Downgrade unit struct match via S(..) warnings to errors

[pnkfelix]

Downgrade unit struct match via S(..) warnings to errors

The error signalling was introduced in #29383

It was noted as a warning-cycle-less regression in #30379

Fix#30379

[rust-highfive]

r? @jroesch

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

[pnkfelix]

r? @nikomatsakis

(This is high priority regression fix)

[nikomatsakis]

Hmm, I think we should modify bad_struct_kind_err to look at is_warning and add a lint in the case of is_warning == true. This will also require adding a new lint corresponding to this future incompatibility. That said, I was going to do a quick PR to improve that tooling so as to address #30746, so maybe I could just roll that change into that.

[pnkfelix]

I debated about doing this but decided I was better off keeping this PR minimal to ease backporting to beta... what do you think?

[pnkfelix]

(I think I should have pushed back harder here. But then again I didn't expect bors to take four days to get to this)

[pnkfelix]

@nikomatsakis Okay I think this is what you want w.r.t. the lint structure. How does this look to you?

[nikomatsakis]

Looks great!

@bors r+

[bors]

📌 Commit 40e2ac2 has been approved by nikomatsakis

[pnkfelix]

failures:
[compile-fail] compile-fail/match-pattern-field-mismatch-2.rs
[compile-fail] compile-fail/pattern-error-continue.rs

hmm, why did that arise on the travis build

[pnkfelix]

@bors r-

i need to look at the two failure cases

[pnkfelix]

@bors r=nikomatsakis fa027d1

[nikomatsakis]

@bors r+

[bors]

📌 Commit fa027d1 has been approved by nikomatsakis

[nikomatsakis]

@bors r-

[nikomatsakis]

@bors r+

[bors]

📌 Commit fa027d1 has been approved by nikomatsakis

[bors]

⌛ Testing commit fa027d1 with merge 41f7bf7...

[bors]

💔 Test failed - auto-win-msvc-32-opt

[alexcrichton]

@bors: retry

On Mon, Jan 11, 2016 at 4:53 PM, bors notifications@github.com wrote:

[image: 💔] Test failed - auto-win-msvc-32-opt
http://buildbot.rust-lang.org/builders/auto-win-msvc-32-opt/builds/1878

—
Reply to this email directly or view it on GitHub
#30753 (comment).

[pnkfelix]

@bors p=1

(upping priority for regression fixing or beta-nominated PR's)

[bors]

⌛ Testing commit fa027d1 with merge 5cf69aa...

[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-x-android-t, auto-linux-cross-opt, auto-linux-musl-64-opt, 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

[pnkfelix]

@nikomatsakis I added #30724 to the beta-nominated set because it seemed like a natural candidate for a back port to beta (and also because 40e2ac2 in this PR depends on infrastructure added by that PR)

however, given that the beta is going to be cut very soon, perhaps back porting PR #30724 would be unwise, and thus I should make a separate PR against beta branch that only goes up to 8aed830 (inclusive)

Thoughts?

Cc @brson @rust-lang/compiler

[nikomatsakis]

@pnkfelix and I discussed on IRC -- seems like best thing is to backport this PR and just the (very few) bits of infrastructure needed to make it work.