fix(portals): absent is not empty, and every start claims the generation

Four more from review, three of them defects in last round's fixes.

The restore normalizer materialized `sourcePins: []` for archives that never
had the field, so "absent means no opinion" became "this archive says there
are no pins" and a merge cleared the user's. Absent now stays absent. My test
for that behaviour had passed for the wrong reason — it stubbed an empty pin
list, so the clear was skipped whether or not the guard worked.

`startGeneration` was claimed only by the switch path, so a plain Play, Resume
or Restart could be overtaken by a switch still awaiting its close. Every
start claims it now.

Raw and normalized tokens were paired by position, which breaks when
normalization drops a whole word: "FR: Ça" normalizes to "ca" and got handed
the raw token "FR:", sending it down the ASCII branch it cannot match from.
They are paired by normalized form instead.

And the ambiguous yearless alias (`title:dune:`) is no longer written or
retired beside a precise key — it may hold another remake's pre-enrichment
pin. It stays available when it is the only key there is, since refusing to
pin at all would be worse.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

apps/electron-backend/src/app/database/operations/title-sources.operations.spec.ts

@@ -93,6 +93,18 @@ describe('title-sources.operations', () => {
             expect(scanQuery.params).toContain('Ça');
         });
 
+        it('pairs raw tokens by meaning, not by position', async () => {
+            // Normalization drops whole words, so "FR: Ça" becomes just "ca".
+            // Pairing positionally would hand "ca" the raw token "FR:" and
+            // send it down the ASCII branch, where it cannot match "Ça".
+            const scan = createDbMock([]);
+            await findTitleSources(scan.db, { title: 'FR: Ça' });
+            const scanQuery = compiledQuery(scan.all);
+
+            expect(scanQuery.sql).toContain('instr');
+            expect(scanQuery.params).toContain('Ça');
+        });
+
         it('keeps the word boundary for ASCII tokens', async () => {
             // The looser substring test must not leak into the ASCII path,
             // where it would let "it" match "Titanic".

apps/electron-backend/src/app/database/operations/title-sources.operations.ts

@@ -160,10 +160,20 @@ function scanCandidateQuery(
     // multiword titles like "I Am" — matching on "i" alone would return most
     // of the catalog and hand it all to the TypeScript pass to throw away.
     const tokens = base.split(' ').filter(Boolean);
-    const rawTokens = rawTitle.split(/\s+/).filter(Boolean);
+    // Paired by NORMALIZED form, not by position: normalization drops whole
+    // words ("FR: Ça" -> "ca"), so a positional pairing hands "ca" the raw
+    // token "FR:" and picks the wrong branch for it.
+    const rawByBase = new Map<string, string>();
+    for (const raw of rawTitle.split(/\s+/).filter(Boolean)) {
+        const normalized = normalizeTitleKeys(raw).base;
+        if (normalized && !rawByBase.has(normalized)) {
+            rawByBase.set(normalized, raw);
+        }
+    }
+
     const wordMatches = sql.join(
-        tokens.map((token, index) =>
-            tokenPredicate(token, rawTokens[index] ?? token)
+        tokens.map((token) =>
+            tokenPredicate(token, rawByBase.get(token) ?? token)
         ),
         sql` AND `
     );