snek

if anyone wants to fix this before i find the motivation to get around to it, the problem is here: https://github.com/rp-rs/pio-rs/blob/0a6d22f083633b0ee8cd520714d917bb68e0c359/pio-parser/src/pio.lalrpop#L35 each `DefineDirective` should have a `NEWLINE+` after it.

Sorry this slipped off my radar. I have no idea if this fixes the problem. I'm not sure that the changes except for line 515 need to be there.

i think it would probably be better to change this to store `process.stdout`, `process.stderr`, etc., internally for stdio instead of using their FDs.

oh very nice! we're actually integrating this module into node.js itself, so you might want to wait until we finish that. check out https://github.com/nodejs/node/pull/27850 for tracking. I'm planning to merge...

it's behind a flag, and also has "unstable" in the name. from a development perspective we do this quite often, and releases of node will just have the experimental feature...

If someone on windows wants to fix this I will gladly accept a PR. Alternatively, a better solution might be to change fds 0, 1, and 2 to be streams,...

fwiw i can't reproduce any context-related crashing/errors with ModuleWrap. as i know it the only thing that should fail is dynamic import which we currently lock to the main context....

@smotaal i believe this issue was inadvertently already fixed. in ModuleJob#instantiate the `this` value would be lost in certain cases i don't anyone completely understands resulting in it returning `undefined`...

the renderer should use chromium's loader right? this seems like a really confusing situation.

in reference to that node pr, could work be done at this point to normalise a spec which hopefully also includes sniffing? i'd be happy to help however i can.