01:27
<Domenic>
So many leftover popover issues. And review already took quite a while. Sigh.
There are some Google folks brainstorming on how we can do better in cases like this. I think there are ways we can make the process less painful (a major part of their concern) but ultimately I also think for large features like this you sometimes just have to accept you'll discover spec bugs in the second implementation. I think that's happened with many other large features too, e.g. service workers.
01:31
<annevk>
This isn't anything like introducing a new environment though. Or at least, I don't think it is.
01:32
<annevk>
I think automated formatting would go a long way toward making the process less painful.
01:36
<Domenic>
I agree it would be less painful, but many of the issues you all are finding aren't formatting issues
01:36
<Domenic>
Some of them are theoretically tooling-detectable (e.g. mismatched algorithm declarations vs. call sites) but several are just solid consistency finds, better APIs, ...
13:52
<annevk>
That's true, but at least I notice I find it hard to spot those when there's formatting errors. And I worry that once I've put a bunch of effort into pointing out various errors, I might no longer be as vigilant the next time around, but that might vary depending on how much time has passed.
15:39
<Dominic Farolino>
Heh, any chance specfmt could help there? (specifically with the formatting)
16:17
<annevk>
Dominic Farolino: if it gets to a stage where it's part of CI somehow, yes