Re: [whatwg/streams] alert underlyingSink when [[strategySize]]() fails (#628)

> If an error occurs in a transformstream that's part of a pipe chain, is the error not supposed to be propagated both backwards and forwards?

I see. So this is specifically a transform stream problem, of how to get the error from the writable side to the readable side. (Although the most natural solution would indeed involve notifying the underlying sink inside the general writable stream mechanisms.)

And you're saying that the other instances of an error not being delivered to `abort()` are not a problem, like https://github.com/whatwg/streams/issues/620, because our underlying sink `write()` inside transform streams is entirely under our control---whereas the queueing strategy is not.

Off the top of my head I can think of two solutions: waiting for finally, and creating some kind of wrapper around writableStrategy. (Are we still planning to have both strategies? :-/.) Maybe there are better ones.

I would feel better if there were other reasons finally was important for transform streams.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/streams/issues/628#issuecomment-268098441