main(): Fix deprecated causes() w/iter_causes()

[userzimmermann]

Fixes #104

Hi :) Since I'm completely new to Rust, it took me a while to figure this out... But that's the best way to learn! :D

Nevertheless I'm still a bit confused ;) causes() is defined in Fail. Just like iter_causes(). The first can be directy used as method on the Error trait (even though being deprecated), the second can't... o.O Why?

I also don't know if my solution really works, since I wasn't be able to raise a simple error somehow for testing :D How to do that?

[fitzgen]

Thanks @userzimmermann!

This needs to be tweaked a little bit before we merge it, see inline comment below.

[fitzgen]

Two things:

1. The iter_causes method already skips the first error in the chain, so we can remove the skip(1) bit now (yay!)
2. Rather than use unified function calling syntax (UFCS) we should use method calling syntax: e.iter_causes()

[userzimmermann]

Hi @fitzgen :) Regarding 1: Oops! Forgot to re-push after reading that iter_causes() "...does the skip(1) dance for us..." ;)

[userzimmermann]

Regarding 2 from #106 (comment)

Using method calling style e.iter_causes() strangely results in:

error[E0599]: no method named `iter_causes` found for type `traits::Error` in the current scope
  --> twiggy\./twiggy.rs:22:20
   |
22 |         for c in e.iter_causes() {
   |                    ^^^^^^^^^^^

This is that confusing thing I mentioned in my PR description, since there also doesn't seem to be any explicit traits::Error::causes() defined, only failure::Fail::causes(), but it can nevertheless be used via e.causes()

I already feel this PR will give me a huge initial boost in understanding Rust... ;P

[fitzgen]

Oh wow this is funky, it is a method on the trait object, not a provided trait method: https://docs.rs/failure/0.1.2/src/failure/lib.rs.html#227-229

I... don't know what the impetus for that was.

It should work if we do:

let e = &e as &Fail;
for c in e.iter_causes() {
    ...
}

If that doesn't do the trick, then we can just merge the current version of the PR!

[userzimmermann]

I also just fought my way to the trait Fail and impl Fail definitions :D And definitely need to get more understanding about those concepts...

[userzimmermann]

@fitzgen, is there no simpler way of casting than let e = &e as &Fail; ?

[userzimmermann]

@fitzgen, your cast suggestion does the trick! :) And can be even simplified:

for c in (&e as &Fail).iter_causes() {
    ...
}

But IMO, Fail::iter_causes(&e) looks somehow nicer ;)

[fitzgen]

Thanks for sticking with this @userzimmermann!

Sorry about the churn on the method calling syntax. I'm still confused why they opted to go for methods on trait objects rather than trait methods! I would have been tripped up by that too, when writing this the first time :-/

[fitzgen]

But IMO, Fail::iter_causes(&e) looks somehow nicer ;)

agreed

[userzimmermann]

Thanks @fitzgen for merging and the nice discussion! But my poll https://twitter.com/zimmermanncode/status/1026540101202059264 about which calling style to use wasn't over yet ;P Luckily everybody voted for Fail::iter_causes(&e) so far :D

I learned a lot from this PR! And will try to get even more involved in the Rust WebAssembly community...