[talumbau]

JIRA ID: AIHPBLAS-5250

Reverts #14195

Motivation

#14195 ("fix(hipblaslt): order dict-format solution keys after merge and geko dump") imposed a canonical key order on dict-format library logic written by the TensileLite merge and by the geko library writer. This PR backs that change out of develop so the previous emit order is restored while the ordering work is re-evaluated under AIHPBLAS-5250.

A full revert is preferred over a forward fix here because the change only affected the serialization order of keys that are semantically unordered, so restoring the prior output is the lowest-risk way to unblock consumers of these YAML files.

Technical Details

Clean revert of commit 043a49e — no conflict resolution, nothing partially retained. Four files, 1 insertion / 178 deletions:

• projects/hipblaslt/tensilelite/tensilelite/TensileMergeLibrary.py — drops the LibraryIO.reorderSolutionDictForDictMerge call from avoidRegressions, so merged solutions are written back in the order reNameSolutions / removeDefaultInitParams leave them in.
• projects/hipblaslt/utilities/geko/geko/library/library.py — drops the local _reorder_dict_solution helper and its use in Library.dump, so libs, final_libs, and full_libs again emit plain insertion order.
• projects/hipblaslt/tensilelite/tensilelite/Tests/unit/test_TensileMergeLibrary.py and projects/hipblaslt/utilities/geko/tests/test_library_unit.py — remove the unit tests added with the reverted change (test_merge_writes_dict_solutions_in_canonical_order, test_dict_dump_orders_appended_solution_keys, and the ordering assertions in test_dict_library_supports_property_access_and_dump).

No public API, kernel, or runtime behavior is touched. The only observable difference is the key order inside generated dict-format logic YAML.

Test Plan

• Verify the revert is an exact inverse of 043a49e rather than a hand-edited approximation.
• Rely on the repository's unit-test policy check over the TensileLite and geko unit suites to confirm nothing depends on the removed reordering helpers.

Test Result

• Verified exact: the diff of this branch against its parent is, per file, the exact line-for-line inverse of the diff of 043a49e against its parent (179 changed lines on each side, matching multisets).
• The 🧪 Unit Test policy check reports Pass on this PR, so no remaining test references reorderSolutionDictForDictMerge or _reorder_dict_solution.

🤖 Generated with Claude Code

[therock-pr-bot]

✅ All Checks Passed — Ready for Review

Check	Status	Details
📝 PR Description	✅ Pass	—
⛔ Forbidden Files	✅ Pass	—
🧪 Unit Test	✅ Pass	—
🔎 pre-commit	✅ Pass	—
🚫 Draft PR	🔜 To Be Enabled	—
🚩 Feature Flag	🔜 To Be Enabled	—
📊 Code Coverage	🔜 To Be Enabled	—
🤖 therock-pr-bot	✅ Pass	—

🎉 All checks passed! This PR is ready for review.

📖 Need help? See the Policy FAQ for details on every check and how to fix failures.

🙋 Wish to Override Policy?

[therock-pr-bot]

🎉 All checks passed! This PR is ready for review.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop   #14565      +/-   ##
===========================================
- Coverage    77.29%   77.29%   -0.00%     
===========================================
  Files         2889     2889              
  Lines       484853   484852       -1     
  Branches     73234    73234              
===========================================
- Hits        374740   374739       -1     
  Misses       85381    85381              
  Partials     24732    24732              

Flag	Coverage Δ		*Carryforward flag
TensileLite-CPP	50.81% <ø> (ø)
TensileLite-Unit	84.02% <ø> (-<0.01%)	⬇️
hipBLAS	90.62% <ø> (ø)		Carriedforward from 03a05bd
hipBLASLt	68.99% <ø> (ø)
hipCUB	82.68% <ø> (ø)		Carriedforward from 03a05bd
hipDNN	86.86% <ø> (ø)		Carriedforward from 03a05bd
hipFFT	38.02% <ø> (ø)		Carriedforward from 03a05bd
hipRAND	76.12% <ø> (ø)		Carriedforward from 03a05bd
hipSOLVER	69.67% <ø> (ø)		Carriedforward from 03a05bd
hipSPARSE	87.03% <ø> (ø)		Carriedforward from 03a05bd
rocBLAS	68.95% <ø> (ø)		Carriedforward from 03a05bd
rocFFT	47.34% <ø> (ø)		Carriedforward from 03a05bd
rocRAND	57.31% <ø> (ø)		Carriedforward from 03a05bd
rocSOLVER	76.74% <ø> (ø)		Carriedforward from 03a05bd
rocSPARSE	83.54% <ø> (ø)		Carriedforward from 03a05bd
rocThrust	93.15% <ø> (ø)		Carriedforward from 03a05bd

*This pull request uses carry forward flags. Click here to find out more.

Files with missing lines	Coverage Δ
...slt/tensilelite/tensilelite/TensileMergeLibrary.py	97.13% <ø> (-0.01%)	⬇️

... and 2 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.

[assistant-librarian]

TheRock Submodule Bump Activity

Newest first

• 2026-10-30 — Included in TheRock via ROCm/TheRock#9594.