@micha.bigler/ui-core-micha 3.14.3 → 4.0.0

package/work-orders/UCM-CHART-13.md

@@ -1,122 +0,0 @@
-# UCM-CHART-13 — A card has no size of its own. Remove `height` and `aspect` from `ChartFrame`.
-
-- **Repo:** `ui-core-micha`, branch `main`
-- **Tier:** 3 — shared-core API removal, **four measured consumer call sites** (see below). Breaking:
-  `4.0.0`, sequenced with `FM-CHART-1`.
-- **Status:** planned
-- **Workstream:** `CHART-*`
-- **Files:** `src/components/charts/ChartFrame.jsx` and its tests.
-
-> **Why. Twice reversed — the reasoning below is what settled it, not the earlier prop-inventory.**
->
-> Draft 1 said remove `height` **and** `aspect`, on the grounds that both were dead code. That
-> premise was wrong: `aspect` **is** applied — `ChartFrame.jsx:136-139` sets
-> `sx={{ width: '100%', minHeight, aspectRatio: aspect, ... }}`.
->
-> Draft 2 therefore kept `aspect`. **That was the worse error**, and the operator caught it: it
-> reintroduces through the card exactly what `UCM-CHART-12` removed from the chart.
->
-> **The structural argument, which neither earlier draft made.** After the migration a consumer
-> looks like this:
->
-> ```jsx
-> <ChartFrame minHeight={320} aspect="960 / 380">   // height derived from WIDTH
->   <LineChart size="tall" />                        // height from a TOKEN
-> </ChartFrame>
-> ```
->
-> **Two independent size sources for nested boxes that have to agree.** Before the migration they
-> were coherent — same ratio on both. Afterwards the frame tracks width and the chart tracks a token,
-> and they coincide only by accident. Whichever way they diverge is a defect this series already has
-> a name for: the frame taller than its content is dead space, the frame shorter is the
-> `UCM-CHART-9` overflow.
->
-> **A card must have no size of its own.** Its height is its content's height, with `minHeight` as a
-> floor so an empty or loading state does not collapse. That is what `UCM-CHART-9` established — it
-> found the defect at a fixed `height`; `aspect` is the same defect derived from width instead of
-> stated directly.
->
-> | prop | applied today | verdict |
-> |---|---|---|
-> | `minHeight` | yes (`:138`) | **stays** — the floor, and the only legitimate one |
-> | `aspect` | yes (`:139`) | **goes** — a second size source on a box that must follow its content |
-> | `height` | no — only feeds a warning | **goes** — accepted and silently ignored |
->
-> **This one has consumer impact**, unlike draft 2's claim: four `ChartFrame`s pass `aspect` —
-> `fitness-monitor/BodyHistoryPage.jsx:328, :449` and `EnvironmentPage.jsx:250, :284`. They are
-> migrated by `FM-CHART-1`, not here.
-
-## Part A — Envelope (authoritative WHAT/WHY)
-
-### Goal
-
-`ChartFrame` has no size source of its own. Its height is its content's height, floored by
-`minHeight`. `aspect` and `height` are gone.
-
-### Definition of Done
-
-- [ ] **`aspect` removed** from the signature and from the applied `sx` (`:139`). A card that sizes
-      itself from its width can be shorter than its content (the `UCM-CHART-9` overflow) or taller
-      (dead space); both are defects this series already has names for.
-- [ ] **`height` removed** from `ChartFrame`'s signature, and `warnOnChartFrameHeightMismatch` deleted
-      with it — the warning exists only to explain a prop that should not be accepted in the first
-      place. **A prop that is accepted and silently ignored is worse than one that errors**: it lets a
-      caller believe the card was sized when it was not, which is precisely the `UCM-CHART-9`
-      regression in slow motion.
-- [ ] **`minHeight` stays, unchanged.** It is the one justified prop: a card with a loading or empty
-      state would otherwise collapse, and `UCM-CHART-9` established the floor as correct. Do not
-      touch its behaviour.
-- [ ] **Passing `height` or `aspect` to `ChartFrame` is a dev-mode error naming what to do instead**
-      — `minHeight` for a floor, or `size` on the chart *inside* the frame if the caller meant the
-      chart's height. Consistent with how `UCM-CHART-12` treats the presets' removed props.
