ci: add npm run build step to unit-tests job

The prerender fix only prevents regression if the build is actually run in CI. Without this gate, a future prerendered route that becomes unreachable behind auth would fail silently until someone runs the build manually. Fits after the test step in the existing unit-tests job — no new job needed since node_modules is already cached for the Playwright container. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
fix(build): add prerender entry for /hilfe/transkription
2026-05-09 08:52:20 +02:00 · 2026-05-09 08:51:45 +02:00
32 changed files with 96 additions and 929 deletions
--- a/.gitea/workflows/ci.yml
+++ b/.gitea/workflows/ci.yml
@@ -39,8 +39,6 @@ jobs:
      - name: Run unit and component tests
        run: npm test
        working-directory: frontend
-        env:
-          TZ: Europe/Berlin

      - name: Build frontend
        run: npm run build
@@ -80,8 +78,6 @@ jobs:
    runs-on: ubuntu-latest
    env:
      DOCKER_API_VERSION: "1.43"  # NAS runner runs Docker 24.x (max API 1.43); Testcontainers 2.x defaults to 1.44
-      DOCKER_HOST: unix:///var/run/docker.sock
-      TESTCONTAINERS_RYUK_DISABLED: "true"
    steps:
      - uses: actions/checkout@v4

--- a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentRepository.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentRepository.java
@@ -100,45 +100,7 @@ public interface DocumentRepository extends JpaRepository<Document, UUID>, JpaSp
            ORDER BY ts_rank(d.search_vector, q.pq) DESC,
                     d.meta_date DESC NULLS LAST
            """)
-    // Unpaged path — for bulk-edit "select all" and density chart
-    List<UUID> findAllMatchingIdsByFts(@Param("query") String query);
-
-    /**
-     * Returns one page of FTS-ranked document IDs with the total match count.
-     *
-     * <p>Each row contains (in column order):
-     * <ol>
-     *   <li>UUID   — document id</li>
-     *   <li>double — ts_rank score</li>
-     *   <li>long   — COUNT(*) OVER () — full match count, not page count</li>
-     * </ol>
-     *
-     * <p>Returns an empty list when the query matches no documents (including
-     * stopword-only queries where websearch_to_tsquery returns an empty tsquery).
-     * Use findAllMatchingIdsByFts for the unpaged bulk-edit path.
-     */
-    @Query(nativeQuery = true, value = """
-            WITH q AS (
-                SELECT CASE WHEN websearch_to_tsquery('german', :query)::text <> ''
-                            THEN to_tsquery('simple', regexp_replace(
-                                     websearch_to_tsquery('german', :query)::text,
-                                     '''([^'']+)''',
-                                     '''\\1'':*',
-                                     'g'))
-                       END AS pq
-            ), matches AS (
-                SELECT d.id, ts_rank(d.search_vector, q.pq) AS rank
-                FROM documents d, q
-                WHERE d.search_vector @@ q.pq
-            )
-            SELECT id, rank, COUNT(*) OVER () AS total
-            FROM matches
-            ORDER BY rank DESC, id
-            OFFSET :offset LIMIT :limit
-            """)
-    List<Object[]> findFtsPageRaw(@Param("query") String query,
-                                  @Param("offset") int offset,
-                                  @Param("limit") int limit);
+    List<UUID> findRankedIdsByFts(@Param("query") String query);

    /**
     * Returns match-enrichment data for a set of documents identified by their IDs.
--- a/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/DocumentService.java
@@ -162,7 +162,7 @@ public class DocumentService {
     */
    private List<UUID> resolveFtsIds(String text) {
        if (!StringUtils.hasText(text)) return null;
-        return documentRepository.findAllMatchingIdsByFts(text);
+        return documentRepository.findRankedIdsByFts(text);
    }

    /** Loads matching documents and projects to non-null {@link LocalDate}s. */
@@ -485,7 +485,7 @@ public class DocumentService {
        boolean hasText = StringUtils.hasText(text);
        List<UUID> rankedIds = null;
        if (hasText) {
-            rankedIds = documentRepository.findAllMatchingIdsByFts(text);
+            rankedIds = documentRepository.findRankedIdsByFts(text);
            if (rankedIds.isEmpty()) return List.of();
        }

@@ -645,43 +645,39 @@ public class DocumentService {
    // 1. Allgemeine Suche (für das Suchfeld im Frontend)
    public DocumentSearchResult searchDocuments(String text, LocalDate from, LocalDate to, UUID sender, UUID receiver, List<String> tags, String tagQ, DocumentStatus status, DocumentSort sort, String dir, TagOperator tagOperator, Pageable pageable) {
        boolean hasText = StringUtils.hasText(text);
-
-        // Pure-text RELEVANCE: push pagination into SQL — skip findAllMatchingIdsByFts entirely (ADR-008).
-        if (isPureTextRelevance(hasText, sort, from, to, sender, receiver, tags, tagQ, status)) {
-            return relevanceSortedPageFromSql(text, pageable);
-        }
-
        List<UUID> rankedIds = null;
+
        if (hasText) {
-            rankedIds = documentRepository.findAllMatchingIdsByFts(text);
+            rankedIds = documentRepository.findRankedIdsByFts(text);
            if (rankedIds.isEmpty()) return DocumentSearchResult.of(List.of());
        }

