Fixing intermittent failures due to js-quic stream errors

[tegefaulkes]

Description

This PR addressed #722

The main problem was during certain circumstances an error that should've been thrown through a stream was escaping the call stack and crashing the program. This was due to improper usage of throw within the stream handlers. This has been resolved within js-quic.

Note that the branch name doesn't describe the problem scope very well. It was created while I was still exploring what the actual issue was. While mainnet connections in tests could still happen, they shouldn't be causing this issue regardless. No amount of interacting with random network traffic should crash the program.

#461 was removed from this PR, it turns out to not actually be a requirement.

Issues Fixed

• Fixes Fix random test failures caused by mainnet connections #722
• Ref ENG-317

Tasks

• 1. find the problem causing intermittent errors.
• 2. Apply the fix to js-quic and update the dependency version pulling in the fix
• 3. Applied any other minor fixes

Final checklist

• Domain specific tests
• Full tests
• Updated inline-comment documentation
• Lint fixed
• Squash and rebased
• Sanity check the final build

[linear]

ENG-317 Fix random test failures caused by `mainnet` connections

[CMCDragonkai]

What are the implications of doing this:

• MatrixAI/js-quic@0c9bd4f
• MatrixAI/js-quic@1f5de39

Given that throw reason; and related were supposed to be "proper".

[tegefaulkes]

Everything functions the same way, we just remove the chance of that throw inside the stream handler bubbling up and crashing. It's not a good idea to throw the errors inside the stream like that, I think I pointed it out at the time as well. So I've removed the throw in the cases where the errors are only expected to be thrown through the stream.

[CMCDragonkai]

If it still happens on CI, are you sure the fixes actually fix this? If not, perhaps it's better to leave it as is, and find the actual deterministic root cause.