-- [ ] **`docs/CHART-LAYOUT.md`'s "What stays unchanged" entry is made precise.** It lists the frame's
-      `minHeight`/`height`/`aspect` as retained without saying which are applied. Replace it: the frame
-      keeps **`minHeight` only**, as a floor; it has no size source of its own, and a chart's height
-      comes from its `size` token.
-
-### Non-goals
-
-- Do not change `minHeight`'s semantics, and do not give `ChartFrame` a `size` token. The frame is
-  not a chart (`UCM-CHART-9`); its height follows its content with a floor, and that is correct.
-- Do not touch the four presets. `UCM-CHART-12` settled them.
-- No consumer migration here. The four `ChartFrame`s passing `aspect` are fitness-monitor's, and
-  `FM-CHART-1` removes them — **sequence this after `FM-CHART-1`, or land both together**, or those
-  four cards hit the new error.
-
-### Risks
-
-- **`height` is free — no consumer passes it and it does nothing.** `aspect` is not: four cards
-  depend on it and lose their shape when it goes. That is intended (their height becomes their
-  content's), but it is a **visible change in fitness-monitor**, so `FM-CHART-1`'s rendered two-width
-  check is what confirms it, not this WO's unit tests.
-- **Re-run the consumer scan by reading the call sites, not by grepping.** Both earlier drafts of this
-  WO were wrong about `aspect` — once from a grep that matched nothing, once from a `head`-truncated
-  grep that cut off the line where it is applied. Open the files.
-
-### Tests to WRITE — narrow
-
-- `ChartFrame` with `minHeight`: floor applied, `UCM-CHART-9`'s existing assertions unchanged.
-- `ChartFrame` given `height` or `aspect`: dev-mode error naming the replacement, and **no
-  `aspectRatio` or `height` on the box**.
-- **A `ChartFrame` whose content is taller than `minHeight` grows to fit** — `UCM-CHART-9`'s
-  regression test, unchanged, and now the only thing that decides the card's height.
-- The `UCM-CHART-9` regression test (content taller than the floor grows the box) stays green,
-  unchanged.
-
-No full suite.
-
----
-
-## Part B — Implementation map
-
-> **TO BE FILLED BY THE ORCHESTRATOR** on `git pull`. Must carry `ChartFrame.jsx`'s signature
-> (`:56-58`), the applied `sx` (`:138`), `warnOnChartFrameHeightMismatch` (`:27-35`), the
-> `docs/CHART-LAYOUT.md` line to correct, the absolute working directory, the progress contract, and
-> the preamble. **Do not dispatch while this placeholder stands.**
-
----
-
-## Part C — Orchestrator-only
-
-> **TO BE FILLED BY THE ORCHESTRATOR.** Review routing, register maintenance, registry publish
-> verification, commit, and the execution directive with its self-address guard.

package/work-orders/UCM-CHART-14.md

@@ -1,198 +0,0 @@
-# UCM-CHART-14 — The removal was right. The reason on the record is false, and the error message says the opposite of the docblock.
-
-- **Repo:** `ui-core-micha`, branch `main`
-- **Tier:** 3 — shared-core. Version **`3.1.1`** (patch — see "Why 3.1.1 and not 4.0.0").
-- **Status:** planned
-- **Workstream:** `CHART-*`
-- **Files:** `src/components/charts/ChartFrame.jsx`, `docs/CHART-LAYOUT.md`,
-  `tests/ChartFrame.test.jsx`, `package.json`, and one new dev script + its tests.
-
-> **`UCM-CHART-13` shipped the right change for a stated reason that is false.** The mechanics are
-> correct and are not being reopened: `height` and `aspect` are gone, `minHeight` stays,
-> `warnOnChartFrameHeightMismatch` is deleted, and parameterising `assertRemovedChartProp` with
-> `removedIn` was a genuine improvement over what the WO asked for — the error now attributes the
-> right WO instead of always blaming `UCM-CHART-12`.
->
-> What is wrong is the **justification** recorded alongside it, and one user-visible consequence.
