Public: anyone can read this; the crew can change it.

Home tells the truth in few words: figures agree with their sources, blurbs become lines #10449

Parked opened this in answer to
Closed without merging

Home's figures are now one line each and agree with their sources, but I have no before/after screenshots. Chromium can't open a socket in this sandbox (socket() failed: Operation not permitted). I rendered the page to HTML on the old and new code but couldn't capture it.

Notes

stopped parked by the crew deploy d80c95a

opened this

Home's figures are now one line each and agree with their sources, but I have no before/after screenshots. Chromium can't open a socket in this sandbox (socket() failed: Operation not permitted). I rendered the page to HTML on the old and new code but couldn't capture it. Please take the phone and desktop shots from this pull's preview.

gate.ts, tsc, design-lint and the home, week, alarm and Later-gate test files all pass. I did not run the full suite; the forge's check will.

What changed

  • One line per figure. Each is a label, a number and at most four words of context, and each links to where it comes from. The new src/flow/home_figures.ts holds the figures as plain data, so a test can check them. kpi_row.ts draws them as row() list lines in place of the tile grid.
  • Builders show running of crew.concurrency, so "of 16" is gone. The "nothing ready" context appears only when the map's ready count is 0. It links to /map; other idle reasons keep their own link.
  • Ready is a new figure, for example 49 with "7 startable", and matches the /map API's ready and startable counts. src/crew/ready_counts.ts reads it, cached once per write.
  • Spend and Burn share one window, since 00:00Z. "Burn 0% an hour" can no longer sit beside $511.
  • Week used says "behind" only when used% is below elapsed%, with a 3-point tolerance for "on pace". Elapsed% comes from crewPace in week_read.ts. The old pace paragraphs are removed from home, and the unused homeWeek is deleted. weekLine stays for the Spend page.
  • Dropped from home: Accounts, Reporting browsers, Traffic, Reported page speed and their sparklines. Their boards and the terminal still show them.
  • Shipped list: it excludes merged pulls whose issue title starts with Sentinel:, since those are alarms. Titles are cut at a word boundary at 60 characters, using the new clipAtWord in src/text/clip.ts.
  • Long blurbs: the stuck-pull alarm reads "repo#N approved, waits for you" and the "sat approved" line is shorter too. The Later-gate line reads "176 held behind Later in forge", with who set it in the link's title. This changes the same alarm text in the terminal dashboard.
  • Tests: test/home-figures.test.ts is the consistency test.

- It checks the figures against a fixture: slots equal crew.concurrency, "nothing ready" only at 0 ready, spend and burn in one window, "behind" only when used% is below elapsed%, and at most four words of context. - It also runs the page against a real app and checks ready and startable against /map. - Tests for the removed tiles are rewritten or deleted; none was loosened to pass.

Not done

  • Shipped has no live-release rows. Nothing in the repo reads releases as a list for home, so I did not add one.
  • I did not touch the "Needs you" ordering. Alarms already render there.

How to try it Sign in as the owner on the preview and open /. Open "Figures" and each line links to its source.

🤖 Generated with Claude Code

Closes #10440

2 earlier comments · 1 event
commented

Home's figures are now one line each and agree with their sources, but I have no before/after screenshots. Chromium can't open a socket in this sandbox (socket() failed: Operation not permitted). I rendered the page to HTML on the old and new code but couldn't capture it. Please take the phone and desktop shots from this pull's preview.

gate.ts, tsc, design-lint and the home, week, alarm and Later-gate test files all pass. I did not run the full suite; the forge's check will.

What changed

  • One line per figure. Each is a label, a number and at most four words of context, and each links to where it comes from. The new src/flow/home_figures.ts holds the figures as plain data, so a test can check them. kpi_row.ts draws them as row() list lines in place of the tile grid.
  • Builders show running of crew.concurrency, so "of 16" is gone. The "nothing ready" context appears only when the map's ready count is 0. It links to /map; other idle reasons keep their own link.
  • Ready is a new figure, for example 49 with "7 startable", and matches the /map API's ready and startable counts. src/crew/ready_counts.ts reads it, cached once per write.
  • Spend and Burn share one window, since 00:00Z. "Burn 0% an hour" can no longer sit beside $511.
  • Week used says "behind" only when used% is below elapsed%, with a 3-point tolerance for "on pace". Elapsed% comes from crewPace in week_read.ts. The old pace paragraphs are removed from home, and the unused homeWeek is deleted. weekLine stays for the Spend page.
  • Dropped from home: Accounts, Reporting browsers, Traffic, Reported page speed and their sparklines. Their boards and the terminal still show them.
  • Shipped list: it excludes merged pulls whose issue title starts with Sentinel:, since those are alarms. Titles are cut at a word boundary at 60 characters, using the new clipAtWord in src/text/clip.ts.
  • Long blurbs: the stuck-pull alarm reads "repo#N approved, waits for you" and the "sat approved" line is shorter too. The Later-gate line reads "176 held behind Later in forge", with who set it in the link's title. This changes the same alarm text in the terminal dashboard.
  • Tests: test/home-figures.test.ts is the consistency test.

- It checks the figures against a fixture: slots equal crew.concurrency, "nothing ready" only at 0 ready, spend and burn in one window, "behind" only when used% is below elapsed%, and at most four words of context. - It also runs the page against a real app and checks ready and startable against /map. - Tests for the removed tiles are rewritten or deleted; none was loosened to pass.

Not done

  • Shipped has no live-release rows. Nothing in the repo reads releases as a list for home, so I did not add one.
  • I did not touch the "Needs you" ordering. Alarms already render there.

How to try it Sign in as the owner on the preview and open /. Open "Figures" and each line links to its source.

🤖 Generated with Claude Code

the crew's run was interrupted: restart, resumes on the next pass — cancelled
commented

The preview deploy didn't fail on my code. The log you sent ends with the forge draining on SIGTERM (a restart), and its only complaint was "no tests declared, none run", which isn't a test failure. The preview answers 200 now.

The working tree is clean and HEAD (d80c95a1b) is already on the remote branch. git push reports "Everything up-to-date", so I made no new changes.

What the pull does: Home's figures (Builders, Ready, Spend, Burn, Week used) are one line each and match the pages they come from. The shipped list no longer shows alarms and cuts titles at a word boundary.

What I chose not to do: I have no before/after screenshots. Chromium can't open a socket in my sandbox. Please take the phone and desktop shots from the preview at https://lab--10449--forge-trevato-apps.open-platform.sh.

How to try it: Sign in as the owner and open /. The Figures list reads one line per figure, and each line opens its source.

🤖 Generated with Claude Code

commented

Parked by the crew: the check stayed red through 2 automatic tries: the releases page > the home links the live release

Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs