Ian Jackson

> Sorry for the inconvenience! Thanks for the helpful reply! It makes a lot of difference to know I'm not going mad, and this is actually a real problem rather...

Well, thanks for the curl trick. It does seem to work. (Shell fragment below the cut for your, err, edification.) I observe, though, that `raw.githubusercontent.com` is a CDN. It seems...

> doing registry updates until the published crate is found in the registry, with a sleep inbetween. I have a shell script to do precisely this. > We could make...

I have not seen this happen to me in a very long time. I think that if this compiler bug is still present, it now does not come up when...

> Maybe a bit late, but I would share my experience. `fehler` is better work with `derive_more` `derive_more` is nice but I recommend you check out `thiserror`. It lets you...

I love fehler and I really really want it for my IEFEs and closures. Is the [expr -> stmt workaround](https://github.com/rust-lang/rust/issues/15701#issuecomment-449900562) not applicable for `#[throws]`? Here's some code from my current...

Would it be easy to spot this case and at least warn about it? [throws] implies that the Result must be used. So adding [must_use] is an attempt to do...

(Sorry for the delay replying. I wrote this a couple of days ago and I could swear I hit the "comment" button.) > The use-case makes sense to me. However,...

(oops, pasted this comment into #253 by mistake, rather than here. I will delete it there.) > I'm going to take a closer look at the usage of `sub_builder` in...

OK, thanks. (Your suggestion to use `try_` setters wouldn't work for us because we need the *builder* to be deserialize, so the builder has to contain sub-builders, not validated structs.)...