Integrate privacy into field and method selection

[jseyfried]

This PR integrates privacy checking into field and method selection so that an inaccessible field/method cannot stop an accessible field/method from being used (fixes #12808 and fixes #22684).
r? @eddyb

[jseyfried]

cc @nikomatsakis @petrochenkov

[nikomatsakis]

On Sat, Feb 27, 2016 at 01:23:08AM -0800, Jeffrey Seyfried wrote:

cc @nikomatsakis @petrochenkov

you're on a roll! Just FYI I sat down over last few days to read over
some of your changes to resolve in detail, because I realize I've kind
of lost track of how things are working, and am starting to feel
uncomfortable reviewing :) hoping to finish that up tonight though

[jseyfried]

@nikomatsakis Ok, let me know if you'd like me to explain or comment anything!

[jseyfried]

Fixes #22684

[eddyb]

@jseyfried you need to put that in the description.

[jseyfried]

@eddyb I updated the description.

[jseyfried]

I rebased and added a commit that refactors away most of the privacy visitor.

[bors]

☔ The latest upstream changes (presumably #32432) made this pull request unmergeable. Please resolve the merge conflicts.

[nikomatsakis]

@jseyfried can you add some tests that show that the logic about private fields/methods is working? I was thinking it'd be nice to have a test where the private field appears in an outer deref. In other words, if the autoderef chain of types is A -> B -> C, then the field/method is on type A or B.

r=me with such a test.

[nikomatsakis]

can you convert this to a run-pass test and somehow check the return value of f, so that we know the right function is being called?

[jseyfried]

Will do

[nikomatsakis]

Sorry, let me clarify my previous comment. I'd like to see two things:

1. Modify the existing test for private fields/methods to be run-pass, and using some logic that proves we are picking the right fields.
   • Maybe also extend the test to read from the fields in a context where it should pick the private one, and prove that this logic is working correctly.
2. A test specifically targeting the case where only a private field/method is available, which would target the logic you put in around that case.

[jseyfried]

@nikomatsakis

1. Modify the existing test for private fields/methods to be run-pass, and using some logic that proves we are picking the right fields.
   • Maybe also extend the test to read from the fields in a context where it should pick the private one, and prove that this logic is working correctly.

Done.

A test specifically targeting the case where only a private field/method is available

Private inherent methods are tested in private-method-inherited (and others) and I amended struct-field-privacy to test private positional fields as well as private named fields.

I believe these tests are sufficient to check the logic you referenced, but if you want I can add a compile-fail test where there is only a private field/method in a non-trivial autoderef chain.

[nikomatsakis]

@jseyfried I'll leave it to your judgement. I certainly lean towards more exhaustive testing but there is no reason to add duplicate tests either. R=me either way.

[jseyfried]

Ok, thanks for reviewing!

@bors r=nikomatsakis

[bors]

📌 Commit 48c20b0 has been approved by nikomatsakis

[bors]

⌛ Testing commit 48c20b0 with merge 4f46ecb...

[bors]

💔 Test failed - auto-mac-32-opt

[jseyfried]

@bors retry

[alexcrichton]

@bors: retry force clean

[alexcrichton]

@bors: retry force clean

maybe second time's the charm?

[bors]

⌛ Testing commit 48c20b0 with merge 3399d19...

[bors]

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-debug-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-opt-mir, auto-linux-64-opt-rustbuild, auto-linux-64-x-android-t, auto-linux-64-x-armhf, auto-linux-64-x-armsf, auto-linux-64-x-freebsd, auto-linux-64-x-netbsd, auto-linux-cross-opt, auto-linux-musl-64-opt, auto-mac-32-opt, auto-mac-64-nopt-t, auto-mac-64-opt, auto-mac-ios-opt, auto-win-gnu-32-nopt-t, auto-win-gnu-32-opt, auto-win-gnu-32-opt-rustbuild, auto-win-gnu-64-nopt-t, auto-win-gnu-64-opt, auto-win-msvc-32-opt, auto-win-msvc-64-opt, auto-win-msvc-64-opt-mir, auto-win-msvc-64-opt-rustbuild