workflow: forward x-ip-token so zerogpu spaces bill the caller

[hannahblair]

Description

@AK391's issue around busy zerogpu spaces was caused because the x-ip-token header wasn't reaching the target spaces, so ZeroGPU couldn't identify the caller and fell back to the workflow Space's shared IP quota, which was exhausted almost instantly.

the fix is to set LocalContext.request around the fn call in /component_server, matching what get_function_with_locals already does for normal event listeners. only /component_server was broken. every other user func dispatch already goes through get_function_with_locals.

[gradio-pr-bot]

🪼 branch checks and previews

•	Name	Status	URL
	Spaces	ready!	Spaces preview
	Website	ready!	Website preview
🦄	Changes	detected!	Details

---

Install Gradio from this PR

pip install https://huggingface.co/buckets/gradio/pypi-previews/resolve/da76169521194fc0722bdd05d0382964f3e0b92a/gradio-6.24.0-py3-none-any.whl

Install Gradio Python Client from this PR

pip install "gradio-client @ git+https://github.com/gradio-app/gradio@da76169521194fc0722bdd05d0382964f3e0b92a#subdirectory=client/python"

Import Gradio JS Client from this PR via CDN

import { Client } from "https://huggingface.co/buckets/gradio/npm-previews/resolve/da76169521194fc0722bdd05d0382964f3e0b92a/browser.js";

[gradio-pr-bot]

🦄 change detected

This Pull Request includes changes to the following packages.

Package	Version
gradio	minor

---

• workflow: forward x-ip-token so zerogpu spaces bill the caller

---

Something isn't right?

• Maintainers can change the version label to modify the version bump.
• If the bot has failed to detect any changes, or if this pull request needs to update multiple packages to different versions or requires a more comprehensive changelog entry, maintainers can update the changelog file directly.

[abidlabs]

LGTM @hannahblair!

[AK391]

Tested this on my workflow Space
(https://huggingface.co/spaces/akhaliq/workflow-test) using the PR preview
wheel, called the workflow via gradio_client both logged in (OAuth) and
logged out, and the ZeroGPU operator (Z-Image-Turbo) completed successfully
both times instead of failing with the busy/quota error. LGTM!

[hannahblair]

great, thanks for testing @AK391!