Use std::process::Command throughout compile-test

[euclio]

Resubmission of #43798.

Fixes#43762.

r? @alexcrichton

[alexcrichton]

@bors: r+

[bors]

📌 Commit 88bac1f has been approved by alexcrichton

[bors]

⌛ Testing commit 88bac1f with merge 8ba7852589b76b9c92d29a09f90777d9b7869cf2...

[bors]

💔 Test failed - status-travis

[kennytm]

All GDB tests on android-arm failed with "Can't bind address: Address already in use". Possibly related to this PR since the testing code is modified.

Details

Check compiletest suite=debuginfo mode=debuginfo-gdb (x86_64-unknown-linux-gnu -> arm-linux-androideabi)
[00:58:24] arm-linux-androideabi debug-info test uses tcp 5039 port.please reserve it
[00:58:24] [00:58:24] running 108 tests
[00:58:24] test [debuginfo-gdb] debuginfo-gdb/associated-types.rs ... [100%] /data/tmp/work/associated-types.stage2-arm-linux-androideabi
[00:58:24] /checkout/obj/build/x86_64-unknown-linux-gnu/test/debuginfo/associated-types.stage2-arm-linux-androideabi: 1 file pushed. 1.8 MB/s (18292 bytes in 0.010s)
[00:58:24] FAILED
[00:58:24] test [debuginfo-gdb] debuginfo-gdb/basic-types-globals-metadata.rs ... Process /data/tmp/work/associated-types.stage2-arm-linux-androideabi created; pid = 31093
[00:58:24] Listening on port 5039
[00:58:24] [100%] /data/tmp/work/basic-types-globals-metadata.stage2-arm-linux-androideabi
[00:58:24] /checkout/obj/build/x86_64-unknown-linux-gnu/test/debuginfo/basic-types-globals-metadata.stage2-arm-linux-androideabi: 1 file pushed. 1.5 MB/s (12668 bytes in 0.008s)
[00:58:24] FAILED
[00:58:24] test [debuginfo-gdb] debuginfo-gdb/basic-types-globals.rs ... Process /data/tmp/work/basic-types-globals-metadata.stage2-arm-linux-androideabi created; pid = 31098
[00:58:24] Can't bind address: Address already in use.
[00:58:24] Killing process(es): 31098
[00:58:24] [100%] /data/tmp/work/basic-types-globals.stage2-arm-linux-androideabi
[00:58:24] /checkout/obj/build/x86_64-unknown-linux-gnu/test/debuginfo/basic-types-globals.stage2-arm-linux-androideabi: 1 file pushed. 2.0 MB/s (12332 bytes in 0.006s)
[00:58:24] FAILED
<snip>
[00:58:53] test [debuginfo-gdb] debuginfo-gdb/vec.rs ... Process /data/tmp/work/vec-slices.stage2-arm-linux-androideabi created; pid = 31573
[00:58:53] Can't bind address: Address already in use.
[00:58:53] Killing process(es): 31573
[00:58:53] [100%] /data/tmp/work/vec.stage2-arm-linux-androideabi
[00:58:53] /checkout/obj/build/x86_64-unknown-linux-gnu/test/debuginfo/vec.stage2-arm-linux-androideabi: 1 file pushed. 1.6 MB/s (9412 bytes in 0.005s)
[00:58:53] FAILED
[00:58:53] [00:58:53] failures:
[00:58:53] [00:58:53] ---- [debuginfo-gdb] debuginfo-gdb/associated-types.rs stdout ----
[00:58:53] NOTE: compiletest thinks it is using GDB without native rust support
[00:58:53] thread '[debuginfo-gdb] debuginfo-gdb/associated-types.rs' panicked at 'called `Option::unwrap()` on a `None` value', /checkout/src/libcore/option.rs:335:20
[00:58:53] note: Run with `RUST_BACKTRACE=1` for a backtrace.
[00:58:53] [00:58:53] ---- [debuginfo-gdb] debuginfo-gdb/basic-types-globals-metadata.rs stdout ----
[00:58:53] NOTE: compiletest thinks it is using GDB without native rust support
[00:58:53] thread '[debuginfo-gdb] debuginfo-gdb/basic-types-globals-metadata.rs' panicked at 'called `Option::unwrap()` on a `None` value', /checkout/src/libcore/option.rs:335:20
[00:58:53] <snip>
[00:58:53] ---- [debuginfo-gdb] debuginfo-gdb/vec.rs stdout ----
[00:58:53] NOTE: compiletest thinks it is using GDB without native rust support
[00:58:53] thread '[debuginfo-gdb] debuginfo-gdb/vec.rs' panicked at 'called `Option::unwrap()` on a `None` value', /checkout/src/libcore/option.rs:335:20
[00:58:53] [00:58:53] [00:58:53] failures:
[00:58:53] [debuginfo-gdb] debuginfo-gdb/associated-types.rs
[00:58:53] [debuginfo-gdb] debuginfo-gdb/basic-types-globals-metadata.rs
<snip>
[00:58:53] [debuginfo-gdb] debuginfo-gdb/vec.rs
[00:58:53] [00:58:53] test result: FAILED. 0 passed; 98 failed; 10 ignored; 0 measured; 0 filtered out

[euclio]

Hm, looking at the GDB and ADB code, I don't see anything glaringly different.

[Mark-Simulacrum]

hm, I think passing &env here (and in other calls to envs) should work, but I'm not certain.

[euclio]

This won't work because envs requires the Item to be a tuple, not a tuple reference.

[Mark-Simulacrum]

Yeah, not clear to me why this would have failed CI. Could you try running locally and seeing if you can reproduce? ./src/ci/docker/run.sh arm-android.

[alexcrichton]

I think that the stdout for the spawned adb shell process isn't captured like it was before, which I think can cause problems like this in this situation.

[euclio]

Thanks for the hints. I am on vacation through Tuesday but I'll take a look when I can.

…

On Sun, Aug 20, 2017, 1:43 PM Alex Crichton ***@***.***> wrote: I think that the stdout for the spawned adb shell process isn't captured like it was before, which I think can cause problems like this in this situation. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#43907 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABTxFp2MrzGH1tehPUJAttylqpHdYRqSks5saH5sgaJpZM4O5Mar> .

[euclio]

I don't have a computer powerful enough to run the android tests locally for now, but I pushed a commit that reattaches the stdout and stderr for adb shell.

[Mark-Simulacrum]

You can temporarily add ALLOW_PR=1 to the android builder so that it runs on Travis and we can get feedback without going through bors.

[alexcrichton]

Er the stdout here should be Stdio::piped where the stderr is Stdio::inherit, we try to capture the stdout just below the call to spawn.

[euclio]

@alexcrichton Fixed.

[alexcrichton]

@bors: r+

Let's see what happens!

[bors]

📌 Commit 91bfe3f has been approved by alexcrichton

[bors]

⌛ Testing commit 91bfe3f with merge a125ec9...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: alexcrichton
Pushing a125ec9 to master...