Don't build rustdoc in more than one stage

[Mark-Simulacrum]

This also introduces a check that if we're attempting to build rustdoc in more than one stage we bail out and abort the build. It shouldn't ever be necessary to do so within one build (it may be necessary locally, primarily, where you don't want a full stage 2 compiler), so just bail out if it happens.

r? @alexcrichton

[Mark-Simulacrum]

This might not pass CI -- I've not thoroughly tested -- but it'll give us information on the failures which we can then fix. I think I've fixed the case we know about, though.

[alexcrichton]

Hm this doesn't feel quite right because in theory we should be able to request documentation from any stage of rustdoc, but is it a bug that stage1 docs are requested somewhere or something like that?

[rust-highfive]

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:0225c4a8:start=1554914663959517361,finish=1554914664852855989,duration=893338628
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---
make -j 4 check
[01:02:45]     Finished dev [unoptimized] target(s) in 0.36s
[01:02:46] thread 'main' panicked at 'assertion failed: `(left == right)`
[01:02:46]   left: `2`,
[01:02:46]  right: `1`: Only ever build one copy of rustdoc per run, currently built in stages [2, 1]', src/bootstrap/tool.rs:464:17
[01:02:46] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test
[01:02:46] Build completed unsuccessfully in 0:00:02
[01:02:46] Build completed unsuccessfully in 0:00:02
[01:02:47] make: *** [check] Error 1
[01:02:47] Makefile:48: recipe for target 'check' failed
The command "stamp sh -x -c "$RUN_SCRIPT"" exited with 2.
travis_time:start:00bd1ccf
$ date && (curl -fs --head https://google.com | grep ^Date: | sed 's/Date: //g' || true)
Wed Apr 10 17:47:24 UTC 2019

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

[Mark-Simulacrum]

Hm this doesn't feel quite right because in theory we should be able to request documentation from any stage of rustdoc, but is it a bug that stage1 docs are requested somewhere or something like that?

Hm, yeah, now that I think a little more about it I think that this change is wrong. We should be using the requested stage, not just stage 2. (To be clear, prior to #58897, we would always use top_stage for rustdoc which then did the right thing for documentation as well). I'll look into the relevant fix (and what caused CI to fail).

[rust-highfive]

The job x86_64-gnu-llvm-6.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

travis_time:end:05e485c0:start=1555199546374667017,finish=1555199547111449582,duration=736782565
$ git checkout -qf FETCH_HEAD
travis_fold:end:git.checkout

Encrypted environment variables have been removed for security reasons.
See https://docs.travis-ci.com/user/pull-requests/#pull-requests-and-security-restrictions
$ export SCCACHE_BUCKET=rust-lang-ci-sccache2
$ export SCCACHE_REGION=us-west-1
$ export GCP_CACHE_BUCKET=rust-lang-ci-cache
$ export AWS_ACCESS_KEY_ID=AKIA46X5W6CZEJZ6XT55
---
travis_time:start:test_assembly
Check compiletest suite=assembly mode=assembly (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
[01:13:41] 
[01:13:41] running 9 tests
[01:13:41] iiiiiiiii
[01:13:41] 
[01:13:41]  finished in 0.154
[01:13:41] travis_fold:end:test_assembly

---
travis_time:start:test_debuginfo
Check compiletest suite=debuginfo mode=debuginfo-both (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
[01:13:58] 
[01:13:58] running 121 tests
[01:14:23] .iiiii...i.....i..i...i..i.i.i..i.ii...i.....i..i....i..........iiii..........i...ii...i.......ii.i. 100/121
[01:14:28] i.i......iii.i.....ii
[01:14:28] 
[01:14:28]  finished in 30.513
[01:14:28] travis_fold:end:test_debuginfo

---
travis_time:start:test_stage2-test
Testing test stage2 (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
[01:21:56]    Compiling getopts v0.2.17
[01:21:59]    Compiling test v0.0.0 (/checkout/src/libtest)
[01:21:59] error[E0523]: found two different crates with name `getopts` that are not distinguished by differing `-C metadata`. This will result in symbol conflicts between the two.
[01:21:59]    |
[01:21:59] 34 | use getopts;
[01:21:59]    |     ^^^^^^^
[01:21:59] 
---
[01:21:59] warning: build failed, waiting for other jobs to finish...
[01:22:09] error: build failed
[01:22:09] 
[01:22:09] 
[01:22:09] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "test" "--target" "x86_64-unknown-linux-gnu" "-j" "4" "--release" "--locked" "--color" "always" "--manifest-path" "/checkout/src/libtest/Cargo.toml" "-p" "test" "--" "--quiet"
[01:22:09] 
[01:22:09] 
[01:22:09] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test
[01:22:09] Build completed unsuccessfully in 0:20:12
[01:22:09] Build completed unsuccessfully in 0:20:12
[01:22:09] Makefile:48: recipe for target 'check' failed
[01:22:09] make: *** [check] Error 1
The command "stamp sh -x -c "$RUN_SCRIPT"" exited with 2.
travis_time:start:06ef1ad4
$ date && (curl -fs --head https://google.com | grep ^Date: | sed 's/Date: //g' || true)
Sun Apr 14 01:14:47 UTC 2019

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

[Dylan-DPC-zz]

ping from triage @Mark-Simulacrum tests are failing can you address that?