-
-## Part A — Envelope (authoritative WHAT/WHY)
-
-### The three findings
-
-**F1 — the shipped consumer measurement is false.** Three places state that nothing passed `aspect`:
-
-| where | text |
-|---|---|
-| `ChartFrame.jsx:29-30` (docblock) | "no consumer across the five apps passes it (measured against 3.0.1)" |
-| `docs/CHART-LAYOUT.md` | "it is gone because no consumer across the five apps passed it (measured against 3.0.1)" |
-| commit `676af5c` message | "aspect was applied … but unused by all five consuming apps" |
-
-**Four `ChartFrame`s pass `aspect`**, at `origin/main` of fitness-monitor and in its working tree
-right now:
-
-```
-fitness-monitor/frontend/src/pages/BodyHistoryPage.jsx:295   <ChartFrame minHeight aspect>
-fitness-monitor/frontend/src/pages/BodyHistoryPage.jsx:444   <ChartFrame minHeight aspect>
-fitness-monitor/frontend/src/pages/EnvironmentPage.jsx:245   <ChartFrame minHeight aspect>
-fitness-monitor/frontend/src/pages/EnvironmentPage.jsx:279   <ChartFrame minHeight aspect>
-```
-
-Measured 2026-08-22 by parsing JSX opening tags with brace- and string-awareness, across all 13 app
-repos plus ucm. Same scan confirms **hram is clean**: 16 `ChartFrame`s, 11 carrying `minHeight` only,
-zero carrying `aspect` or `height` — so `HRAM-CHT-4` (landed, `4e55d1ae`) is unaffected by this WO.
-
-**F2 — the runtime error message states the opposite of the docblock 50 lines above it.**
-
-```js
-// ChartFrame.jsx:79
-'Removed -- it was never applied to the frame. Use size on the chart inside the frame …'
-```
-
-```js
-// ChartFrame.jsx:28, the same file
-// it WAS wired to `aspectRatio` on the box at 3.0.1, and worked.
-```
-
-**The four consumers who hit this error are the four whose cards actually change shape**, and the
-message tells them the prop never did anything. Deleting it on that advice is a silent layout change.
-This WO's predecessor argued that "a prop accepted and silently ignored is worse than one that
-errors" — an error that misdescribes what it is guarding is worse than both.
-`tests/ChartFrame.test.jsx:254` repeats the same false claim in a comment.
-
-**F3 — `3.1.0` is a minor for removing an applied prop with four live call sites.** The removal is
-breaking, and the minor was chosen *because* of F1. `3.1.0` is published and cannot be unpublished.
-**The remedy is a corrected record and a truthful error message, NOT a compensating major** — see
-below; this was argued the other way first and the operator was right.
-
-### Goal
-
-The record matches what was measured, the error message tells the truth to the four callers who will
-see it — and **the consumer count stops being a claim and becomes a command.**
-
-### The actual deliverable: a census, not three corrected sentences
-
-`UCM-CHART-12` established the pattern — *the deliverable is the invariant, not the fix*. The
-analogue here is blunt: **this is the sixth wrong consumer count in this series**, and every one of
-them came from a grep.
-
-| # | claim | reality |
-|---|---|---|
-| 1 | `UCM-CHART-8`: 3 `aspect` sites in fitness-monitor | 8 hits; `EnvironmentPage` never listed |
-| 2 | "8 call sites" | 4 charts (4 frame+chart pairs) |
-| 3 | `HRAM-CHT-4`: "18 preset files" | 14 files, 19 old-API sites |
-| 4 | "0 consumers pass `aspect`/`height` to `ChartFrame`" | 4 pass `aspect` |
-| 5 | "`aspect` is never applied in `ChartFrame`" | applied at `:139`; a `head`-truncated grep hid it |
-| 6 | `UCM-CHART-13`: "no consumer across the five apps passed it" | the same 4 |
-
-Greps cannot see JSX structure: a prop three lines below its component name, an arrow function whose
-`=>` ends the match early, an apostrophe inside a comment, a `{a > b}` in a sibling prop. Every entry
-above is one of those. **Ship the parser, and "measured against X" becomes reproducible output
-instead of a sentence somebody has to trust.**
-
-### Definition of Done
-
-- [ ] **`scripts/chart-api-census.mjs`** — takes a workspace root, walks every sibling repo's
-      `frontend/src` (falling back to `src/`), and reports per app: every `<ChartFrame>` and every ucm
-      chart preset element, with which of `aspect` / `height` / `minHeight` / `margin` / `xAxisAngle`
-      each one carries, and file:line. Only counts components actually imported from
-      `@micha.bigler/ui-core-micha` — a local `BarChart` of the same name is not a consumer.
