Skip to content

chore: roll Playwright to 1.66.0-alpha-1793916328000 - #523

Merged
Pavel Feldman (pavelfeldman) merged 1 commit into
microsoft:mainfrom
pavelfeldman:roll_166
Nov 5, 2026
Merged

Pavel Feldman (pavelfeldman) merged 1 commit into
microsoft:mainfrom
pavelfeldman:roll_166

Conversation

@pavelfeldman

Copy link
Copy Markdown
Member

Summary

  • Roll playwright, playwright-core and @playwright/test to 1.66.0-alpha-1793916328000.
  • Parse the arguments once with parseArgs() and pass them to program(), the embedder contract from feat(cli): let embedders reuse parsed arguments playwright#43941. The update check reads the command from the parsed arguments, so a global option with a value before the command (--session foo list) no longer hides it.
  • Regenerate the skill: running-code.md notes that event listener errors are reported as console errors.

Parse the arguments once with parseArgs() and pass them to program(), as
microsoft/playwright#43941 allows. The update check now reads the command
from the parsed arguments, so a global option with a value before the
command no longer hides it.
@pavelfeldman
Pavel Feldman (pavelfeldman) merged commit 745fbc7 into microsoft:main Nov 5, 2026
4 checks passed
Pavel Feldman (pavelfeldman) pushed a commit that referenced this pull request Nov 5, 2026
The update and skill checks took the first non-option token as the
command, so `-s foo list` skipped them and `-s list snapshot` ran them.
Since #523 the command comes from the parsed arguments. Document the
space-separated session forms and extend the update-check and
skill-warning tests.

Fixes: #505
Pavel Feldman (pavelfeldman) added a commit that referenced this pull request Nov 5, 2026
## Summary
- The entry point discarded the promise from `program()`, so argument
parsing, startup and daemon validation failures surfaced as unhandled
rejections with a stack trace, and `--json` produced no output.
- Await the program and report a failure as the message on stderr, or `{
"isError": true, "error": "..." }` on stdout in JSON mode, with exit
code 1. The output mode comes from the arguments parsed with the core's
`parseArgs()` (#523), read before `program()` runs; when parsing itself
fails, it falls back to the `--json` flag.
- Extend the open-data-URL test with the text and JSON parse error and a
daemon-side validation error in JSON mode, and document the behavior.

Fixes #502

---------

Co-authored-by: microsoft-playwright-automation[bot] <203992400+microsoft-playwright-automation[bot]@users.noreply.github.com>
Co-authored-by: Pavel Feldman <pavel.feldman@gmail.com>
Pavel Feldman (pavelfeldman) pushed a commit that referenced this pull request Nov 5, 2026
## Summary
- The update and skill checks took the first non-option token as the
command, so `-s foo list` skipped them and `-s list snapshot` ran them.
Since #523 the command comes from the arguments parsed with the core's
`parseArgs()`, which fixes this.
- Document the `-s name` and `--session name` forms, and extend the
update-check and skill-warning tests: a session before the command warns
about a stale skill and checks for updates, `-s list snapshot` makes no
registry request and writes no cache, and `--session list install
--skills` keeps the install exemption.

Fixes #505

Co-authored-by: microsoft-playwright-automation[bot] <203992400+microsoft-playwright-automation[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant