r? @alexcrichton

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

@bors r+ p=1 since this may cause PRs to fail

📌 Commit ec2f3e13e035d13985abf2c6f6aee8059b1bf0c2 has been approved by Mark-Simulacrum

I would personally prefer if this were only applied to the location necessary for llvm, all the others should be inconsequential and/or handled already. Could the comment be expanded as well as to why it is applied, not just what it is?

@alexcrichton

Could the comment be expanded as well as to why it is applied, not just what it is?

The reason is that CMAKE_INSTALL_MESSAGE=LAZY is almost always what you want, during local development or on CI, doesn't matter.
It's not the default in CMake because CMAKE_INSTALL_MESSAGE is a relatively recent addition (introduced in CMake 3.1 https://cmake.org/pipermail/cmake/2014-December/059418.html).
Perhaps cmake-rs can even set it by default.

The Travis issue with overflowing logs is a trigger for enabling it rather than the reason.

The option is useless if the install directory is certainly fresh and empty.
I don't know whether that's the case for cmake that's run from src/ci, so I enabled it everywhere.

For cmake that's run from *.rs that's not the case.

It makes sense yeah that it's a better default, and I'm not disagreeing at all. All our build scripts are already complicated and this is just one more thing to understand whenever reading them, and it's basically easier to read zero code than it is to reason later "did someone add this for correctness?" or having to preserve this through all future changes. I would rather than this be surgical and only apply, with a comment, to where it matters. Changing cmake-rs itself also seems totally fine by me.

OK, the messages in the failed job were produced by LLVM and LLD builds, so I reduced the changes to those only.

@bors r=Mark-Simulacrum

📌 Commit 3646b3c has been approved by Mark-Simulacrum

⌛ Testing commit 3646b3c with merge 7158ed9...

☀️ Test successful - checks-travis, status-appveyor
Approved by: Mark-Simulacrum
Pushing 7158ed9 to master...