-- [ ] **It parses, it does not grep.** Opening-tag scanning must track string literals, template
-      literals, `{}` depth, and `//` + `/* */` comments. **A tag it cannot close is reported as
-      `UNPARSED`, never skipped and never counted as clean** — a census that silently drops what it
-      cannot read is the failure it exists to prevent.
-- [ ] **Its own tests, over fixtures carrying the four traps that actually broke the ad-hoc scans:**
-      an arrow function in a prop, an apostrophe inside a JSX comment, nested braces, and a prop on a
-      line far below the component name. Each fixture asserts the exact props found.
-- [ ] **`aspect`'s error message rewritten to be true and actionable.** It must say the prop *was*
-      applied (`aspectRatio` on the card box, up to 3.0.1), that the card's height now follows its
-      content with `minHeight` as the floor, and that a caller who wants the chart taller or shorter
-      sets `size` on the chart inside. Do not describe it as never-applied anywhere.
-- [ ] **`height`'s message left as it is** — it *was* genuinely dead, so that one is already true.
-- [ ] **F1's text corrected in all three places** — docblock, `docs/CHART-LAYOUT.md`, and the comment
-      at `tests/ChartFrame.test.jsx:254`. Replace "no consumer passed it" with what the census
-      reports: four call sites in fitness-monitor, removed by `FM-CHART-1`. **Do not quietly delete
-      the sentence** — the point is that a measurement was wrong, and that is worth leaving legible.
-- [ ] **Version `3.1.1`**, with a CHANGELOG entry saying plainly that **`3.1.0` removed an applied
-      prop under a minor**, that four call sites were affected, and that the dev-mode error is what
-      makes the break loud. Do not publish a compensating `4.0.0`.
-- [ ] **Run the census and paste its output into this WO's register Notiz.** A WO about unverified
-      counts that lands with an unverified count is self-refuting.
-
-### Why `3.1.1` and not `4.0.0`
-
-**This WO changes no API.** Corrected prose, one error-message string, one dev script that never
-ships. On its own diff it is a patch, and labelling it `4.0.0` would announce a breaking change it
-does not contain.
-
-The counter-argument, made first in this WO and **withdrawn**: no consumer has crossed `3.1.0` —
-hram is on `3.0.1` (`HRAM-CHT-4`, landed), fitness-monitor on `2.37.0` committed / `3.0.1` in its
-working tree, jg-ferien on `2.41.1` — so the break is still ahead of all three, and a major would put
-the boundary where they will actually meet it.
-
-**What settles it: the break is not silent.** `assertRemovedChartProp` throws on first render in dev
-and names the replacement. The protection a major version buys — "read the notes before upgrading" —
-is already delivered by a hard error at the exact call site, which is strictly better than a version
-number nobody reads. A major would buy a second copy of a signal that already exists, at the cost of
-a version number that lies about its own contents.
-
-**This is conditional, and the condition is F2.** The guard is only a safety net if it tells the
-truth. An error saying `aspect` "was never applied" invites precisely the wrong action from the four
-callers whose cards change shape. **`3.1.1` is defensible because F2 is fixed in it — not
-independently of that.** If F2 were deferred, this argument collapses and the major comes back.
-
-`3.1.0` may additionally be deprecated on npm (`npm deprecate`, operator-only — a registry mutation).
-Low value here since nobody is on it and `3.1.1` supersedes it immediately; noted as available, not
-recommended.
-
-### Non-goals
-
-- **Do not reinstate `aspect`, and do not touch `minHeight`.** The removal stands; only its stated
-  reason and its error text are wrong.