        Specification<Document> spec = buildSearchSpec(
                hasText, rankedIds, from, to, sender, receiver, tags, tagQ, status, tagOperator);

-        // SENDER and RECEIVER sorts load the full match set and slice in-memory.
+        // SENDER, RECEIVER and RELEVANCE sorts load the full match set and slice in memory.
        // JPA's Sort.by("sender.lastName") generates an INNER JOIN that silently drops
-        // documents with null sender/receivers. Cost scales with match count —
-        // acceptable while documents stays under ~10k rows. (ADR-008)
+        // documents with null sender/receivers; RELEVANCE maps a DB order to an external
+        // rank list. Cost scales linearly with match count — acceptable while documents
+        // stays under ~10k rows. Past that, replace with SQL-level LEFT JOIN sort.
        if (sort == DocumentSort.RECEIVER) {
-            // In-memory sort on page slice (≤ page size rows) — acceptable
            List<Document> sorted = sortByFirstReceiver(documentRepository.findAll(spec), dir);
            return buildResultPaged(pageSlice(sorted, pageable), text, pageable, sorted.size());
        }
        if (sort == DocumentSort.SENDER) {
-            // In-memory sort on page slice (≤ page size rows) — acceptable
            List<Document> sorted = sortBySender(documentRepository.findAll(spec), dir);
            return buildResultPaged(pageSlice(sorted, pageable), text, pageable, sorted.size());
        }

-        // RELEVANCE with active filters: load filtered subset and sort in-memory by rank.
+        // RELEVANCE: default when text present and no explicit sort given
        boolean useRankOrder = hasText && (sort == null || sort == DocumentSort.RELEVANCE);
        if (useRankOrder) {
+            List<Document> results = documentRepository.findAll(spec);
            Map<UUID, Integer> rankMap = new HashMap<>();
            for (int i = 0; i < rankedIds.size(); i++) rankMap.put(rankedIds.get(i), i);
-            List<Document> sorted = documentRepository.findAll(spec).stream()
-                    .sorted(Comparator.comparingInt(doc -> rankMap.getOrDefault(doc.getId(), Integer.MAX_VALUE)))
+            List<Document> sorted = results.stream()
+                    .sorted(Comparator.comparingInt(
+                            doc -> rankMap.getOrDefault(doc.getId(), Integer.MAX_VALUE)))
                    .toList();
            return buildResultPaged(pageSlice(sorted, pageable), text, pageable, sorted.size());
        }
@@ -692,39 +688,6 @@ public class DocumentService {
        return buildResultPaged(page.getContent(), text, pageable, page.getTotalElements());
    }

-    private static boolean isPureTextRelevance(boolean hasText, DocumentSort sort,
-            LocalDate from, LocalDate to, UUID sender, UUID receiver,
-            List<String> tags, String tagQ, DocumentStatus status) {
-        return hasText && (sort == null || sort == DocumentSort.RELEVANCE)
-                && from == null && to == null && sender == null && receiver == null
-                && (tags == null || tags.isEmpty()) && (tagQ == null || tagQ.isBlank()) && status == null;
-    }
-
-    /**
-     * Pure-text RELEVANCE path — pagination and ts_rank ordering pushed into SQL.
-     * Called when no non-text filters are active (ADR-008).
-     */
-    private DocumentSearchResult relevanceSortedPageFromSql(String text, Pageable pageable) {
-        long rawOffset = pageable.getOffset();
-        if (rawOffset > Integer.MAX_VALUE) return DocumentSearchResult.of(List.of());
-        int offset = (int) rawOffset;
-        int limit = pageable.getPageSize();
-        FtsPage ftsPage = toFtsPage(documentRepository.findFtsPageRaw(text, offset, limit));
-        if (ftsPage.hits().isEmpty()) return DocumentSearchResult.of(List.of());
-
-        // Preserve ts_rank order from SQL across the JPA findAllById call.
-        Map<UUID, Integer> rankMap = new HashMap<>();
-        List<UUID> pageIds = new ArrayList<>();
-        for (int i = 0; i < ftsPage.hits().size(); i++) {
-            rankMap.put(ftsPage.hits().get(i).id(), i);
-            pageIds.add(ftsPage.hits().get(i).id());
-        }
-        List<Document> docs = documentRepository.findAllById(pageIds).stream()
-                .sorted(Comparator.comparingInt(d -> rankMap.getOrDefault(d.getId(), Integer.MAX_VALUE)))
-                .toList();
-        return buildResultPaged(docs, text, pageable, ftsPage.total());
-    }
-
    private static <T> List<T> pageSlice(List<T> sorted, Pageable pageable) {
        int from = Math.min((int) pageable.getOffset(), sorted.size());
        int to = Math.min(from + pageable.getPageSize(), sorted.size());
@@ -1050,28 +1013,6 @@ public class DocumentService {
        return result;
    }

-    private static final int COL_ID = 0;
-    private static final int COL_RANK = 1;
-    private static final int COL_TOTAL = 2;
-
-    /**
-     * Maps raw Object[] rows from {@link DocumentRepository#findFtsPageRaw} to an
-     * {@link FtsPage}. Uses pattern-matching UUID cast to guard against driver-level
-     * type variance (some JDBC drivers return UUID as String).
-     */
-    private static FtsPage toFtsPage(List<Object[]> rows) {
-        if (rows.isEmpty()) return new FtsPage(List.of(), 0);
-        long total = ((Number) rows.get(0)[COL_TOTAL]).longValue();
-        List<FtsHit> hits = rows.stream()
-                .map(r -> {
-                    UUID id = r[COL_ID] instanceof UUID u ? u : UUID.fromString(r[COL_ID].toString());
-                    double rank = ((Number) r[COL_RANK]).doubleValue();
-                    return new FtsHit(id, rank);
-                })
-                .toList();
-        return new FtsPage(hits, total);
-    }
-
    /** Clean text + highlight offsets parsed from a {@code ts_headline} sentinel-delimited string. */
    public record ParsedHighlight(String cleanText, List<MatchOffset> offsets) {}

--- a/backend/src/main/java/org/raddatz/familienarchiv/document/FtsHit.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/FtsHit.java
@@ -1,6 +0,0 @@
-package org.raddatz.familienarchiv.document;
-
-import java.util.UUID;
-
-/** A single document hit from a paginated FTS query — id and its ts_rank score. */
-record FtsHit(UUID id, double rank) {}
--- a/backend/src/main/java/org/raddatz/familienarchiv/document/FtsPage.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/FtsPage.java
@@ -1,6 +0,0 @@
-package org.raddatz.familienarchiv.document;
-
-import java.util.List;
-
-/** One page of FTS results — the ranked hit list for this page and the total match count. */
-record FtsPage(List<FtsHit> hits, long total) {}
--- a/backend/src/main/java/org/raddatz/familienarchiv/document/comment/CommentController.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/document/comment/CommentController.java
@@ -27,9 +27,7 @@ public class CommentController {
    // ─── Block (transcription) comments ────────────────────────────────────────

    @GetMapping("/api/documents/{documentId}/transcription-blocks/{blockId}/comments")
-    public List<DocumentComment> getBlockComments(
-            @PathVariable UUID documentId,
-            @PathVariable UUID blockId) {
+    public List<DocumentComment> getBlockComments(@PathVariable UUID blockId) {
        return commentService.getCommentsForBlock(blockId);
    }

@@ -50,7 +48,6 @@ public class CommentController {
    @RequirePermission({Permission.ANNOTATE_ALL, Permission.WRITE_ALL})
    public DocumentComment replyToBlockComment(
            @PathVariable UUID documentId,
-            @PathVariable UUID blockId,
            @PathVariable UUID commentId,
            @RequestBody CreateCommentDTO dto,
            Authentication authentication) {
--- a/backend/src/main/java/org/raddatz/familienarchiv/user/AppUser.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/user/AppUser.java
@@ -88,8 +88,7 @@ public class AppUser {
    };

    public static String computeColor(UUID id) {
-        // Math.floorMod avoids the Integer.MIN_VALUE overflow trap in Math.abs(hashCode())
-        return PALETTE[Math.floorMod(id.hashCode(), PALETTE.length)];
+        return PALETTE[Math.abs(id.hashCode()) % PALETTE.length];
    }

    @PrePersist
--- a/backend/src/main/java/org/raddatz/familienarchiv/user/UserService.java
+++ b/backend/src/main/java/org/raddatz/familienarchiv/user/UserService.java
@@ -271,10 +271,9 @@ public class UserService {

    @Transactional
    public UserGroup createGroup(GroupDTO dto) {
-        UserGroup group = UserGroup.builder()
-                .name(dto.getName())
-                .permissions(dto.getPermissions() != null ? dto.getPermissions() : new HashSet<>())
-                .build();
+        UserGroup group = new UserGroup();
+        group.setName(dto.getName());
+        group.setPermissions(dto.getPermissions());
        return groupRepository.save(group);
    }

