Fast-path framework bookkeeping fields in state attribute access

[Alek99]

Linear: ENG-10096

Description

_get_attribute was ~43% of profiled event-handling time, and much of that was the framework's own reads of dirty_vars, parent_state, substates, and _backend_vars paying the full inherited-vars/event-handler/proxying chain on every access from _mark_dirty/_clean/get_delta.

• Add those names (plus dirty_substates) to the CLASS_VAR_NAMES fast path. They are never state vars, never inherited, and were never proxied (not in base_vars/backend_vars), so resolution via object.__getattribute__ is behavior-identical; assignments already ended in object.__setattr__ with no dirty-marking.
• Reorder the proxying condition so base_vars/backend_vars membership is checked before is_mutable_type, sparing non-var values the type check.
• Cache get_skip_vars() per class as a frozenset, invalidated when inherited_vars changes (_update_substate_inherited_vars, add_var).
• Cache the frontend computed var names used by get_delta per class, invalidated when computed_vars changes (_evaluate, dynamic route vars).

Benchmarks (GitHub Actions runner, run, 2 passes each)

Case	main	this PR	speedup
4 bookkeeping attr reads	3.66 / 3.56 us	0.61 / 0.66 us	~5.7x
setattr + get_delta + _clean cycle	84.4 / 71.4 us	54.5 / 51.4 us	~1.4x
get_skip_vars()	0.49 / 0.48 us	0.15 / 0.15 us	~3.3x

cProfile (5000 bookkeeping reads + 500 event cycles): _get_attribute tottime drops 0.120s -> 0.061s; total function calls drop 256k -> 207k.

Type of change

• Bug fix (non-breaking change which fixes an issue)

Changes To Core Features:

• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your core changes, as applicable?
  • test_get_skip_vars_cached_per_class: per-class caching with inherited vars included and no cross-class leakage.
  • test_frontend_computed_var_names_cached_per_class: backend computed vars excluded; cached identity.
  • test_framework_bookkeeping_fields_not_proxied: bookkeeping containers stay raw while real state vars still get proxied.
• Have you successfully ran tests with your changes locally?

🤖 Generated with Claude Code

https://claude.ai/code/session_01G8cXh3TUbNtbjm2ERqE62X

---

Generated by Claude Code

[linear-code]

ENG-10096

[codspeed-hq]

Merging this PR will not alter performance

⚡ 1 improved benchmark
❌ 1 regressed benchmark
✅ 24 untouched benchmarks
⏩ 8 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	test_var_access[non_mutable_scalar]	57.7 ms	65.9 ms	-12.41%
⚡	test_process_event	7.6 ms	6.6 ms	+14.32%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing claude/reflex-perf-optimizations-01l7a3-eng-10096 (97e9e37) with main (48f9e6c)}

Footnotes

1. 8 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[greptile-apps]

Greptile Summary

This PR speeds up access to internal state bookkeeping. The main changes are:

• Routes framework bookkeeping fields through direct attribute access.
• Caches serialization skip names and frontend computed-variable names per class.
• Reorders mutable-proxy checks to avoid unnecessary type inspection.
• Adds unit coverage and a performance changelog entry.

Confidence Score: 4/5

Descendant skip-cache invalidation needs to be fixed before merging.

• The bookkeeping fast path preserves the current internal field behavior.
• Dynamic variable registration can leave a grandchild with stale serialization metadata.
• Existing tests do not cover an already-cached grandchild when an ancestor gains a variable.

reflex/state.py

Important Files Changed

Filename	Overview
reflex/state.py	Adds bookkeeping fast paths and per-class caches; dynamic variable registration can leave a descendant skip cache stale.
tests/units/test_state.py	Covers cache identity and raw bookkeeping access, but not ancestor variable registration after a grandchild cache is populated.
news/6757.performance.md	Documents the state attribute and serialization-cache performance changes.

_{Reviews (1): Last reviewed commit: "Add changelog fragment" | Re-trigger Greptile}

[greptile-apps]

Grandchild Skip Cache Stays Stale

When add_var runs after a grandchild has cached get_skip_vars(), this loop clears only the direct substates while the new variable remains visible farther down through the aliased vars and inherited_vars dictionaries. The grandchild can then serialize the inherited variable as local state because its cached skip set was not invalidated.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 483d2d1253

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".

[chatgpt-codex-connector]

Recursively invalidate skip-var caches for dynamic vars

When add_var() is called on a state that already has nested substates, this clears the cache only for immediate children. Grandchildren have their inherited_vars updated indirectly because it aliases the child vars dict, but their _skip_var_names frozenset was already populated during class creation and remains stale, so Grandchild.get_skip_vars() omits the newly inherited dynamic var. Please invalidate descendants as well when propagating a dynamically added var.

Useful? React with 👍 / 👎.