-- Do not migrate fitness-monitor's four call sites — `FM-CHART-1` owns them.
-- Do not rewrite history, unpublish `3.1.0`, or publish a compensating `4.0.0`.
-- Do not extend the census beyond chart props. It answers "who passes what to the chart API", not
-  "who uses ucm".
-
-### Risks
-
-- **The census must stay out of the published package.** It is a dev script: keep it off the
-  `exports` map and out of `files`, and confirm the existing `packageTreeShaking` test still passes —
-  that test has been flaky under concurrent load before (`UCM-CHART-12`'s note), so a failure there
-  needs re-running alone before it is believed.
-- **The census reads sibling repos and must degrade gracefully** when a repo is absent, has no
-  frontend, or is a fresh clone — report "not present", never throw. It must never write anything.
-- **`FM-CHART-1` is mid-flight** with fitness-monitor's pin at `3.0.1` and all four `aspect` props
-  still in place. Until it lands, the census correctly reports 4 — that is the expected output, not a
-  failure. Do not "fix" fitness-monitor from here.
-- The corrected error message will be seen by that session. Getting F2 right is what makes its
-  remaining work safe, so **do F2 before anything else in this WO.**
-
-### Tests to WRITE — narrow
-
-- `ChartFrame` + `aspect`: the thrown message names `minHeight` and `size`, and **does not contain
-  "never applied"**. Assert on content, not just that it throws — `13`'s test matched
-  `/ChartFrame.*aspect.*size/s`, which the false message passes.
-- `ChartFrame` + `height`: unchanged, still green.
-- The census fixtures above.
-- Nothing else. `13`'s suite stays as it is apart from the corrected comment and the tightened
-  message assertion.
-
-No full suite.
-
----
-
-## Part B — Implementation map
-
-> **TO BE FILLED BY THE ORCHESTRATOR** on `git pull`. Must carry: `ChartFrame.jsx:22-35` (docblock)
-> and `:72-81` (the two `assertRemovedChartProp` calls); `assertRemovedChartProp` itself at
-> `chartDefaults.js:566`; the `docs/CHART-LAYOUT.md` "What stays unchanged" block; the test file's
-> `describe` at `:257` and comment at `:254`; the absolute working directory; the progress contract;
-> and the preamble. **Do not dispatch while this placeholder stands.**
-
----
-
-## Part C — Orchestrator-only
-
-> **TO BE FILLED BY THE ORCHESTRATOR.** Review routing (Tier 3: `reviewer` + `ui_reviewer`
-> concurrent), register maintenance including the pasted census output, publish verification for
-> `3.1.1`, commit, and the execution directive with its self-address guard.

package/work-orders/UCM-CHART-15.md

@@ -1,187 +0,0 @@
-# UCM-CHART-15 — Shift the chart size scale up one step, and raise the ceiling
-
-## Part A — Envelope
-
-*Authored by the Expertenchat. Authoritative WHAT/WHY.*
-
-### Goal
-
-The named chart sizes become one step taller, and the scale gains two steps above `tall`.
-
-| token | today | new |
-|---|---|---|
-| `compact` | 30 → 240 px | **40 → 320 px** |
-| `standard` | 40 → 320 px | **50 → 400 px** |
-| `tall` | 50 → 400 px | **60 → 480 px** |
-| `extra_tall` | — | **70 → 560 px** |
-| `super_tall` | — | **80 → 640 px** |
-
-The 10-unit step is preserved throughout, and every value stays a clean multiple of the theme's 8 px
-spacing unit.
-
-### Why
-
-Two independent pressures, both from real call sites rather than speculation.
-
-**The ceiling is too low.** `tall` (400 px) is the largest token, and hram has already had to use the
-documented `height` escape to get past it — `StructuralReachabilityPanel` sits at `height={480}` with
-a comment recording exactly this: *"no size token goes taller than `tall`"*. An escape used because
-the scale ends is a scale that ends too early.
