Fix to_edge_transform_and_lower to respect preserve_ops without parti…

[psiddh]

…tioner

Summary:
to_edge_transform_and_lower silently ignores EdgeCompileConfig.preserve_ops when no partitioner is provided — ops like aten.linear.default get decomposed to addmm despite being listed in preserve_ops. This is inconsistent with to_edge, which correctly pops preserved ops from the decomposition table. The fix adds 3 lines to _gen_edge_manager_for_partitioners to pop config.preserve_ops from the table in the no-partitioner path, matching the existing to_edge behavior. When preserve_ops is empty (the default for all existing callers), the behavior is unchanged.

Test:
python3 -m unittest exir.program.test.test_program.TestProgramManagers.test_to_edge_transform_and_lower_with_preserve_ops_no_partitioner -v

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17493

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 1 Unrelated Failure

As of commit eebc204 with merge base a24d3e7 ():

NEW FAILURE - The following job has failed:

• Test CUDA Builds / test-model-cuda-e2e (mistralai, Voxtral-Mini-3B-2507, quantized-int4-tile-packed) / linux-job (gh)
  RuntimeError: Command docker exec -t c2e5b57728840be6bed22e883992e215d0b427afd0c448fb55fd35e58cb11ebc /exec failed with exit code 1

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• pull / test-coreml-bc-macos (macos-m2-stable) / macos-job (gh) (matched macos rule in flaky-rules.json)
  File doesn't exist

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

Fixes to_edge_transform_and_lower so it respects EdgeCompileConfig.preserve_ops even when no partitioner is provided, aligning behavior with to_edge and preventing unintended decompositions (e.g., aten.linear.default → addmm).

Changes:

• Update _gen_edge_manager_for_partitioners to remove config.preserve_ops entries from the default decomposition table in the no-partitioner path.
• Add a regression test ensuring preserve_ops prevents decomposition in to_edge_transform_and_lower without a partitioner.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
exir/program/_program.py	Pops config.preserve_ops from the decomposition table when partitioners_for_program is empty, preventing unwanted decompositions.
exir/program/test/test_program.py	Adds a unit test covering preserve_ops behavior in to_edge_transform_and_lower with no partitioner.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

This test duplicates a count_nodes helper that already exists in _test_to_edge_with_preserved_ops above. Consider extracting count_nodes into a shared helper (e.g., a private method on the test class) to avoid repeating the same graph-walk logic across tests.

[psiddh]

Good suggestion! Will do this in follow up PR..

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[psiddh]

cc @rascani @AdrianLundell
This was the reason why aten.linear.default was decomposing to addmm. I will update the PR : #17075 accordingly removing the addmm impl from it (and extra pass that was added as result)