Bryan English

Is there a reason why _all_ errors shouldn't be thrown up to the publisher? @Qard from what I can tell, passing exceptions to `'uncaughtException'` was done as a response to...

@targos That would require protection on the publisher side from `abort` being called more than once. That's certainly possible, but means more logic on the publisher side, which I think...

But with `try/catch` the publisher can opt to do nothing about it, and instead defer to calling code. This should simplify code on the publishing side to effectively just the...

@Qard For sure. In cases where simply bubbling an error up isn't an option, publisher code would need to be wrapped and handled accordingly. That's true with either "throw everything"...

Upon further investigation, it looks like this doesn't quite work yet, as I hadn't touched `Datastore.prototype.getCandidates`, so the compound indexes aren't yet being used. I've now got that working locally....

It should be good to go now :+1:

The failing test ensures that IITM preserves a behaviour of Node.js core, when IITM's loader hooks are enabled. I believe it _does_, in fact, still preserve the behaviour of Node.js...

I hadn't seen the comment from @merceyz prior to bisecting, but I did arrive at the same PR, nodejs/node#49869. This PR seems to have _introduced_ the ability to import extensionless...

> When the package fails frequently, ping the module author. If this is going to block releases, it may be worth collecting alternative contact information, since GitHub notifications can be...

Landed in https://github.com/bengl/sbffi/commit/608aea568bd7674d8acdf988ea66ea2d6ddcb1f0. Sorry it took almost 2 years. :sweat_smile: