ICH: Fix and test foreign mod hashing.

[michaelwoerister]

r? @nikomatsakis

[eddyb]

Shouldn't this be part of FnDecl?

[michaelwoerister]

visit_fn() is not called for foreign items, and it's not clear that it should be (since foreign items have no body like the other cases). As far as I could tell, this is the only way to get at a foreign item's FnDecl without changing the visitor implementation.

[eddyb]

Oh there is no visit_fn_decl? Maybe you should add that instead.

[michaelwoerister]

I'm afraid, you're right :)

[michaelwoerister]

I've added a visit_fn_decl method to intravisit::Visitor and looked through all occurrences of visit_fn. None of them seemed like it would need to change or could be implemented much more nicely via visit_fn_decl. So the only use of the new method is in the ICH hasher.

[michaelwoerister]

I removed hashes for individual foreign items, which caused no trouble for the incremental tests locally. Let's see what Travis says.

[nikomatsakis]

@michaelwoerister looks good. 👍

[nikomatsakis]

@michaelwoerister though I feel like this caused some sort of ICE -- do you think you could use this branch to build syntex with RUSTFLAGS="-Z incremental" or something like that?

[michaelwoerister]

do you think you could use this branch to build syntex with RUSTFLAGS="-Z incremental" or something like that?

Sure.

[michaelwoerister]

So, I did a full bootstrap on this branch, with both stage1 and stage2 being built with -Zincremental and haven't run into any ICEs.

[nikomatsakis]

@bors r+

[bors]

📌 Commit f0a630b has been approved by nikomatsakis

[bors]

⌛ Testing commit f0a630b with merge a9ab778...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: nikomatsakis
Pushing a9ab778 to master...