--- a/backend/src/main/resources/db/migration/V63__deduplicate_group_permissions.sql
+++ b/backend/src/main/resources/db/migration/V63__deduplicate_group_permissions.sql
@@ -1,7 +0,0 @@
-- Remove duplicate (group_id, permission) rows that accumulated without a UNIQUE constraint.
-- Keeps the row with the smallest ctid (earliest physical insertion order).
-DELETE FROM group_permissions a
-USING group_permissions b
-WHERE a.ctid < b.ctid
-  AND a.group_id = b.group_id
-  AND a.permission = b.permission;
--- a/backend/src/main/resources/db/migration/V64__group_permissions_primary_key.sql
+++ b/backend/src/main/resources/db/migration/V64__group_permissions_primary_key.sql
@@ -1,11 +0,0 @@
-- Add NOT NULL and PRIMARY KEY to group_permissions.
-- Requires V63 to have run first (no duplicates can remain).
--
-- After this migration, future seed migrations can use:
--   INSERT INTO group_permissions ... ON CONFLICT DO NOTHING
-- instead of the INSERT ... WHERE NOT EXISTS pattern used before V64.
-ALTER TABLE group_permissions
-    ALTER COLUMN permission SET NOT NULL;
-
-ALTER TABLE group_permissions
-    ADD CONSTRAINT pk_group_permissions PRIMARY KEY (group_id, permission);
--- a/backend/src/main/resources/db/migration/V65__tbmp_promote_unique_to_pk.sql
+++ b/backend/src/main/resources/db/migration/V65__tbmp_promote_unique_to_pk.sql
@@ -1,8 +0,0 @@
-- Promote the de-facto unique constraint on transcription_block_mentioned_persons to a named PK.
-- uq_tbmp_block_person (added in V57) is backed by a B-tree index identical to a PK;
-- this rename makes the naming convention explicit (pk_* vs uq_*).
-ALTER TABLE transcription_block_mentioned_persons
-    DROP CONSTRAINT uq_tbmp_block_person;
-
-ALTER TABLE transcription_block_mentioned_persons
-    ADD CONSTRAINT pk_tbmp PRIMARY KEY (block_id, person_id);
--- a/backend/src/test/java/org/raddatz/familienarchiv/MigrationIntegrationTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/MigrationIntegrationTest.java
@@ -399,68 +399,6 @@ class MigrationIntegrationTest {
              AND dc.annotation_id IS NOT NULL
            """;

-    // ─── V63+V64: group_permissions dedup + primary key ──────────────────────
-
-    @Test
-    void v64_pk_group_permissions_exists() {
-        Integer count = jdbc.queryForObject(
-                """
-                SELECT COUNT(*) FROM pg_catalog.pg_constraint c
-                JOIN pg_catalog.pg_class t ON c.conrelid = t.oid
-                WHERE t.relname = 'group_permissions'
-                  AND c.conname = 'pk_group_permissions'
-                  AND c.contype = 'p'
-                """,
-                Integer.class);
-        assertThat(count).isEqualTo(1);
-    }
-
-    @Test
-    void v64_permission_column_isNotNullable() {
-        Integer count = jdbc.queryForObject(
-                """
-                SELECT COUNT(*) FROM information_schema.columns
-                WHERE table_schema = 'public'
-                  AND table_name = 'group_permissions'
-                  AND column_name = 'permission'
-                  AND is_nullable = 'NO'
-                """,
-                Integer.class);
-        assertThat(count).isEqualTo(1);
-    }
-
-    @Test
-    @Transactional(propagation = Propagation.NOT_SUPPORTED)
-    void v64_rejectsDuplicateGroupPermission() {
-        UUID groupId = createUserGroup("DuplicateTestGroup-" + UUID.randomUUID());
-        try {
-            jdbc.update("INSERT INTO group_permissions (group_id, permission) VALUES (?, 'READ_ALL')", groupId);
-
-            assertThatThrownBy(() ->
-                    jdbc.update("INSERT INTO group_permissions (group_id, permission) VALUES (?, 'READ_ALL')", groupId)
-            ).isInstanceOf(DataIntegrityViolationException.class);
-        } finally {
-            jdbc.update("DELETE FROM group_permissions WHERE group_id = ?", groupId);
-            jdbc.update("DELETE FROM user_groups WHERE id = ?", groupId);
-        }
-    }
-
-    // ─── V65: tbmp UNIQUE promoted to PRIMARY KEY ─────────────────────────────
-
-    @Test
-    void v65_pk_tbmp_exists() {
-        Integer count = jdbc.queryForObject(
-                """
-                SELECT COUNT(*) FROM pg_catalog.pg_constraint c
-                JOIN pg_catalog.pg_class t ON c.conrelid = t.oid
-                WHERE t.relname = 'transcription_block_mentioned_persons'
-                  AND c.conname = 'pk_tbmp'
-                  AND c.contype = 'p'
-                """,
-                Integer.class);
-        assertThat(count).isEqualTo(1);
-    }
-
    // ─── helpers ─────────────────────────────────────────────────────────────

