Fix --nocapture for run-make tests

[jyn514]

This was confusing because there are three layers of output hiding.

1. libtest shoves all output into a buffer and does not print it unless the test fails or --nocapture is passed.
2. compiletest chooses whether to print the output from any given process.
3. run-make-support chooses what output to print.

This modifies 2 and 3.

• compiletest: Don't require both --verbose and --nocapture to show the output of run-make tests.
• compiletest: Print the output from rmake processes if they succeed. Previously this was only printed on failure.
• compiletest: Distinguish rustc and rmake stderr by printing the command name (e.g. "--stderr--" to "--rustc stderr--").
• run-make-support: Unconditionally print the needle/haystack being searched. Previously this was only printed on failure.

Before:

$ x t tests/run-make/linker-warning --force-rerun -- --nocapture
running 1 tests
.

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 377 filtered out; finished in 281.64ms
$ x t tests/run-make/linker-warning --force-rerun -v -- --nocapture 2>&1 | wc -l
1004
$ x t tests/run-make/linker-warning --force-rerun -v -- --nocapture | tail -n40
running 1 tests

------stdout------------------------------

------stderr------------------------------
warning: unused import: `std::path::Path`
 --> /home/jyn/src/rust2/tests/run-make/linker-warning/rmake.rs:1:5
  |
1 | use std::path::Path;
  |     ^^^^^^^^^^^^^^^
  |
  = note: `#[warn(unused_imports)]` on by default

warning: unused import: `run_make_support::rfs::remove_file`
 --> /home/jyn/src/rust2/tests/run-make/linker-warning/rmake.rs:3:5
  |
3 | use run_make_support::rfs::remove_file;
  |     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

warning: 2 warnings emitted

------------------------------------------
test [run-make] tests/run-make/linker-warning ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 377 filtered out; finished in 285.89ms

After:

Testing stage1 compiletest suite=run-make mode=run-make (x86_64-unknown-linux-gnu)

running 1 tests
------rmake stdout------------------------------

------rmake stderr------------------------------
assert_contains_regex:
=== HAYSTACK ===
error: linking with `./fake-linker` failed: exit status: 1
  |
  = note: LC_ALL="C" PATH="/home/jyn/src/rust2/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/bin:...:/bin" VSLANG="1033" "./fake-linker" "-m64" "/tmp/rustcYqdAZT/symbols.o" "main.main.d17f5fbe6225cf88-cgu.0.rcgu.o" "main.2uoctswmurc6ir5rvoay0p9ke.rcgu.o" "-Wl,--as-needed" "-Wl,-Bstatic" "-Wl,-Bdynamic" "-lgcc_s" "-lutil" "-lrt" "-lpthread" "-lm" "-ldl" "-lc" "-B/home/jyn/src/rust2/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/bin/gcc-ld" "-fuse-ld=lld" "-Wl,--eh-frame-hdr" "-Wl,-z,noexecstack" "-L" "/home/jyn/src/rust2/build/x86_64-unknown-linux-gnu/test/run-make/linker-warning/rmake_out" "-L" "/home/jyn/src/rust2/build/x86_64-unknown-linux-gnu/stage1/lib/rustlib/x86_64-unknown-linux-gnu/lib" "-o" "main" "-Wl,--gc-sections" "-pie" "-Wl,-z,relro,-z,now" "-nodefaultlibs" "run_make_error"
  = note: error: baz

error: aborting due to 1 previous error

=== NEEDLE ===
fake-linker.*run_make_error
assert_not_contains_regex:
=== HAYSTACK ===

=== NEEDLE ===
fake-linker.*run_make_error

------------------------------------------
.

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 377 filtered out; finished in 314.81ms

r? @jieyouxu

[rustbot]

The run-make-support library was changed

cc @jieyouxu

Some changes occurred in src/tools/compiletest

cc @jieyouxu

[jieyouxu]

I'll do some local testing tmrw, thanks for the fix!

[jieyouxu]

FTR, local manual testing indicates there seems to be some stdout/stderr interleaving, possibly due to some combination of #134040 + #134123. Will look into it a bit later.

EDIT: fallout from #134040, unrelated to this PR, which will be addressed by revert #134207.

[jieyouxu]

Thanks! I've done some local testing and have some feedback.

[jieyouxu]

@rustbot author

[jyn514]

@rustbot ready

[jieyouxu]

@rustbot blocked (on #134207)

[jieyouxu]

#134207 has merged so this is no longer blocked. However, I think there's one issue with the proc_name logic.
@rustbot author

[jyn514]

@rustbot ready

[jieyouxu]

Based on some local testing, this seems to work as expected (for the most part, On My Machine™). Example:

Testing stage1 compiletest suite=run-make mode=run-make (x86_64-pc-windows-msvc)

running 1 tests
------compiletest-self-test/rmake.exe stdout------------------------------
meow

------compiletest-self-test/rmake.exe stderr------------------------------
bark

------------------------------------------
.

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 382 filtered out; finished in 243.20ms

Thanks!

[jieyouxu]

@bors r+ rollup

[bors]

📌 Commit 056eb75 has been approved by jieyouxu

It is now in the queue for this repository.