Install sccache

[aneeshusa]

Part of servo/servo#17106.

---

This change is 

[aneeshusa]

r? @larsbergstrom
cc @luser

[larsbergstrom]

Just one question - where does this come from? Do we want our own for Servo? cc @edunham

[aneeshusa]

Official binary releases are not yet available for sccache, so per suggestion I copied these links from rust-lang/rust CI scripts.

[larsbergstrom]

Makes sense - sounds good to me! This is r=me, then. Looks like the travis failure is just atimeout to repo.saltstack.com?

[aneeshusa]

I wish, there's a real test failure on the cross build due to the Android SDK check I added, this looks like it might require some yak shaving. Investigating.

[bors-servo]

☔ The latest upstream changes (presumably #694) made this pull request unmergeable. Please resolve the merge conflicts.

[larsbergstrom]

Does this just need a rebase now that all the Android SDK build stuff has been updated?

[aneeshusa]

@bors-servo retry
I've updated this with a workaround for the failing Android SDK test; I think there is a slightly better way possible but it requires a newer Salt and/or a bunch of backporting, so this is simpler.
This should clean up the builds that failed due to force pushing and being unable to find the previous SHA.
Looking at the saltfs build history, we also seem to be hitting travis-ci/travis-ci#8363. I can look for a temporary workaround for that if we don't want to wait.

[jdm]

@bors-servo: r=larsberg

[bors-servo]

📌 Commit 63f02a6 has been approved by larsberg

[bors-servo]

⌛ Testing commit 63f02a6 with merge faf753b...

[bors-servo]

💔 Test failed - status-travis

[larsbergstrom]

hrm:

$ export GHTOKENLINE=[secure]

The previous command failed, possibly due to a malformed secure environment variable.

Please be sure to escape special characters such as ' ' and '$'.

For more information, see https://docs.travis-ci.com/user/encryption-keys.

Does that look familiar?

[aneeshusa]

@larsbergstrom that's a separate problem; I thought I had an issue tracking that, but not yet it seems. We're not using this variable (I think it was for a PR we decided not to merge), so it should be unset at some point, but that isn't causing the test failure.

The actual test failure is caused by travis-ci/travis-ci#8363; we want to get Python 3.5 but are getting a weird Python 3.4 instead, causing the test failure. If you want sccache now we can add a workaround, or we can wait for Travis to fix this.

[aneeshusa]

@bors-servo r=larsbergstrom

[bors-servo]

💡 This pull request was already approved, no need to approve it again.

• This pull request previously failed. You should add more commits to fix the bug, or use retry to trigger a build again.
• There's another pull request that is currently being tested, blocking this pull request: #716

[bors-servo]

📌 Commit 63f02a6 has been approved by larsbergstrom

[bors-servo]

⌛ Testing commit 63f02a6 with merge e4a1d01...

[bors-servo]

☀️ Test successful - status-travis
Approved by: larsbergstrom
Pushing e4a1d01 to master...