    private UUID createPerson(String firstName, String lastName) {
@@ -544,10 +482,4 @@ class MigrationIntegrationTest {
                """, id, recipientId, docId, commentId);
        return id;
    }
-
-    private UUID createUserGroup(String name) {
-        UUID id = UUID.randomUUID();
-        jdbc.update("INSERT INTO user_groups (id, name) VALUES (?, ?)", id, name);
-        return id;
-    }
 }
--- a/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentFtsPagedIntegrationTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentFtsPagedIntegrationTest.java
@@ -1,109 +0,0 @@
-package org.raddatz.familienarchiv.document;
-
-import jakarta.persistence.EntityManager;
-import org.junit.jupiter.api.BeforeEach;
-import org.junit.jupiter.api.Test;
-import org.raddatz.familienarchiv.PostgresContainerConfig;
-import org.raddatz.familienarchiv.config.FlywayConfig;
-import org.raddatz.familienarchiv.document.DocumentRepository;
-import org.raddatz.familienarchiv.document.Document;
-import org.raddatz.familienarchiv.document.DocumentStatus;
-import org.springframework.beans.factory.annotation.Autowired;
-import org.springframework.boot.data.jpa.test.autoconfigure.DataJpaTest;
-import org.springframework.boot.jdbc.test.autoconfigure.AutoConfigureTestDatabase;
-import org.springframework.context.annotation.Import;
-import org.springframework.test.annotation.DirtiesContext;
-
-import java.util.List;
-import java.util.UUID;
-
-import static org.assertj.core.api.Assertions.assertThat;
-import static org.assertj.core.api.Assertions.assertThatNoException;
-
-/**
- * Repository-level integration tests for {@code findFtsPageRaw}: verifies that the
- * paginated FTS query returns exactly page-size rows and that the window-function
- * total reflects the full match count, not just the page count.
- *
- * <p>Uses real Postgres via Testcontainers so the GIN index, tsvector trigger, and
- * {@code websearch_to_tsquery} semantics are identical to production.
- *
- * <p>{@code AFTER_CLASS} dirty-context keeps the Spring context alive for all tests
- * in this class and rebuilds it once at the end, rather than after every test.
- */
-@DataJpaTest
-@AutoConfigureTestDatabase(replace = AutoConfigureTestDatabase.Replace.NONE)
-@Import({PostgresContainerConfig.class, FlywayConfig.class})
-@DirtiesContext(classMode = DirtiesContext.ClassMode.AFTER_CLASS)
-class DocumentFtsPagedIntegrationTest {
-
-    @Autowired DocumentRepository documentRepository;
-    @Autowired EntityManager em;
-
-    // 60 docs match "Walter"; 10 docs with "Hans" do not.
-    private static final int WALTER_COUNT = 60;
-    private static final int PAGE_SIZE = 50;
-
-    @BeforeEach
-    void seed() {
-        documentRepository.deleteAll();
-        em.flush();
-        for (int i = 0; i < WALTER_COUNT; i++) {
-            documentRepository.saveAndFlush(doc("Brief von Walter Nr. " + i));
-        }
-        for (int i = 0; i < 10; i++) {
-            documentRepository.saveAndFlush(doc("Brief von Hans Nr. " + i));
-        }
-        em.clear();
-    }
-
-    @Test
-    void findFtsPageRaw_firstPage_returnsPageSizeRows() {
-        List<Object[]> rows = documentRepository.findFtsPageRaw("Walter", 0, PAGE_SIZE);
-
-        assertThat(rows).hasSize(PAGE_SIZE);
-    }
-
-    @Test
-    void findFtsPageRaw_windowTotal_equalsFullMatchCount_notPageSize() {
-        List<Object[]> rows = documentRepository.findFtsPageRaw("Walter", 0, PAGE_SIZE);
-
-        long total = ((Number) rows.get(0)[2]).longValue();
-        assertThat(total).isEqualTo(WALTER_COUNT);
-    }
-
-    @Test
-    void findFtsPageRaw_lastPage_returnsRemainder() {
-        int remainder = WALTER_COUNT % PAGE_SIZE; // 60 % 50 = 10
-        List<Object[]> rows = documentRepository.findFtsPageRaw("Walter", PAGE_SIZE, PAGE_SIZE);
-
-        assertThat(rows).hasSize(remainder);
-        long total = ((Number) rows.get(0)[2]).longValue();
-        assertThat(total).isEqualTo(WALTER_COUNT);
-    }
-
-    @Test
-    void findFtsPageRaw_noMatches_returnsEmptyList() {
-        List<Object[]> rows = documentRepository.findFtsPageRaw("XYZ_KEIN_TREFFER", 0, PAGE_SIZE);
-
-        assertThat(rows).isEmpty();
-    }
-
-    @Test
-    void findFtsPageRaw_stopwordOnlyQuery_returnsEmptyList_noException() {
-        assertThatNoException().isThrownBy(() -> {
-            List<Object[]> rows = documentRepository.findFtsPageRaw("der die das und", 0, PAGE_SIZE);
-            assertThat(rows).isEmpty();
-        });
-    }
-
-    // ─── Helper ───────────────────────────────────────────────────────────────
-
-    private Document doc(String title) {
-        return Document.builder()
-                .title(title)
-                .originalFilename(title.replace(" ", "_") + ".pdf")
-                .status(DocumentStatus.UPLOADED)
-                .build();
-    }
-}
--- a/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentFtsTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentFtsTest.java
@@ -69,7 +69,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Alter Brief"));
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Brief");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Brief");

        assertThat(ids).hasSize(1);
    }
@@ -79,7 +79,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Alter Brief"));
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Briefe");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Briefe");

        assertThat(ids).hasSize(1);
    }
@@ -89,7 +89,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Ein furchtbarer Brief"));
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("furchtb");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("furchtb");

        assertThat(ids).hasSize(1);
    }
@@ -99,7 +99,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Familienfoto"));
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Brief");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Brief");

        assertThat(ids).isEmpty();
    }
@@ -115,7 +115,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("schreiben");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("schreiben");

        assertThat(ids).contains(doc.getId());
    }
@@ -125,14 +125,14 @@ class DocumentFtsTest {
        Document doc = documentRepository.saveAndFlush(document("Leeres Dokument"));
        em.clear();

