Compute client storage dependents once per hydrate

[FarhanAliRaza]

Fixes #7508. A smaller alternative to #7515.

Problem

hydrate_and_load sends the snapshot first. Then it passes the browser's client storage values through get_delta overrides again. It marked those values dirty with _mark_dirty. That invalidated their computed dependents, so each dependent was evaluated a second time. The snapshot had already sent them, computed from the same values. In reflex-local-auth, this runs the authenticated_user database query twice on every page load.

Fix

The second pass now builds a delta that holds only the dirty fields. While _send_fields_only is set, build_delta does not invalidate or send computed vars. The pass runs inside the handler, so the flag does not reach any other delta.

The behavior stays the same as on main:

• Overrides see each browser value once.
• What an override returns still reaches browser storage.
• An override still changes only the value sent to the browser, as on every other event.

Testing

A new test, test_hydrate_computes_client_storage_dependents_once, fails on main and passes with the fix. It checks a cached dependent in the same state and a cached dependent in a sibling state. It also checks that an uncached var is not evaluated more often than in a hydrate without stored values.

I also ran an example app in Chrome against main, with stored values set. The app has a cached dependent, a sibling dependent, an uncached var, an override that rewrites a value, and an override that drops a key. Per reload:

	main	this PR
Cached dependent, same state	2	1
Cached dependent, sibling state	2	1
Uncached var	3	3
Override calls	1	1

The displayed values, the stored values and the backend values stay the same as on main.

The pass now sends its own delta, so a hydrate with stored values sends one more websocket message than on main. This PR does not change that.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-14T18:41:08.599559Z	6a7213d	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[cubic-dev-ai]

All reported issues were addressed across 4 files

<sub>Reply with feedback, questions, or to request a fix.

View guided diff | Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6a7213d24b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[codspeed]

Merging this PR will not alter performance

⚠️ 9 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

• test_cli_startup[run_help]
• test_cli_startup[cloud_help]
• test_cli_startup[deploy_help]
• test_import_cli
• test_cli_startup[version]
• test_cli_startup[help]
• test_set_state[disk]
• test_modify_state[disk]
• test_python_process_startup

✅ 148 untouched benchmarks

---

_{Comparing FarhanAliRaza:farhan/hydrate-storage-dependents (e90bff3) with main (6efb227)¹}

Footnotes

1. No successful run was found on main (2fc4a9c) during the generation of this report, so 6efb227 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable issues remain.

Summary

The PR avoids computing client-storage dependents again after sending the hydration snapshot.

• Hydration evaluates client-storage dependents once.

The previous unused-set finding is fixed. The normal build_delta path now creates delta_vars once.

_{Reviews (3) · Last reviewed commit: "Merge remote-tracking branch 'upstream/m..." · Reviewed by Greptile}

[masenf]

nice fix, thanks for simplifying