Public: anyone can read this; the crew can change it.
The roadmap list is keyboard-first: j/k move between rows, enter opens one #10424
The roadmap page now joins the shell's j/k row traversal: j and k move between rows, and enter opens the one you're on. I didn't open the page in a running app. test/keys-roadmap.test.ts and test/roadmap.test.ts pass, bun x tsc --noEmit is clean, and bun scripts/design-lint.ts is clean.
stopped parked by the crew deploy 0bf5af3
3 earlier comments · 1 event
The change is small and matches the pattern, but j/k can stall on the folded pillars. One fix first: make j/k skip rows inside a closed <details>. Also add a line on j/k and enter to docs/features/roadmap.md (not README).
The roadmap's j/k walk now skips rows inside a closed <details>, so it no longer stalls at the end of the open pillar. I haven't run the full suite; only the 16 test/keys-*.test.ts files plus client-list ran, and they pass. bun x tsc --noEmit and bun scripts/gate.ts both ran clean.
What changed
- Fix: in
src/web/client.ts,rows()now drops any[data-nav]row whose enclosing<details>(at any nesting depth) is closed. This is in the shared shell script, so every page's j/k walk benefits, not just the roadmap. Opening a pillar by mouse or keyboard adds its rows back to the walk. I took your first option, the runtime skip, rather than markingdata-navonly in the open pillar, because the second would leave a pillar the reader just opened unreachable. - Tests (
test/keys-roadmap.test.ts): one test extracts the realfolded()/rows()source fromshellJsand runs it on a fake DOM with an open pillar and a folded one. It checks that the folded rows are excluded. A second test checks that a non-current pillar's rows are still rendered asdata-nav(the skip happens at runtime), so a regression in either place fails. - Docs: added a "Keyboard" section on j/k and enter to
docs/features/roadmap.md, not the README.
What I left alone
- The tests use a fake DOM, not a browser, so focus behaviour on a real closed
<details>is not exercised. scripts/fast.tsnamed 1392 test files, which is too broad to run here, so I picked the key-traversal tests that cover the change. The forge's check runs the rest on this push.
To try it: open /roadmap with several pillars, press j repeatedly past the last row of the open pillar (focus should not stick), then open another pillar and press j again.
I merged main into the branch and pushed. The "every module under src/ is imported by another" failure comes from main, not from this pull, so the check will probably fail on it again.
- Unreached module:
src/kernel/why_now.tsis imported by nothing undersrc/. Commit54dacba00onmainretired the direction that used it. The same test fails on a clean checkout oforigin/main, and this pull doesn't touch that module. I left it alone, because deleting or allow-listing it belongs with whoever retired it. - Idle-preview sleep test: "a pull touched an hour ago keeps its preview…" passed when I ran
test/apps-sleep.test.tshere, so I treated it as a flake. I did not find a cause for the failure in the forge's run. - Checks here: after the merge,
bun x tsc --noEmitis clean andtest/keys-roadmap.test.tspasses. I did not re-run the sleep test after the merge. - Push: I merged rather than rebased because force pushes are refused.
This pull makes the roadmap list keyboard-first: j/k move between linked rows, skipping rows in a folded pillar. The keys are documented in docs/features/roadmap.md. To try it, open /roadmap, press j then k to move between rows and Enter to open one.
This pull makes j/k and the other list keys work on the roadmap. Roadmap rows, including the "needs you" rows, now carry data-nav. The walk skips rows inside a folded pillar, since a closed <details> can't take focus. The keys are documented in docs/features/roadmap.md, with tests in test/keys-roadmap.test.ts.
The check failed twice. I fixed the first failure and did not fix the second.
- Fixed: "every module under src/ is imported by another, or allowed" failed because
src/kernel/why_now.tswas orphaned. The earlier commit that retired the "why now" direction (54dacba00) left that module andtest/why-now.test.tsbehind, and main has the same orphan. I deleted both.test/unreached.test.ts,test/keys-roadmap.test.ts,bun x tsc --noEmitandbun scripts/gate.tspass here. - Not fixed: the "looks of a public app" preview test timed out at about 90 s in one shard. It is a browser-screenshot test that needs the preview, and I can't run it here. My changes don't touch any route that test covers, so I think it was a slow shard, but I haven't confirmed that. The re-run on this push will show whether it fails again.
Parked by the crew: the check stayed red through 2 automatic tries
The roadmap page now joins the shell's j/k row traversal: j and k move between rows, and enter opens the one you're on. I didn't open the page in a running app.
test/keys-roadmap.test.tsandtest/roadmap.test.tspass,bun x tsc --noEmitis clean, andbun scripts/design-lint.tsis clean.src/web/pages/roadmap.ts, the "Needs you" rows and the issue rows inside each pillar now carrynav: true.<summary>fold toggles with no link, so enter would do nothing on them. The empty page carries nodata-nav.test/keys-roadmap.test.tsrendersroadmapBodydirectly.- One case seeds a needs-you row, an issue with a pull and two plain issues. It asserts exactly four
data-navrows, that the pull-linked row points at the pull, and that no<summary>hasdata-nav. - The other case checks that an empty roadmap and a pillar with no issues draw nodata-nav. The first case fails without the change./roadmapon the preview, expand a pillar, press j then k, and press enter to open the focused row.Closes #10406