-        assertThat(documentRepository.findAllMatchingIdsByFts("Grundbuch")).isEmpty();
+        assertThat(documentRepository.findRankedIdsByFts("Grundbuch")).isEmpty();

        UUID annotationId = annotation(doc.getId());
        blockRepository.saveAndFlush(block(doc.getId(), annotationId, "Grundbuch Eintrag 1923", 0));
        em.flush();
        em.clear();

-        assertThat(documentRepository.findAllMatchingIdsByFts("Grundbuch")).contains(doc.getId());
+        assertThat(documentRepository.findRankedIdsByFts("Grundbuch")).contains(doc.getId());
    }

    @Test
@@ -144,13 +144,13 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        assertThat(documentRepository.findAllMatchingIdsByFts("Grundbuch")).contains(doc.getId());
+        assertThat(documentRepository.findRankedIdsByFts("Grundbuch")).contains(doc.getId());

        blockRepository.deleteById(block.getId());
        em.flush();
        em.clear();

-        assertThat(documentRepository.findAllMatchingIdsByFts("Grundbuch")).doesNotContain(doc.getId());
+        assertThat(documentRepository.findRankedIdsByFts("Grundbuch")).doesNotContain(doc.getId());
    }

    // ─── Ranking ───────────────────────────────────────────────────────────────
@@ -166,7 +166,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Grundbuch");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Grundbuch");

        assertThat(ids).hasSize(2);
        assertThat(ids.get(0)).isEqualTo(docA.getId());
@@ -179,7 +179,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Ein Brief von der Oma"));
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("der die das und");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("der die das und");

        assertThat(ids).isEmpty();
    }
@@ -195,7 +195,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Wille");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Wille");

        assertThat(ids).contains(doc.getId());
    }
@@ -205,7 +205,7 @@ class DocumentFtsTest {
        documentRepository.saveAndFlush(document("Brief"));
        em.clear();

-        assertThatNoException().isThrownBy(() -> documentRepository.findAllMatchingIdsByFts("((("));
+        assertThatNoException().isThrownBy(() -> documentRepository.findRankedIdsByFts("((("));
    }

    // ─── Weight C: sender/receiver names ───────────────────────────────────────
@@ -223,7 +223,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Schmidt");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Schmidt");

        assertThat(ids).contains(doc.getId());
    }
@@ -241,7 +241,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Raddatz");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Raddatz");

        assertThat(ids).contains(doc.getId());
    }
@@ -260,7 +260,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> ids = documentRepository.findAllMatchingIdsByFts("Familiengeschichte");
+        List<UUID> ids = documentRepository.findRankedIdsByFts("Familiengeschichte");

        assertThat(ids).hasSize(1);
    }
@@ -278,7 +278,7 @@ class DocumentFtsTest {
        em.flush();
        em.clear();

-        List<UUID> rankedIds = documentRepository.findAllMatchingIdsByFts("Grundbuch");
+        List<UUID> rankedIds = documentRepository.findRankedIdsByFts("Grundbuch");
        Specification<Document> spec = Specification.where(hasIds(rankedIds))
                .and(hasStatus(DocumentStatus.UPLOADED));

--- a/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentServiceSortTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentServiceSortTest.java
@@ -21,22 +21,17 @@ import org.springframework.data.domain.Pageable;
 import org.springframework.data.jpa.domain.Specification;

 import java.time.LocalDate;
-import java.util.ArrayList;
 import java.util.List;
 import java.util.UUID;

 import static org.assertj.core.api.Assertions.assertThat;
 import static org.mockito.ArgumentMatchers.any;
-import static org.mockito.ArgumentMatchers.anyInt;
-import static org.mockito.ArgumentMatchers.anyString;
-import static org.mockito.Mockito.never;
-import static org.mockito.Mockito.verify;
 import static org.mockito.Mockito.when;

@ExtendWith(MockitoExtension.class)
 class DocumentServiceSortTest {

-    private static final Pageable PAGE = org.springframework.data.domain.PageRequest.of(0, 10_000);
+    private static final Pageable UNPAGED = org.springframework.data.domain.PageRequest.of(0, 10_000);

    @Mock DocumentRepository documentRepository;
    @Mock PersonService personService;
@@ -48,12 +43,12 @@ class DocumentServiceSortTest {
    @Mock TranscriptionBlockQueryService transcriptionBlockQueryService;
    @InjectMocks DocumentService documentService;

-    // ─── DATE sort ────────────────────────────────────────────────────────────
+    // ─── searchDocuments — DATE sort ──────────────────────────────────────────

