This is an automated email from the ASF dual-hosted git repository.
ruifengz pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/spark.git
The following commit(s) were added to refs/heads/master by this push:
new 1b1904ec9f79 [SPARK-55405][PYTHON][TESTS][FOLLOW-UP] Add macOS ARM
overrides for float-to-int unsafe cast golden file mismatches
1b1904ec9f79 is described below
commit 1b1904ec9f790398e53d95fe6d0e3f3bec18d5a3
Author: Yicong-Huang <[email protected]>
AuthorDate: Fri Feb 27 13:16:17 2026 +0800
[SPARK-55405][PYTHON][TESTS][FOLLOW-UP] Add macOS ARM overrides for
float-to-int unsafe cast golden file mismatches
### What changes were proposed in this pull request?
Add macOS ARM-specific overrides in `_overrides_unsafe()` for 7
float-to-int unsafe cast cases that differ between Linux ARM and macOS ARM due
to LLVM code generation differences. This fixes the `build_python_3.12_macos26`
CI workflow which has been failing on every run.
### Why are the changes needed?
Fix macOS 26 CI failures.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Tested on macOS 26 (Darwin 25.2.0, arm64) with PyArrow 22.0.0. Also tested
in CI: https://github.com/Yicong-Huang/spark/actions/runs/22465733224
### Was this patch authored or co-authored using generative AI tooling?
No.
Closes #54525 from
Yicong-Huang/SPARK-55405/fix/macos26-arm-float-int-cast-overrides.
Authored-by: Yicong-Huang <[email protected]>
Signed-off-by: Ruifeng Zheng <[email protected]>
---
.../upstream/pyarrow/test_pyarrow_array_cast.py | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/python/pyspark/tests/upstream/pyarrow/test_pyarrow_array_cast.py
b/python/pyspark/tests/upstream/pyarrow/test_pyarrow_array_cast.py
index b182e254b90f..6242a6d784da 100644
--- a/python/pyspark/tests/upstream/pyarrow/test_pyarrow_array_cast.py
+++ b/python/pyspark/tests/upstream/pyarrow/test_pyarrow_array_cast.py
@@ -587,6 +587,27 @@ class PyArrowScalarTypeCastTests(_PyArrowCastTestBase):
}
)
+ if platform.system() == "Darwin":
+ # macOS ARM differs from Linux ARM for some unsafe casts due to
+ # differences in LLVM code generation between the two
platforms.
+ overrides.update(
+ {
+ # negative float → uint8/uint16: macOS ARM wraps
(255/65535),
+ # Linux ARM saturates to 0
+ ("float16:standard", "uint8"): "[0, 1, 255,
None]@uint8",
+ ("float16:standard", "uint16"): "[0, 1, 65535,
None]@uint16",
+ ("float32:standard", "uint8"): "[0, 1, 255,
None]@uint8",
+ ("float64:standard", "uint8"): "[0, 1, 255,
None]@uint8",
+ ("float64:standard", "uint16"): "[0, 1, 65535,
None]@uint16",
+ # negative float → uint32: macOS ARM saturates to 0,
+ # Linux ARM wraps to 4294967295 (matching x86 golden)
+ ("float64:standard", "uint32"): "[0, 1, 0,
None]@uint32",
+ # special float → int32: macOS ARM saturates
(INT32_MAX/MIN),
+ # Linux ARM gives -1/0
+ ("float64:special", "int32"): "[2147483647,
-2147483648, 0, None]@int32",
+ }
+ )
+
return overrides
# ----- test methods -----
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
