Reduce verbosity of build logs

[Mark-Simulacrum]

This does two separate things.

• Sets sccache logging to warn instead of info.
• Makes tests when running for a given PR (not on auto branch) quiet. (cc @eddyb)

r? @alexcrichton

[alexcrichton]

@bors: r+

Thanks @Mark-Simulacrum!

[bors]

📌 Commit 56f448d has been approved by alexcrichton

[Mark-Simulacrum]

Travis completed with 6,714 lines cleanly. This makes the build log nicely viewable!

[alexcrichton]

I'd love to combine test suites as well if we can eventually, right now there's hundreds of "running {0,1} tests" for all the markdown files in doc that we test but getting that into one giant list of tests would be great!

[Mark-Simulacrum]

@bors rollup

[kennytm]

So the rollup build #42367 has failed because the emscripten build has no output in 30 minutes during run-pass test.

I've checked the last successful auto build, the emscripten run-pass test,

[00:52:59] Check compiletest suite=run-pass mode=run-pass (x86_64-unknown-linux-gnu -> asmjs-unknown-emscripten)
...
[01:41:24] test result: ok. 2564 passed; 0 failed; 139 ignored; 0 measured; 0 filtered out
[01:41:24] 
[01:41:24] 	finished in 2904.985

the whole thing runs in 48m 25s. So indeed if the output is not flushed, Travis will normally cancel the job.

Since stamp won't print anything until seeing a \n, the quiet test should output a \n every 120 dots or so.

(Is it expected that emscripten is so slow? The run-pass test for x86_64 only runs for 6m 42s. The cost of asm.js is 7× native?)

[Mark-Simulacrum]

@bors r-

I'll investigate what we can do. Might be worth just not having stamp for PR builds...

[alexcrichton]

@bors: r+

[bors]

📌 Commit 3447aa7 has been approved by alexcrichton