Skip to content

fix(web): show the real reopen command and fix stale copy - #80

Merged
voyagi merged 2 commits into
mainfrom
fix/judge-sim-wording
Oct 25, 2026
Merged

voyagi merged 2 commits into
mainfrom
fix/judge-sim-wording

Conversation

@voyagi

@voyagi voyagi commented Oct 25, 2026

Copy link
Copy Markdown
Owner

Closes #79.

Reading the repo the way a judge would, I found four places that say something the product doesn't do.

The replay's last step showed a raw aws sesv2 put-configuration-set-sending-options call, but a lane reopens through the reopen function. It now shows the same command the operator email and docs/reopen-a-lane.md print, and the replay test checks it against reopenCommandFor. The fake replay page the built-check tests use still drew the old readout tag, so it counted words the real page doesn't. It now matches replay.astro, and the real page stays under its word cap.

The limits row now says numbers wait for more than three live runs, which is the rule in docs/measurement-protocol.md. The README opening no longer hints at an email approval step, and its layout says the agent calls Claude Sonnet 4.6 on Bedrock.

I left the design mockups alone: they're the old comps and aren't shipped.

Tests pass, and the updated replay test fails on main.

After merge: redeploy the site.

The replay's step 4 printed a raw `aws sesv2
put-configuration-set-sending-options` call, but the product reopens a
lane through its reopen function. The fixture now holds
reopenCommandFor('fothering-bulk'), the text the operator email and
docs/reopen-a-lane.md print, and the replay test pins it to that.

The planted replay in the built-check tests still drew the old lockout
tag, so it counted about 15 words the built page does not. It now
counts only the lane and its legend, as replay.astro does. The built
replay counts 234 of its 238 words.

The limits row now reads "Numbers wait for more than three live runs",
the rule docs/measurement-protocol.md states.

README: the opening says what the email carries instead of implying an
approval step, and the layout lists the five Lambda entries and says
the agent calls Claude Sonnet 4.6 on Bedrock.
@coderabbitai

coderabbitai Bot commented Oct 25, 2026 •

Copy link
Copy Markdown

Important

Review skipped

Auto reviews are limited based on label configuration.

🏷️ Required labels (at least one): 1
  • review-ready

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

Run configuration
  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: Organization UI
  • Review profile: CHILL
  • Plan: Essentials
  • Run ID: 03738fb1-d868-4013-80bf-5cf497e46646

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Comment @coderabbitai help to get the list of available commands.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 25, 2026 •

Copy link
Copy Markdown

Codex Review Summary

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

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-25T06:35:46.431651Z 01cf856 Manual request
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.

@voyagi

voyagi commented Oct 25, 2026

Copy link
Copy Markdown
Owner Author

@codex review

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Hooray!

Reviewed commit: 966bf0f0c0

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".

@voyagi-house-review voyagi-house-review Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

House review of 966bf0f, its first house review.

The PR updates the replay fixture so the reopen step shows the product's real `aws lambda invoke`
reopen command, and pins it in the test with `reopenCommandFor`. It also changes the fake replay
page's readout so it counts the same words as `replay.astro`, rewords the limits row and fixes the
README opening and layout. The copy changes match the code: the five Lambda handlers exist and the
agent pins Claude Sonnet 4.6. One small test-fidelity gap remains: the new command has a raw `<...>`
placeholder, so the fake test page no longer matches the real one in step 4.

2 findings: 2 minor. Each is a thread on this pull request, marked resolved once it is handled:

1. minor, test-gap packages/web/src/test/planted-dist.ts:195 The fake replay page drops the new command's '<ReopenFunctionName stack output>' placeholder into HTML unescaped, so parse5 reads it as an element and step 4 counts fewer words than the real page., a thread on its file

2. minor, docs-mismatch packages/web/src/fixtures/replay-incident.json:183 The replay shows the reopen command with an unfilled '<ReopenFunctionName stack output>' slot and leaves out the step that says how to fill it., a thread on its line

Not reviewed

  • packages/web/src/site/command-units.test.ts
The only change renames a local variable from `reopen` to `command`. The issue does not ask for it
and nothing else depends on it.

Comment thread packages/web/src/test/planted-dist.ts Resolved
The replay's reopen command holds a <ReopenFunctionName stack output>
slot, and nothing on the page said where its value comes from. The
email's reopen line in the replay now says to run the command with the
ReopenFunctionName stack output, the way the operator email does.

The fake replay page put the command into HTML unescaped, so the slot
parsed as a tag and its words dropped out of the count. It is escaped
now, and a test checks the command reads whole there. The built replay
counts 237 of its 238 words.
@voyagi

voyagi commented Oct 25, 2026

Copy link
Copy Markdown
Owner Author

@codex review

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Another round soon, please!

Reviewed commit: 01cf856d01

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".

@voyagi-house-review voyagi-house-review Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

House review of 01cf856, a re-review after 966bf0f, review 2 of 3.

The PR shows the product's own reopen command (`aws lambda invoke` on the reopen function) in the
replay instead of a raw SES call. It also makes the fake replay page match replay.astro and fixes
the wording in the README and the limits row. The new commits fix earlier finding 1: the fake page
now escapes the command with the file's existing `escapeHtml`, and a new test checks that the whole
command, `<ReopenFunctionName stack output>` slot included, appears as text. They partly fix finding
2: the email's reopen line now says the function name comes from the ReopenFunctionName stack
output, but the replay still leaves out the email's lookup command. I found nothing new to report.
One thing I couldn't confirm: I can't run the build, so I haven't checked that the real page is
still under its 238-word cap now that step 4 has a few more words.

No findings.

Not reviewed

  • packages/web/src/site/command-units.test.ts
Only renames a local variable from `reopen` to `command`. The issue doesn't ask for it, and the test
behaves the same.

@voyagi
voyagi merged commit 089147e into main Oct 25, 2026
5 checks passed
@voyagi
voyagi deleted the fix/judge-sim-wording branch October 25, 2026 06:54
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.

README and replay say things the product doesn't do

1 participant