    @Test
    void searchDocuments_with_DATE_sort_and_text_sorts_chronologically_not_by_relevance() {
-        UUID id1 = UUID.randomUUID();  // higher relevance, older doc
-        UUID id2 = UUID.randomUUID();  // lower relevance, newer doc
+        UUID id1 = UUID.randomUUID();  // rank position 0 (higher relevance, older doc)
+        UUID id2 = UUID.randomUUID();  // rank position 1 (lower relevance, newer doc)

        Document older = Document.builder().id(id1)
                .title("Brief").status(DocumentStatus.UPLOADED)
@@ -62,48 +57,38 @@ class DocumentServiceSortTest {
                .title("Brief").status(DocumentStatus.UPLOADED)
                .documentDate(LocalDate.of(1960, 1, 1)).build();

-        when(documentRepository.findAllMatchingIdsByFts("Brief")).thenReturn(List.of(id1, id2));
+        // FTS returns id1 first (higher rank), id2 second
+        when(documentRepository.findRankedIdsByFts("Brief")).thenReturn(List.of(id1, id2));
+        // findAll(spec, pageable) — the correct date path — returns date-DESC order
        when(documentRepository.findAll(any(Specification.class), any(Pageable.class)))
                .thenReturn(new PageImpl<>(List.of(newer, older)));

        DocumentSearchResult result = documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null, DocumentSort.DATE, "DESC", null, PAGE);
+                "Brief", null, null, null, null, null, null, null, DocumentSort.DATE, "DESC", null, UNPAGED);

+        // Expect: date order (newer 1960 first), NOT rank order (older 1940 first)
        assertThat(result.items()).hasSize(2);
-        assertThat(result.items().get(0).document().getId()).isEqualTo(id2);  // newer first
+        assertThat(result.items().get(0).document().getId()).isEqualTo(id2);  // newer doc first
    }

-    // ─── RELEVANCE sort — pure text (no filters) ──────────────────────────────
-
-    @Test
-    void searchDocuments_relevance_pureText_calls_findFtsPageRaw_not_findAllMatchingIds() {
-        UUID id1 = UUID.randomUUID();
-        List<Object[]> ftsRows = ftsRows(id1, 0.5d, 1L);
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(ftsRows);
-        when(documentRepository.findAllById(any()))
-                .thenReturn(List.of(doc(id1)));
-
-        documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null, DocumentSort.RELEVANCE, null, null, PAGE);
-
-        verify(documentRepository).findFtsPageRaw(anyString(), anyInt(), anyInt());
-        verify(documentRepository, never()).findAllMatchingIdsByFts(anyString());
-    }
+    // ─── searchDocuments — RELEVANCE sort ─────────────────────────────────────

    @Test
    void searchDocuments_with_RELEVANCE_sort_and_text_preserves_fts_rank_order() {
-        UUID id1 = UUID.randomUUID();  // higher rank — must appear first
-        UUID id2 = UUID.randomUUID();  // lower rank
+        UUID id1 = UUID.randomUUID();  // rank position 0
+        UUID id2 = UUID.randomUUID();  // rank position 1

-        List<Object[]> ftsRows = new ArrayList<>();
-        ftsRows.add(new Object[]{id1, 0.8d, 2L});
-        ftsRows.add(new Object[]{id2, 0.3d, 2L});
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(ftsRows);
-        when(documentRepository.findAllById(any())).thenReturn(List.of(doc(id2), doc(id1))); // unordered from JPA
+        Document doc1 = Document.builder().id(id1).title("Brief").status(DocumentStatus.UPLOADED).build();
+        Document doc2 = Document.builder().id(id2).title("Brief").status(DocumentStatus.UPLOADED).build();
+
+        when(documentRepository.findRankedIdsByFts("Brief")).thenReturn(List.of(id1, id2));
+        when(documentRepository.findAll(any(Specification.class)))
+                .thenReturn(List.of(doc2, doc1));  // unordered from DB

        DocumentSearchResult result = documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null, DocumentSort.RELEVANCE, null, null, PAGE);
+                "Brief", null, null, null, null, null, null, null, DocumentSort.RELEVANCE, null, null, UNPAGED);

+        // Expect: rank order restored (id1 first)
        assertThat(result.items().get(0).document().getId()).isEqualTo(id1);
    }

@@ -112,82 +97,16 @@ class DocumentServiceSortTest {
        UUID id1 = UUID.randomUUID();
        UUID id2 = UUID.randomUUID();

-        List<Object[]> ftsRows = new ArrayList<>();
-        ftsRows.add(new Object[]{id1, 0.8d, 2L});
-        ftsRows.add(new Object[]{id2, 0.3d, 2L});
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(ftsRows);
-        when(documentRepository.findAllById(any())).thenReturn(List.of(doc(id2), doc(id1)));
+        Document doc1 = Document.builder().id(id1).title("Brief").status(DocumentStatus.UPLOADED).build();
+        Document doc2 = Document.builder().id(id2).title("Brief").status(DocumentStatus.UPLOADED).build();
+
+        when(documentRepository.findRankedIdsByFts("Brief")).thenReturn(List.of(id1, id2));
+        when(documentRepository.findAll(any(Specification.class)))
+                .thenReturn(List.of(doc2, doc1));

