Exclude some long-running tests when running under miri

[jhorstmann]

Which issue does this PR close?

Closes #1862.

Rationale for this change

Stacked borrow checking uses a lot of cpu and memory resources. Disabling it should make CI much more reliable.

What changes are included in this PR?

Are there any user-facing changes?

[jhorstmann]

This did not seem to fix CI. Test tests still seem to get killed and then retried by miri.sh:

test compute::kernels::cast_utils::tests::string_to_timestamp_timezone_space ... ok
error: test failed, to rerun pass '-p arrow --lib'
Caused by:
  process didn't exit successfully: `/home/runner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/bin/cargo-miri /home/runner/work/arrow-rs/arrow-rs/target/miri/x86_64-unknown-linux-gnu/debug/deps/arrow-ca24f64806e3ede9 --skip csv --skip ipc --skip json` (exit status: 255)
test compute::kernels::comparison::tests::ilike_utf8_scalar_escape_testing ... Starting Arrow MIRI run...
    Finished test [unoptimized + debuginfo] target(s) in 0.38s
     Running unittests src/lib.rs (target/miri/x86_64-unknown-linux-gnu/debug/deps/arrow-ca24f64806e3ede9)
running 1038 tests
test array::array::tests::test_array_ref_as_array ... ok

[jhorstmann]

error: test failed, to rerun pass '-p arrow --lib'
Caused by:
  process didn't exit successfully: `/home/runner/.rustup/toolchains/nightly-x86_64-unknown-linux-gnu/bin/cargo-miri /home/runner/work/arrow-rs/arrow-rs/target/miri/x86_64-unknown-linux-gnu/debug/deps/arrow-ca24f64806e3ede9 --skip csv --skip ipc --skip json` (exit status: 255)
test compute::kernels::comparison::tests::ilike_utf8_scalar_escape_testing ... Miri finished with exit code 255

That's actually not the error code I expected if it is an out of memory issue

[jhorstmann]

Disabling stacked-borrow checking is not perfect, but probably better to run miri with one less check than not running it at all. The check using a lot of cpu and memory is a known problem: rust-lang/miri#1367

[tustvold]

This seems sensible 👍. We could conceivably run the full MIRI suite as part of release validation, or something, but I also don't feel particularly strongly about this

[codecov-commenter]

Codecov Report

Merging #1863 (0d265c0) into master (fb697ce) will decrease coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1863      +/-   ##
==========================================
- Coverage   83.49%   83.47%   -0.03%     
==========================================
  Files         201      201              
  Lines       56902    56987      +85     
==========================================
+ Hits        47511    47570      +59     
- Misses       9391     9417      +26     

Impacted Files	Coverage Δ
arrow/src/array/array_union.rs	90.92% <ø> (ø)
arrow/src/compute/kernels/filter.rs	88.33% <ø> (ø)
arrow/src/util/bit_chunk_iterator.rs	93.83% <ø> (ø)
parquet_derive/src/parquet_field.rs	65.75% <0.00%> (-0.23%)	⬇️
parquet/src/encodings/encoding.rs	93.46% <0.00%> (-0.20%)	⬇️
parquet/src/arrow/async_reader.rs	0.00% <0.00%> (ø)
arrow/src/datatypes/datatype.rs	65.79% <0.00%> (+0.37%)	⬆️
parquet/src/file/serialized_reader.rs	94.95% <0.00%> (+0.49%)	⬆️
arrow/src/array/array_binary.rs	94.24% <0.00%> (+0.92%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fb697ce...0d265c0. Read the comment docs.

[tustvold]

Looks good to me