Test for appropriate span on second custom derive

[estebank]

Adds test for and closes #36935.

[rust-highfive]

r? @brson

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

[petrochenkov]

@bors r+ rollup

[bors]

📌 Commit dfb6ae6 has been approved by petrochenkov

[bors]

⌛ Testing commit dfb6ae6 with merge 0fbeada...

[bors]

💔 Test failed - status-appveyor

[estebank]

Has there been a change in how proc_macro is used in master?

error[E0463]: can't find crate for `proc_macro`
  --> C:\projects\rust\src/test\ui\custom-derive\auxiliary\plugin.rs:16:1
   |
16 | extern crate proc_macro;
   | ^^^^^^^^^^^^^^^^^^^^^^^^ can't find crate
error: aborting due to previous error

[alexcrichton]

Currently the ui test suite is not configured to have librustc available when it runs, so proc-macro tests in the ui test suite aren't guaranteed to work. Rustbuild is nondeterministic in how it lays out steps, so that may explain why it worked locally.

If we really want to add this test as a ui test then the suite needs to be moved below like the other "fulldeps" suites.

[estebank]

@alexcrichton I'm trying to figure out why I needed to remove the ui assertion from the test_with_a_target test for the test suite to run successfully, and what are the consequences from me doing so.

[alexcrichton]

@estebank no worries that's ok. You can actually change the assertion to testing that it's not present instead of testing that it's present.

That basically means that we only run the tests for hosts, not targets (which is correct)

[alexcrichton]

r=me with that change

[estebank]

@alexcrichton, done. I also removed two duplicated asserts while I was at it.

[alexcrichton]

@bors: r+

[bors]

📌 Commit d8b3a64 has been approved by alexcrichton

[bors]

⌛ Testing commit d8b3a64 with merge db073ca...

[bors]

💔 Test failed - status-travis

[alexcrichton]

@bors: retry * network error

…

On Mon, Jan 9, 2017 at 9:06 AM, bors ***@***.***> wrote: 💔 Test failed - status-travis <https://travis-ci.org/rust-lang/rust/builds/190300898> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#38607 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAD95FP30U7nj8Ib1CyQ69omd3tvLCvUks5rQmkWgaJpZM4LVd9Q> .