missed this important bug

[Eh2406]

In the PR #5000 I finished and we merged yesterday I missed a bug and left in an outdated comment.

@alexcrichton

[rust-highfive]

r? @alexcrichton

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

[alexcrichton]

Oh nice catch! Out of curiosity, would it be possible to easily add a test for this as well? (if not no worries)

[Eh2406]

Looking into it more closely I think activation hits the MissingFeatures before it modified cx. So I think I was wrong, and this is not technically a bug. So I can't add a test. It is more a loaded foot gun for anyone adding to the kinds of activation errors. Which is how I came across it, I am working on a performance related PR that involved adding another activation error, and most of the existing test blew up in my face.

I will get ci green in a sec.

[alexcrichton]

Ah ok! Sounds good to me to merge when gree nthen

[Eh2406]

CI is green except for a spurious network error.

[alexcrichton]

@bors: r+

Oh dear, that test should be updated to not hit the real index...

[bors]

📌 Commit 2fa4fdf has been approved by alexcrichton

[bors]

⌛ Testing commit 2fa4fdf with merge 91739dc2c49d7f1922e9bd606ca266fb2043e7e6...

[bors]

💥 Test timed out

[Eh2406]

Added a commit that should help with the spurious network error.
How do I update the test to not hit the real index?
The test timeout I don't understand, the link is https://ci.appveyor.com/project/rust-lang-libs/cargo/build/1.0.3765 I think. but that seems unrelated to my change.

[alexcrichton]

@bors: r+

Oh I've generally used Pakage::new("foo", "0.1.0").publish() which rewrites the test's home directory to use a per-test index instead.

[bors]

📌 Commit d9c97de has been approved by alexcrichton

[bors]

⌛ Testing commit d9c97de with merge 8f76fde...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: alexcrichton
Pushing 8f76fde to master...