Unbreak build-benchmark-app

[swolchok]

We delete the headers after symlinking them into frameworks. whoops!

[swolchok]

Stack from ghstack (oldest at bottom):

• -> Unbreak build-benchmark-app #8388

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/8388

• 📄 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, 4 Pending

As of commit 3a1600b with merge base d92a763 ():

NEW FAILURE - The following job has failed:

• pull / test-phi-3-mini-runner-linux / linux-job (gh)
  RuntimeError: Command docker exec -t ca86613eaf3dd010f7ff9826461996fa8210f60359c0aa95daae093cc3ee273e /exec failed with exit code 1

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

[shoumikhin]

iirc, here's what worked for me:
ln -s executorch/runtime/core/portable_type/c10 "$HEADERS_PATH/c10"

[swolchok]

I would expect that we are likely to accidentally produce a framework containing a broken symbolic link. I would like to retreat to safety given the amount of breakage we are seeing currently. I don't expect saving disk I/O for a couple dozen text files to be a significant win in 2025.

[shoumikhin]

Happy to unblock, but curious what the root cause is?
Symlinks merely point to some location, and we do have all the headers inside .xcframework, so I'd guess it's all about composing the symlink to point to the eventual headers location?

[swolchok]

the root cause

• we make a symlink (the patched line)
• we delete the linked-to-files
• we attempt to copy the linked-to-files later, but they are gone

I agree that we could get it working with a symlink, I just don't think it is an impactful thing to spend time on.