Ensure first-class support for transpiled & ts code in transports

[castarco]

This PR upgrades the thread-stream dependency, applies a couple of changes on how transports are created, and adds new tests related to typescript & transpiled code.

• Changes to improve support for typescript in transports (both transpiled and untranspiled)
• Tests checking support for untranspiled code
• Tests checking support for transpiled code

Fixes #1243 .

[mcollina]

Could you add some docs?

[castarco]

Could you add some docs?

Sure. I'll check the broken builds as well.

[castarco]

@mcollina I was checking the problems on the big.test.js file, and saw that you added an await sleep(1000) to account for file system synchronisation delays (which seem to be worse on Mac than on Linux).

I'm working on Linux, so I cannot directly check if that will work or not on Mac, but I wonder if relying on the sync command would work, without having to impose arbitrary delays.

Writing this instead:

await execa(`sync`)

or

await execa(`sync`, ['-d', destination])

[mcollina]

That'd be awesome, feel free to add it. We can still rely on the timer if that's not available.

Requested changes have been addressed.

[jsumners]

I have two concerns with this PR:

1. I don't know who is going to maintain this new code as neither @mcollina or myself are TypeScript developers.
2. There are many TODOs littered throughout this new code. I think they should be resolved before this is actually ready to be merged. But I have removed my "request for changes", regardless.

[castarco]

@jsumners I sympathise with your two points, they're valid concerns.

1. Regarding the first point, I see myself maintaining it, I'm a "TypeScript developer" and also a user of Pinojs for my own personal projects. Also, although not formally under Matteo's command, I work at NearForm too. NearForm usually allocates some resources to maintain certain packages of interest.
2. Regarding the second one, while I agree, I didn't want to fall into scope-creeping; and it's already useful for some TS developers as it is now. The PR is also quite big, and some of the highlighted problems in the TODO comments deserve experiments and research of their own. Maybe I should remove the comments and write proper issues instead.

What really bothers me about this PR... is how I still didn't manage to fix the flaky behaviour of the big.test.js test file 😞 .

[mcollina]

@kibertoad could you review this? This might a bit fall on you to help with!

[simoneb]

I know that @castarco cannot work on this any longer, is anybody else looking into this? Any input needed?

[mcollina]

lgtm

[kibertoad]

will take a look shortly, can resolve conflicts as well

[mcollina]

Thanks!

[castarco]

@kibertoad I rebased to resolve the conflicts.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.