[Regression]: evaluate() with larger results ~40% slower since 1.65

Open
#3,362 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
3/5
Estimated time
1-2 days
Newbie friendliness
62/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
python
Domain
performance

Research direction

Start in protocol/validator_primitives.py, reading t_object and the t_optional helper, and trace where SerializedValue's schema keys are iterated. Before editing, run the issue's page.evaluate benchmark script to record a baseline, then re-run it after skipping the optional-key validators when a key is absent. Also run the protocol validator tests in tests/ (the suite that round-trips validated messages) to confirm malformed messages are still rejected. Done means the benchmark is back near 1.30s with unchanged validation behaviour, including for events and initializers.

Written by the indexing model from the issue text.

Description

Last Good Version

1.64.0

First Bad Version

1.65.0

Steps to reproduce

evaluate() with a medium-sized result got about 40% slower:

import time
from playwright.sync_api import sync_playwright

JS = "Array.from({length: 200}, (_, i) => ({id: i, name: 'item ' + i, price: i * 1.5, url: 'https://example.com/' + i, tags: ['a', 'b']}))"

with sync_playwright() as p:
    page = p.chromium.launch().new_page()
    page.evaluate(JS)
    best = 1e9
    for _ in range(3):
        start = time.perf_counter()
        for _ in range(300):
            page.evaluate(JS)
        best = min(best, time.perf_counter() - start)
    print(f"{best:.2f}s")
Expected behavior

1.64.0: 1.30s

Actual behavior

1.65.2: 1.84s

Additional context

cProfile points at the protocol validator: the t_object validator in protocol/validator_primitives.py is about 60% of the profiled time. For every object node it calls the validator of every key in the schema, and SerializedValue has ~15 optional keys, so most calls are t_optional(None) for keys that aren't there. It also calls context.is_under_test() once per node.

Skipping the call when the key is absent and the validator is a t_optional (the result is None in that case anyway) brings the script above to 1.47s in my test, same output. Validated messages also include every event and initializer, so it's not only evaluate.

Environment
- Operating System: macOS 27.2
- CPU: arm64
- Browser: Chromium
- Python Version: 3.14
- Other info:
Dominant language
Python
Stars
15.1k
Forks
1.3k
Avg merge
6h 16m
Merged PRs (30d)
53

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from microsoft/playwright-python

All issues in microsoft/playwright-python

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.