Fix build_android_library.sh

[kimishpatel]

Summary

Qnn backend requires libQnnDlc. Package that in the aar file

Test plan

CI
Plus local testing for llamademo app revealed that without this change it throws error

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13695

• 📄 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

As of commit d9b55d4 with merge base cac1a71 ():

NEW FAILURE - The following job has failed:

• Build documentation / build (buck2) / Build doc (gh)
  At least one of the pre-conditions you specified did not hold

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.

[mergennachin]

@kirklandsign @kimishpatel @cccclai why we didn't our CI catch this error? Is it something we can add regression test in our CI?

[kimishpatel]

actually this one is probably not needed. I erroneously thought this lib is needed. Issue is the qnn delegate tries to load this library in "some" cases. Those cases are not valid for us, so I am not sure why that loading path exist to begin with. Maybe @haowhsu-quic knows

[kimishpatel]

actually this one is probably not needed. I erroneously thought this lib is needed. Issue is the qnn delegate tries to load this library in "some" cases. Those cases are not valid for us, so I am not sure why that loading path exist to begin with. Maybe @haowhsu-quic knows

actually this one, https://github.com/pytorch/executorch/blob/main/examples/qualcomm/utils.py#L146, is doing exactly this thing

[cccclai]

I'm still not sure why it's needed...when I just compile the binary and use adb, it seems fine without this .so. It's good to have for jit use cases anyway (most users don't use this flow)

[JacobSzwejbka]

seems like the branch history is super messed up

[kimishpatel]

seems like the branch history is super messed up

wow ok will fix

[kimishpatel]

@kirklandsign so now this pr adds the lines that you removed earlier. I think for aar build for maven we dont need to include these in executorch.aar, but anyone relying on building aar locally needs these. what do you wanna do?

[kirklandsign]

@kirklandsign so now this pr adds the lines that you removed earlier. I think for aar build for maven we dont need to include these in executorch.aar, but anyone relying on building aar locally needs these. what do you wanna do?

Thank you. I have https://github.com/pytorch/executorch/pull/13912/files to update the tutorial.

[kirklandsign]

@kirklandsign @kimishpatel @cccclai why we didn't our CI catch this error? Is it something we can add regression test in our CI?

We are building and testing the LlamaDemo app with XNNPACK AAR package only. We didn't build QNN AAR CI. Currently we are using emulator to run CI and QNN won't run.

[kirklandsign]

If we copy these files to AAR, and the user continues to use com.qualcomm.qti:qnn-runtime, it will cause duplicated .so right?

[kirklandsign]

So for local build, we need to add com.qualcomm.qti:qnn-runtime manually, per use case. See https://github.com/pytorch/executorch/pull/13912/files

[kimishpatel]

So for local build, we need to add com.qualcomm.qti:qnn-runtime manually, per use case. See https://github.com/pytorch/executorch/pull/13912/files

yeah thats better. abandoning this one