[RFC] "http-types" breakage / additions / rework

arybczak:

How many of them have loose bounds and are going to require revisions? Who’s going to spend time making these revisions?

I find it quite odd that we’re trying to hold library maintainers hostage because someone else is too lazy to manage their bounds.

No, if you have loose bounds (which is ok), then it’s your responsibility to be above-average responsive when things break. Otherwise use tight bounds.

We can argue whether the change is worth it in general.

To me, what’s more concerning is how I make my own libraries compatible with both APIs, because I don’t want to force a build plan on end users. I suspect that might even require a compat package of some sort.