Add object safety to TRPL

[steveklabnik]

Fixes #26938

[rust-highfive]

r? @huonw

(rust_highfive has picked a reviewer for you, use r? to override)

[steveklabnik]

looks like a spurious travis failure?

[Gankra]

Disagree?

run doc-trpl-trait-objects [x86_64-unknown-linux-gnu]
test The_Stack_0 ... �[32mok�(B�[m

running 11 tests
test _7 ... �[33mignored�(B�[m
test _8 ... �[33mignored�(B�[m
test _9 ... �[33mignored�(B�[m
test _10 ... �[31mFAILED�(B�[m
test The_Stack_2 ... �[32mok�(B�[m
test Arguments_and_borrowing_0 ... �[32mok�(B�[m
test The_Stack_1 ... �[32mok�(B�[m
test A_complex_example_0 ... �[32mok�(B�[m
test The_Heap_0 ... �[32mok�(B�[m


...


failures:

---- _10 stdout ----
    <anon>:2:59: 2:60 error: unknown start of token: `
<anon>:2     error: cannot convert to a trait object because trait `core::clone::Clone` is not object-safe [E0038]
                                                                   ^
thread '_10' panicked at 'Box<Any>', src/libsyntax/parse/lexer/mod.rs:167



failures:
    _10

[Gankra]

(had to go into the full log: https://s3.amazonaws.com/archive.travis-ci.org/jobs/74270996/log.txt)

[Gankra]

Shouldn't both of these be marked rust?

[Gankra]

Or rather, this one should be marked text (words are hard).

[steveklabnik]

@gankro ahhh full log

[steveklabnik]

failures should be addressed now, r?

[Gankra]

What does "receiver that has type Self" even mean? (Jargon isn't clicking for me today)

[steveklabnik]

I just copied it from the RFC... come to think of it, I'm not 100% sure either.

[Gankra]

It sounds like it's saying it's fn foo(self) but that's not object safe today (though there is a possible route to making it so). This is why FnBox is a thing.

CC @nikomatsakis

[nikomatsakis]

@gankro ah, I never answered here, sorry. I think yes it was looking forward to a day where fn foo(self) is supported. In particular, at the time of the RFC, I think there was SOME kind of hacky support for by-value self methods in trait objects, though I later removed it because it didn't handle all cases and was hard to generalize.

[steveklabnik]

I've removed the part you were wondering about, @gankro , based on @nikomatsakis 's feedback. r?

[alexcrichton]

@bors: r+ ce1bdc7

[bors]

⌛ Testing commit ce1bdc7 with merge dc9b74f...

[bors]

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-x-android-t, auto-mac-32-opt, auto-mac-64-nopt-t, auto-mac-64-opt, auto-win-gnu-32-nopt-t, auto-win-gnu-32-opt, auto-win-gnu-64-nopt-t, auto-win-gnu-64-opt, auto-win-msvc-32-opt, auto-win-msvc-64-opt