Add memory planning to oss tests

[lucylq]

Summary

add to pytest.ini, and add cmake path-finding import for OSS

Test plan

python -m unittest exir.tests.test_memory_planning

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17596

• 📄 Preview Python docs built from this PR

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

❌ 4 New Failures

As of commit ad56f3e with merge base d17ff74 ():

NEW FAILURES - The following jobs have failed:

• pull / test-openvino-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 44b1d8a28eebd01d884d83acd5b91fbd1d16be3a0305c911be35b2178c71c621 /exec failed with exit code 1
• pull / unittest-nxp-neutron / linux-job (gh)
  RuntimeError: Command docker exec -t b51abc7f071cbf360ca0cdf8b3999b3407c1d87fb358df8fd8e7e32619f248e4 /exec failed with exit code 4
• Test CUDA Builds / export-model-cuda-artifact (mistralai, Voxtral-Mini-3B-2507, quantized-int4-weight-only) / linux-job (gh)
  RuntimeError: Command docker exec -t 2c528d23926c7278a22e3b71d3b13605e21aa795d7530a17a5b912aaffb09eca /exec failed with exit code 1
• Test Metal Backend / test-model-metal-e2e (mistralai, Voxtral-Mini-4B-Realtime-2602, quantized-int4-metal) / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2

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

Enables the exir/tests/test_memory_planning.py suite to run in OSS test environments by removing its pytest ignore and adding a non-Buck shared-library discovery path for custom_ops_generated_lib.

Changes:

• Stop ignoring exir/tests/test_memory_planning.py in pytest.ini so it runs under pytest in OSS.
• Add a try/except around torch.ops.load_library to fall back to locating the built shared library via filesystem globbing when not running under Buck.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
pytest.ini	Un-ignores the memory planning test so it’s included in the OSS pytest run.
exir/tests/test_memory_planning.py	Adds fallback logic to locate/load custom_ops_generated_lib outside Buck (e.g., CMake/pip installs).

---

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

[Copilot]

The addopts section header says these are tests with missing custom_ops_generated_lib dependencies, but exir/tests/test_memory_planning.py is no longer ignored even though it still loads that library. Please update this comment (or move the test out of this section) so the ignore rationale remains accurate.

[Copilot]

If the Buck target load fails and the glob finds no matches, the module import continues without loading custom_ops_generated_lib, which will likely cause later failures in ToOutVarPass (missing out variants) with a less-direct error. Consider failing fast with a clearer error or skipping these tests when the library cannot be located/loaded.

Suggested change

	torch.ops.load_library(str(_libs[0]))
	torch.ops.load_library(str(_libs[0]))
	else:
	# Fail fast or explicitly skip if the custom ops library cannot be found.
	# This avoids obscure downstream errors (e.g., in ToOutVarPass).
	raise unittest.SkipTest(
	"Skipping memory planning tests: could not locate "
	"custom_ops_generated_lib shared library."
	)

[Copilot]

The recursive glob (**/kernels/portable/**/*custom_ops_generated_lib.*) can be expensive and is also imprecise (it may match non-shared-library artifacts and, when multiple matches exist, _libs[0] is not deterministic). Consider filtering to platform shared-library extensions, sorting the matches, and/or searching only known build/install locations to keep imports fast and deterministic.

[JacobSzwejbka]

lol. Didnt realize we had so much stuff disabled