Waffle Maybe

@Avimitin this is mostly irrelevant. We use `serde` via macros, which do not have access to the type information and can't change `Deserialize` implementation depending on if type implements `Default`...

I don't think anyone is working on this, you may go ahead

Sadly, I have no idea why this might be happening, if you can minimize the example more, that would be helpful.

@syrtcevvi polling timeout must be smaller than http client's, which is the case (polling_default = 10s, http = 17s), so I don't think this is the problem...

Just a note: I'm overwhelmed with work, so I'm unlikely to have time to debug this issue anytime soon.

@GoldsteinE unless I misunderstand what `MetaSized`/`MetaAligned` are, isn't `Foo` not just `MetaAligned`, but `Aligned` (i.e. the alignment is statically known even without metadata, similarly to slices). Would also be nice...

The concern about targets with data and function pointers being different is interesting. Note that I don't think difference in address spaces is significant — whatever the "blessed" way will...

I've previously wanted to add support for `.*`, `.&`, `.&mut` so I'll chime in for a second (and then chime out, since I don't have energy for a prolonged discussion)....

How much do you want to allow users to depend on this? I.e. is this a _guarantee_, or just a _hint_? Asking because there might potentially be backends that can't/don't...

I think the consensus in the T-lang meeting today in response to @RalfJung's [concern](https://github.com/rust-lang/rfcs/pull/3484#discussion_r1548068355) (hm, re-reading the comment now I'm not sure it's actually a concern, it actually looks more...