Four UX threads stitched into one commit. Each pulls Studio
toward Stefan's "zero-learning-curve" goal — the feedback
that earned its own memory entry.
1. Flow-Runnability-Indikator
─────────────────────────
The Flows tab now fetches `listFlows` and `listModules`
in parallel. Each card compares the flow's
`requiredCapabilities` against the installed-modules'
capability set; rows with missing modules show a "Needs:
text.extract@^0" red pill row beneath the path and have
their Run button greyed out + tooltip
"Install the missing modules first." Operators stop
hitting Run → cryptic hub error → frustration.
2. Welcome-Checklist Celebration
─────────────────────────────
Once all four checklist signals flip to done, an
`_AllDoneCelebration` card replaces the bare
"All four steps complete" + Hide button. Three concrete
next-threads with action buttons: "Read the audit log",
"Set up the daily Today story" (opens the Flows / Today
doc inline via `_DocReaderSheet`), and "Build your own
module" (opens the architecture doc). Operator who just
got set up sees what to do next instead of an empty
"what now?" feeling.
3. Audit-Page Time-Bucket Headers + Flow-Run Detail
────────────────────────────────────────────────
The flat event list grows tiny "TODAY / YESTERDAY /
EARLIER THIS WEEK / OLDER" section headers — bucket is
computed in the operator's local timezone so an event at
23:55 yesterday in Berlin doesn't end up in "today"
because UTC happened to spill into a new day.
Plus: the event-detail dialog gains a "View flow run"
action when the picked event has a `flow_execution`. It
opens a drill-down that lists every event in the
already-fetched 100-event window sharing the same
execution id, sorted ascending — the operator reads the
run from step.started top to flow.completed bottom.
4. Approvals-Batch-Aktionen
────────────────────────
Each pending approval card grows a checkbox. When ≥1
selected, a floating action bar appears at the bottom
with "N selected · Select all · Clear · Reject all ·
Approve all". The parent loops sequentially through the
per-record SDK calls so a partial failure produces
"X done, Y failed" instead of a confusing all-or-nothing
rollback. Reject prompts for a reason once and applies
to the whole picked set.
13 new ARB keys cover the strings the four features
needed. Studio's tests stay green.
Signed-off-by: flemming-it <stefan.a.flemming@googlemail.com>
2002486828