comparing: cadf94d...bf891da

Critical size changes

Includes critical production bundles, as well as any change greater than 2%:

Name	+/-	Base	Current	+/- gzip	Base gzip	Current gzip
oss-stable/react-dom/cjs/react-dom.production.min.js	+0.02%	130.05 kB	130.08 kB	+0.02%	41.34 kB	41.35 kB
oss-experimental/react-dom/cjs/react-dom.production.min.js	+0.02%	132.88 kB	132.91 kB	=	42.31 kB	42.28 kB
facebook-www/ReactDOM-prod.classic.js	+0.01%	413.66 kB	413.70 kB	=	76.44 kB	76.40 kB
facebook-www/ReactDOM-prod.modern.js	+0.01%	402.24 kB	402.29 kB	=	74.71 kB	74.71 kB
facebook-www/ReactDOMForked-prod.classic.js	+0.01%	413.66 kB	413.70 kB	=	76.45 kB	76.41 kB

Significant size changes

Includes any change greater than 0.2%:

(No significant changes)

Generated by 🚫 dangerJS against bf891da