-
-**The whole scale is too cramped.** The operator's judgement, on existing plots and not only the new
-ones. That judgement lands on a value that was never chosen on its merits: the comment at
-`src/components/charts/chartDefaults.js:295` says `standard` is
-
-> "pinned to the pre-existing deployed default (`TimeSeriesChart`'s old `CHART_HEIGHT = 320`) so this
-> migration does not also silently redraw every already-shipped default-sized chart"
-
-That was a **migration guardrail, not a design verdict** — 320 was kept because it was the previous
-value. The migration it protected is complete, so the pin has served its purpose and what remains is
-a height inherited rather than decided.
-
-### The blast radius is one app, and that is why now
-
-Measured across the workspace on 2026-08-24:
-
-```
-hram                    3.0.1     <- the only 3.x consumer
-14 other apps           2.41.3 / 2.37.0
-```
-
-The size tokens are a **3.0.0 feature** (this repo's own CHANGELOG marks 3.0.0 "Breaking": `size`
-replaced `minHeight`/`height`/`aspect`). On 2.x the scale does not exist. So no chart outside hram
-changes, and the verification surface is one application.
-
-**That inverts the usual caution into urgency:** fourteen apps are still to migrate onto 3.x. Every
-app that migrates onto a scale already known to be too cramped has to be revisited afterwards.
-Correcting the scale before they migrate is strictly cheaper than after.
-
-### Version: minor, by operator decision — and what that obliges
-
-Redefining what an existing public token renders is breaking under semver, and the Expertenchat
-recommended a major. **The operator decided minor on 2026-08-24; that decision stands and this WO
-implements it.**
-
-The consequence has to be carried somewhere, so it is carried here: a minor version can arrive
-through automation. Renovate's weekly run bumps minors without a human reading the diff, and on this
-estate that triage is also the vehicle that opens `develop → main`. A silent minor could therefore
-make every hram chart 80 px taller in a PR nobody read as a visual change.
-
-Two obligations follow, and they are not optional decoration:
-
-1. **The CHANGELOG entry must lead with the redraw**, in the plainest possible words: every existing
-   `compact`/`standard`/`tall` chart becomes 80 px taller. Not a bullet at the bottom — the first
-   line of the entry.
-2. **Release and consumption are coordinated**, not left to the weekly bump. `hram`'s pin bump and
-   its own verification are `HRAM-CHT-5` and land deliberately.
-
-### Scope
-
-**`src/components/charts/chartDefaults.js`**
-
-1. `CHART_SIZE_SPACING_UNITS` becomes `{ compact: 40, standard: 50, tall: 60, extra_tall: 70,
-   super_tall: 80 }`.
-2. **Rewrite the pin comment at `:294-299`, do not delete it.** It explains why 320 was there; without
-   a successor the next reader sees an unexplained shift. The new text should say what the old pin
-   was for, that its migration is complete, and that the scale is now chosen rather than inherited.
-
-**The enumeration sites — all of them.** The valid-size list is written out in more places than the
-one that matters, and a half-updated enumeration is this change's most likely defect:
-
-3. `resolveChartHeightPx`'s error message (`:407`) — `"compact" | "standard" | "tall"`.
-4. `assertRemovedChartProp`'s guidance string in **each** chart component that carries it
-   (`LineChart.jsx:72`, `BarChart.jsx:54`, and the same call in `ScatterChart`).
-5. The `size` JSDoc in **each** preset (`LineChart.jsx:44`, `BarChart.jsx:27`, `ScatterChart.jsx:155`).
-6. `docs/CHART-LAYOUT.md`.
-7. The CHANGELOG entry described above.
-
-Grep for the literal `"tall"` before declaring this done; the list above is what was found on
-2026-08-24 and is not guaranteed complete.
-
-### Non-goals / do not touch
-
-- `resolveChartLayout`'s geometry, the axis-band resolution, margins, `xLabels`, rotation logic.
-- The `height` px escape — it stays exactly as documented.
-- `ChartFrame`'s `minHeight` semantics.
-- Any consuming app. hram's adoption is `HRAM-CHT-5`.
-- Adding a sixth step. If 640 px proves too small later, the scale extends by another 10 units; do
-  not pre-add one now.
-
-### Risks
-
-- **A half-updated enumeration.** Adding the tokens to the resolver but not to the error message
-  produces a runtime error that lists only three valid values while five work — which reads as a bug
-  in the caller's code, not in this package.
