[archana-ramalingam]

Summary

For test_PlaceholderMerge.py:

• In math-ci, the logic files are available, so all tests pass as expected because the xfail condition depends on the presence of those files.
• In TheRock CI, a sparse checkout omits these files, so the xfail condition is triggered for all tests.
  However, 3 of the 5 tests do not rely on the logic files, resulting in unexpected passes (xpass), which in turn cause CI failures.

Technical details

The module-level pytestmark = pytest.mark.xfail(not _LOGIC_ROOT.is_dir(), ...) applied to ALL 5 tests in the module.
However, 3 / 5 tests don't need logic files and pass regardless — causing XPASS(strict) failures in TheRock CI (unexpected pass treated as failure by xfail_strict=True in pytest.ini).

Fix

• Define a reusable _needs_logic_dir xfail marker to verify if logic files exist.
• Apply @_needs_logic_dir only to the 2 tests that need logic files.
• The other 3 tests run and pass normally without any logic file related markers.

Test plan

• TheRock CI passes (no XPASS failures)
• Source-tree tests still pass (pytest Tensile/Tests/unit/test_PlaceholderMerge.py -v)

Ref: #7481

AIGECORE-1997

[talumbau]

For clarity - my understanding is as follows:

In some cases, we run these Placeholder merge tests in an environment where the logic files directories exist, in some cases we have a test environment where they do not exist. We want to xfail them since in rocm-libraries standard CI, the directories do not exist. However, in TheRock artifacts, they do exist, Xfailed test -> Passes == CI failure with XPASS(strict). Thus this PR conditionally sets the xfail. Thanks @archana-ramalingam

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

❌ Your project status has failed because the head coverage (77.83%) is below the target coverage (80.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #7716      +/-   ##
===========================================
- Coverage    61.95%   61.87%   -0.08%     
===========================================
  Files         2086     2086              
  Lines       357056   357253     +197     
  Branches     53784    53858      +74     
===========================================
- Hits        221180   221024     -156     
- Misses      117067   117426     +359     
+ Partials     18809    18803       -6     

Flag	Coverage Δ		*Carryforward flag
TensileLite	26.06% <ø> (-0.39%)	⬇️
hipBLAS	90.65% <ø> (ø)		Carriedforward from 6824cec
hipBLASLt	41.27% <ø> (ø)
hipCUB	82.68% <ø> (ø)		Carriedforward from 6824cec
hipDNN	85.87% <ø> (ø)		Carriedforward from 6824cec
hipFFT	50.00% <ø> (ø)		Carriedforward from 6824cec
hipRAND	76.12% <ø> (ø)		Carriedforward from 6824cec
hipSOLVER	69.24% <ø> (ø)		Carriedforward from 6824cec
hipSPARSE	85.09% <ø> (ø)		Carriedforward from 6824cec
rocBLAS	48.11% <ø> (ø)		Carriedforward from 6824cec
rocFFT	52.07% <ø> (ø)		Carriedforward from 6824cec
rocRAND	57.04% <ø> (ø)		Carriedforward from 6824cec
rocSOLVER	77.83% <ø> (ø)		Carriedforward from 6824cec
rocSPARSE	72.68% <ø> (ø)		Carriedforward from 6824cec

*This pull request uses carry forward flags. Click here to find out more.
see 9 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.