fix(admin): absolute sidebar nav links so /admin URLs don't accumulate (v0.52.2)

The /admin left-rail NavLinks used relative targets (to="users", etc.). Under
the /admin/* nested route a relative link resolves against the current URL, so
each click appended a segment — Users→Allowlist→Graduation yielded
/admin/users/allowlist/graduation instead of /admin/graduation. Use absolute
targets (/admin/<tab>) + `end` for exact active matching.

Patch 0.52.1 → 0.52.2; CHANGELOG updated. Caught by the operator on PPE.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -23,6 +23,18 @@ skip versions are the composition of each intervening adjacent
 release's steps in order — no A-to-B path is pre-computed beyond
 that.
 
+## 0.52.2 — 2026-06-09
+
+**Patch — admin sidebar nav links no longer accumulate URL segments (no
+operator action required).**
+
+- **The `/admin` left-rail links were relative**, so under the `/admin/*`
+  nested route each click resolved against the current URL and *appended* a
+  segment — e.g. clicking Users → Allowlist → Graduation produced
+  `/admin/users/allowlist/graduation` instead of `/admin/graduation`. Fixed:
+  the rail `NavLink`s now use absolute targets (`/admin/<tab>`) with `end`
+  for exact active-state matching.
+
 ## 0.52.1 — 2026-06-08
 
 **Patch — collection-scoped entry-detail fetch fix (no operator action

VERSION

@@ -1 +1 @@
-0.52.1
+0.52.2

e2e/metadata.spec.js

@@ -43,15 +43,22 @@ test('SLICE-4: edit one entry\'s priority from the detail panel', async ({ page
 
   const panel = page.locator('.metadata-fields-panel')
   await expect(panel).toBeVisible()
-  // Seeded at P1; change to P2 and save (direct sidecar commit, D7).
-  await expect(panel.locator('#mf-priority')).toHaveValue('P1')
-  await panel.locator('#mf-priority').selectOption('P2')
+  const select = panel.locator('#mf-priority')
+  // Value-agnostic so the test is idempotent across retries and re-runs: a
+  // prior (committed) edit may have already moved it off the P1 seed. Pick a
+  // target distinct from the current value so Save is a real change.
+  const current = await select.inputValue()
+  const target = current === 'P2' ? 'P1' : 'P2'
+  await select.selectOption(target)
   await panel.getByRole('button', { name: 'Save' }).click()
-  await expect(panel.getByText('Saved')).toBeVisible()
+  // The save is a real git sidecar commit via the bot (D7); over the deployed
+  // edge the round trip can take tens of seconds — wait generously for the
+  // confirmation rather than the default expect timeout.
+  await expect(panel.getByText('Saved')).toBeVisible({ timeout: 60_000 })
 
   // Persisted across a reload (read back from the committed sidecar).
   await page.reload()
-  await expect(page.locator('.metadata-fields-panel #mf-priority')).toHaveValue('P2')
+  await expect(page.locator('.metadata-fields-panel #mf-priority')).toHaveValue(target)
 })
 
 // SLICE-5 (PUC-2) — multi-select + bulk action bar, one commit (authed).
@@ -71,8 +78,10 @@ test('SLICE-5: bulk-set priority on multiple selected entries', async ({ page })
   await expect(bar.getByText('2 selected')).toBeVisible()
 
   // Set priority P1 across both → one commit; a toast reports the result.
+  // Like SLICE-4, the bulk write is a real git commit via the bot; allow the
+  // deployed-edge round trip generous time before the result toast appears.
   await bar.getByLabel('Set Priority').selectOption('P1')
-  await expect(page.getByText(/2 updated/)).toBeVisible()
+  await expect(page.getByText(/2 updated/)).toBeVisible({ timeout: 60_000 })
 
   // The change is reflected server-side: the Priority P1 facet now counts the
   // two newly-updated entries plus the pre-existing P1 (checkout-returning was

frontend/package.json

@@ -1,7 +1,7 @@
 {
   "name": "rfc-app-frontend",
   "private": true,
-  "version": "0.52.1",
+  "version": "0.52.2",
   "type": "module",
   "scripts": {
     "dev": "vite",

frontend/src/components/Admin.jsx

@@ -59,7 +59,8 @@ export default function Admin({ viewer }) {
           {tabs.map(t => (
             <li key={t.path}>
               <NavLink
-                to={t.path}
+                to={`/admin/${t.path}`}
+                end
                 className={({ isActive }) => `admin-rail-link ${isActive ? 'active' : ''}`}
               >
                 {t.label}

testing/seed-ppe.sh

@@ -103,6 +103,17 @@ put_file() {
     201 "create $_repo/$_path"
 }
 
+# delete_file <repo> <path> — remove the file if it exists (no-op if absent).
+delete_file() {
+  _repo="$1"; _path="$2"
+  _sha=$(file_sha "$_repo" "$_path")
+  [ -n "$_sha" ] || { echo "seed-ppe: $_repo/$_path absent, nothing to delete"; return 0; }
+  echo "seed-ppe: deleting $_repo/$_path"
+  mutate DELETE "$GITEA/api/v1/repos/$ORG/$_repo/contents/$_path" \
+    "{\"message\":\"reseed: drop sidecar $_path\",\"sha\":\"$_sha\",\"branch\":\"main\"}" \
+    200 "delete $_repo/$_path"
+}
+
 ensure_repo "$REGISTRY_REPO"
 ensure_repo "$CONTENT_REPO"
 
@@ -182,4 +193,17 @@ tags: [search]
 Faceted search scenario.
 " "$RESEED"
 
+# RESEED also drops any metadata sidecars (<slug>.meta.yaml) left by prior
+# SLICE-4/5 edits. A sidecar takes precedence over the .md frontmatter
+# (dual-read, §22.4a), so without this a re-run inherits the edited
+# priorities (everything ends up P1) and the faceted preconditions drift —
+# SLICE-3 expects P0 count = 2. Dropping the sidecars restores the
+# frontmatter as the source of truth: checkout-guest=P0, search-facets=P0,
+# checkout-returning=P1.
+if [ "$RESEED" = "1" ]; then
+  for _slug in checkout-guest checkout-returning search-facets; do
+    delete_file "$CONTENT_REPO" "bdd/rfcs/$_slug.meta.yaml"
+  done
+fi
+
 echo "seed-ppe: done. Set REGISTRY_REPO=$REGISTRY_REPO on rfc-app-ppe and redeploy."