Rick Richardson

> This pointer is valid for the whole allocation. But NonNull::as_mut returns a &mut u8, which is a reference to a single byte. Ah. Right.. that makes complete sense. "transmuting"...

I think the safest/easiest approach would be to add a `try_resize` method to Bytes/Mut and to the VTable. If the new len is less than the capacity (which is currently...

To clarify: If someone says `X.unsplit(Y)`, what they're really saying is "split_off" X's original buffer, starting at X.ptr with a len of X.len + Y.len. So clone X, then resize...

@Matthias247 wrote: > The thing which we merge into should be taken by &mut self since it will always stay intact but change its properties. The thing which gets merged...

> However, we'll still need to consult the VTable to ensure that the capacity is enough to support the new len. So the VTable will need a new method: get_capacity(data)...

I've been researching a bunch of relevant tickets and solutions for the last couple days, I'm going to attempt to summarize all of the issues/PRs so that we can discuss...

pinging @ipoupaille for comment, since they made a PR (#558) for VTable exposure as well.

> Hey, any news regarding this? I guess its stalled, any plans on a rewrite or some future plans regarding async? I ended up implementing the async rewrite, but it...

Would this include things like `xsi:type` as well? In order to include it as an attribute in an element, I used `#[serde(rename = "@xsi:type")]` which causes it to serialize correctly,...

Looking more closely at the issue and the PR. I think my issue is unrelated. I'll add a ticket.