-- **A test that asserts the token names but not their values** would pass against both the old and
-  new scale and prove nothing. See test 1.
-- The two new tokens have **no call site yet** in this repo. That is intended — `HRAM-CHT-5` adopts
-  them — but it means nothing here exercises them beyond the resolver test.
-
-### Tests to write
-
-1. `resolveChartHeightPx` returns **320 / 400 / 480 / 560 / 640** for the five tokens — asserting the
-   pixel values, not just that a number comes back. Written so it fails against the current scale.
-2. An unknown size still throws, and the message names **all five** valid tokens.
-3. The `height` escape still wins over `size`, for a new token as well as an old one.
-4. Existing chart-layout tests that hard-code 240/320/400 are updated deliberately, each one seen
-   and changed rather than mass-replaced — one of them may be asserting the old value *on purpose*.
-
----
-
-## Part B — Implementation map
-
-`.claude/models.local.json` has no local override, so the default applies: `implementation.runtime`
-is `claude`/`sonnet` — the Orchestrator implemented directly, no Codex dispatch, no preamble/progress
-contract needed.
-
-Context package used:
-
-- `src/components/charts/chartDefaults.js:294-300` (the pin comment + `CHART_SIZE_SPACING_UNITS`)
-  and `:402-419` (`resolveChartHeightPx`, its error message, and the `theme.spacing` unwrap comment).
-- Enumeration sites confirmed by grep for the literal `"tall"` (matched the WO's own 2026-08-24
-  list exactly, nothing extra found): `LineChart.jsx:44,72`, `BarChart.jsx:27,54`,
-  `ScatterChart.jsx:155,198`, `docs/CHART-LAYOUT.md:66,73-75`.
-- Existing hardcoded-value tests located via grep for `240|320|400` across `tests/*.test.{js,jsx}`:
-  `chartDefaults.test.js:119,562,576`, `BarChart.test.jsx:36,83-85`, `LineChart.test.jsx:51-53`,
-  `ScatterChart.test.jsx:242-244`. Distinguished from unrelated same-valued `minHeight`/`ChartFrame`
-  test fixtures (`chartDefaults.test.js:291,298,308,315,318`, `ChartFrame.test.jsx:294,311`,
-  `chart-api-census.test.js:196`), which were left untouched — they assert an arbitrary prop value,
-  not a size-token pixel height.
-
----
-
-## Part C — Orchestrator only
-
-> **STOP — everything below addresses the Orchestrator.**
-
-**Tier 3 · tests: the four resolver cases plus the updated layout tests**
-
-Shared-core — a change *inside* `ui-core-micha` is Tier 3 by the tiering table regardless of diff
-size, and this one redefines rendered output for every consumer that reaches 3.x.
-
-- `reviewer` — the enumeration completeness and the test's ability to fail are its brief.
-- `ui_reviewer` — the scale as a design decision: does the 10-unit step still read as a coherent
-  scale with five members, and is `super_tall` distinguishable from `extra_tall` in practice.
-- `sec_reviewer` — not run. No auth, no data, no exposure surface.
-- Concurrent, full context.
-
-### Release
-
-Minor bump per the operator's decision → **3.2.0** from the current 3.1.1. Follow this repo's release
-procedure. The CHANGELOG obligation in Part A is a release gate, not a nicety: if the entry does not
-lead with "every existing chart becomes 80 px taller", the release is not ready.
-
-Do **not** bump hram from here. `HRAM-CHT-5` owns that, and it carries the visual verification.
-
-### Register
-
-Row `UCM-CHART-15`, review notes in the required shape, the published version, landing SHA(s) in a
-second commit before the push. Record explicitly that a **major was recommended and minor was chosen
-by the operator** — a future reader tracing a silent redraw should find that decision, not infer it.
-
-### Log
-
-- **2026-08-24** — Envelope authored. Operator judged all three existing tokens too cramped and chose
-  a full one-step shift over adding a token at the top, accepting the breaking redraw; added
-  `super_tall` to the two steps proposed. Version set to minor against the Expertenchat's
-  recommendation of major, recorded above with its consequence.