Remove ty_open.

[eddyb]

This type wasn't necessary, as there was no place using it and unsized types not wrapped in it, at the same time.
r? @nikomatsakis

[nikomatsakis]

@eddyb r+ modulo comments, this seems nice overall (and @nick29581 mentioned he was fine with it too). But I'd like to understand what's going on with that change to type-needs-drop.

[eddyb]

@bors r=nikomatsakis 258fa79

[bors]

⌛ Testing commit 258fa79 with merge 6d44972...

[bors]

💔 Test failed - auto-mac-64-nopt-t

[eddyb]

I was going to adjust the tests for the new errors, but it looks like that would be a regression.

I'm tempted to use something like the current debug location to always provide some span to functions like ty::type_is_sized which may need it. However, trans::common::type_is_sized could at most require a CrateContext, as it's used from Datum and type_of and at least the latter has to be usable from outside a FunctionContext.

A global "current Span" is too hacky for my taste but we mainly need that span for such _extra_ordinary overflow errors, and only to point to something better than "somewhere in this crate".
@nikomatsakis, what are your thoughts on this?

[nikomatsakis]

@eddyb I personally think you can just adjust the tests. This is a pre-existing problem. I thought there was an issue for it, but I can't find it now. In any case, I think that in the end we should probably just thread a span down through typeof and so forth, though, annoying as that is.

[nikomatsakis]

(We might want to open an issue and attach a FIXME in type_is_sized and on the tests.)

[eddyb]

It's actually not coming from type_needs_drop so that means I forgot to run the compile-fail tests before that. In any case, if you don't mind a temporary loss of diagnostic information, I won't.

[eddyb]

@bors r=nikomatsakis 85052b4

[eddyb]

@bors r=nikomatsakis 724c207

[bors]

⌛ Testing commit 724c207 with merge e6fdf5b...

[bors]

💔 Test failed - auto-mac-64-nopt-t

[Manishearth]

failures:

---- [run-pass-valgrind] run-pass-valgrind/dst-dtor-1.rs stdout ----

error: test run failed!
status: exit code: 101
command: x86_64-apple-darwin/test/run-pass-valgrind/dst-dtor-1.stage2-x86_64-apple-darwin 
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------
thread '<main>' panicked at 'assertion failed: DROP_RAN', /Users/rustbuild/src/rust-buildbot/slave/auto-mac-64-nopt-t/build/src/test/run-pass-valgrind/dst-dtor-1.rs:34

------------------------------------------

thread '[run-pass-valgrind] run-pass-valgrind/dst-dtor-1.rs' panicked at 'explicit panic', /Users/rustbuild/src/rust-buildbot/slave/auto-mac-64-nopt-t/build/src/compiletest/runtest.rs:1466


---- [run-pass-valgrind] run-pass-valgrind/dst-dtor-2.rs stdout ----

error: test run failed!
status: exit code: 101
command: x86_64-apple-darwin/test/run-pass-valgrind/dst-dtor-2.stage2-x86_64-apple-darwin 
stdout:
------------------------------------------

------------------------------------------
stderr:
------------------------------------------
thread '<main>' panicked at 'assertion failed: DROP_RAN == 3', /Users/rustbuild/src/rust-buildbot/slave/auto-mac-64-nopt-t/build/src/test/run-pass-valgrind/dst-dtor-2.rs:31

------------------------------------------

thread '[run-pass-valgrind] run-pass-valgrind/dst-dtor-2.rs' panicked at 'explicit panic', /Users/rustbuild/src/rust-buildbot/slave/auto-mac-64-nopt-t/build/src/compiletest/runtest.rs:1466

[eddyb]

@bors r=nikomatsakis 8659de0

[Manishearth]

@bors: p=10

[Manishearth]

Mostly everything else will come in the rollup.

[Manishearth]

@bors: p=5