chore: refresh the vendored console to 1bb77aa24514 (#3528 follow-through)

[os-zhuang]

Ships the #3528 fix to actual users. Without this, both merged halves are invisible.

Why this is needed

@objectstack/console publishes a prebuilt objectui SPA pinned by .objectui-sha: the release pipeline clones objectui at that commit, builds it, and ships the dist alongside the framework. A merged frontend fix reaches nobody until the pin moves.

#3771 put a screen field's visibleWhen on the wire. objectui#2899 taught the console to honour it — in rendering and in required enforcement, which is the half that actually unblocks Submit. Right now the server is sending a field that the shipped console still cannot read.

This is not a hypothetical failure mode. It is the exact trap that made the first #3528 repro look like the bug was somewhere else entirely: the console source was already correct while /_console served an older build. #3558 documented it after hitting the same wall.

Pin: 09c6a177bb4a → 1bb77aa24514 (objectui main), carrying the #3528 console half plus 24 other frontend commits merged since the last refresh — i18n gap closures, grid summary aggregations, chart spec conformance, FLS wiring, scoped invitations. The generated changeset lists the full range.

How it was produced

The real tooling, not a hand-edit:

OBJECTUI_ROOT=… scripts/bump-objectui.sh 1bb77aa24514a9bee9014a9194cbe0c1eed8adc4 --no-commit
→ objectui pin: 09c6a177bb4a → 1bb77aa24514
→ wrote changeset console-1bb77aa24514.md (@objectstack/console: minor)

It stamped minor by itself — the range contains feat commits, and the set -o pipefail + grep -q SIGPIPE bug that used to force every refresh to patch regardless was fixed in #3558. Worth noting because recent refreshes before that fix read patch; this is the corrected behaviour, not a deliberate escalation.

Test plan

• pnpm check:console-sha → passes (No console dist at packages/console/dist — skipping SHA check; nothing in this checkout to drift against).
• The pinned commit is objectui main, whose own CI builds the console on every PR (Build & E2E runs the vite build; Bundle Analysis runs the full tsc && vite build and reported the entry at 27.9 KB against a 350 KB budget on docs: third-party developer onboarding and self-hosted deployment guides #2899). So the pin is known-buildable without re-running the ~15-minute console build here.
• End-to-end behaviour this unblocks was verified in Chromium before either half merged: trigger → the visibleWhen fields correctly stay hidden → Submit → resume 200 → flow runs to completion, 0 server errors. Details in fix(automation): put a screen field's visibleWhen on the wire (#3528) #3771.

Ordering

Merge order does not matter for safety — each half is inert rather than wrong without the other. But nothing user-visible changes until this lands, so it is the one remaining blocker on #3528 being genuinely fixed for the reporter.

The app-side fixes in hotcrm#505 (wrong CEL dialect, a required boolean with no default, and an industry enum narrower than the lead's) also only start mattering once this ships.

🤖 Generated with Claude Code

https://claude.ai/code/session_0122xvfanLmniNbsAPtg5hk3

---

Generated by Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
objectstack	Ignored		Jul 28, 2026 5:27am