        DocumentSearchResult result = documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null, null, null, null, PAGE);
+                "Brief", null, null, null, null, null, null, null, null, null, null, UNPAGED);

        assertThat(result.items().get(0).document().getId()).isEqualTo(id1);
    }
-
-    // ─── RELEVANCE sort — overflow guard ─────────────────────────────────────
-
-    @Test
-    void searchDocuments_relevance_returns_empty_when_offset_exceeds_maxInt() {
-        // offset = pageNumber * pageSize; choose values so offset > Integer.MAX_VALUE
-        Pageable hugePage = org.springframework.data.domain.PageRequest.of(Integer.MAX_VALUE / 10 + 1, 10);
-
-        DocumentSearchResult result = documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null,
-                DocumentSort.RELEVANCE, null, null, hugePage);
-
-        assertThat(result.items()).isEmpty();
-        verify(documentRepository, never()).findFtsPageRaw(anyString(), anyInt(), anyInt());
-    }
-
-    // ─── toFtsPage — UUID-as-String JDBC driver variance ────────────────────
-
-    @Test
-    void searchDocuments_relevance_handles_string_uuid_from_jdbc_driver() {
-        String stringId = "11111111-1111-1111-1111-111111111111";
-        UUID uuidId = UUID.fromString(stringId);
-        // Simulate a JDBC driver that returns the id column as String instead of UUID
-        List<Object[]> ftsRows = new ArrayList<>();
-        ftsRows.add(new Object[]{stringId, 0.5d, 1L});
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(ftsRows);
-        when(documentRepository.findAllById(any())).thenReturn(List.of(doc(uuidId)));
-
-        DocumentSearchResult result = documentService.searchDocuments(
-                "Brief", null, null, null, null, null, null, null,
-                DocumentSort.RELEVANCE, null, null, PAGE);
-
-        assertThat(result.items()).hasSize(1);
-        assertThat(result.items().get(0).document().getId()).isEqualTo(uuidId);
-    }
-
-    // ─── RELEVANCE sort — text + active filter ────────────────────────────────
-
-    @Test
-    void searchDocuments_relevance_with_active_filter_uses_inMemory_path() {
-        UUID id1 = UUID.randomUUID();
-        UUID id2 = UUID.randomUUID();
-
-        when(documentRepository.findAllMatchingIdsByFts("Brief")).thenReturn(List.of(id1, id2));
-        when(documentRepository.findAll(any(Specification.class)))
-                .thenReturn(List.of(doc(id2), doc(id1)));
-
-        // sender filter is active → triggers in-memory path, not findFtsPageRaw
-        LocalDate from = LocalDate.of(1900, 1, 1);
-        documentService.searchDocuments(
-                "Brief", from, null, null, null, null, null, null, DocumentSort.RELEVANCE, null, null, PAGE);
-
-        verify(documentRepository, never()).findFtsPageRaw(anyString(), anyInt(), anyInt());
-        verify(documentRepository).findAllMatchingIdsByFts("Brief");
-    }
-
-    // ─── Helpers ──────────────────────────────────────────────────────────────
-
-    private static Document doc(UUID id) {
-        return Document.builder().id(id).title("Brief").status(DocumentStatus.UPLOADED).build();
-    }
-
-    private static List<Object[]> ftsRows(UUID id, double rank, long total) {
-        List<Object[]> rows = new ArrayList<>();
-        rows.add(new Object[]{id, rank, total});
-        return rows;
-    }
 }
--- a/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentServiceTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/document/DocumentServiceTest.java
@@ -1620,10 +1620,9 @@ class DocumentServiceTest {
        // chr(1)=\u0001 marks start, chr(2)=\u0002 marks end of highlighted term
        List<Object[]> rows = Collections.singletonList(new Object[]{docId, "\u0001Brief\u0002 an Anna", null, false, null, null, null});

-        List<Object[]> ftsRows = new java.util.ArrayList<>();
-        ftsRows.add(new Object[]{docId, 0.5d, 1L});
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(ftsRows);
-        when(documentRepository.findAllById(any())).thenReturn(List.of(doc));
+        when(documentRepository.findRankedIdsByFts("Brief")).thenReturn(List.of(docId));
+        when(documentRepository.findAll(any(org.springframework.data.jpa.domain.Specification.class)))
+                .thenReturn(List.of(doc));
        when(documentRepository.findEnrichmentData(any(), eq("Brief"))).thenReturn(rows);

        DocumentSearchResult result = documentService.searchDocuments(
@@ -1655,10 +1654,9 @@ class DocumentServiceTest {
        String snippetHeadline = "Hier ist der \u0001Brief\u0002 aus Berlin";
        List<Object[]> rows = Collections.singletonList(new Object[]{docId, "Dok", snippetHeadline, false, null, null, null});

-        List<Object[]> snippetFtsRows = new java.util.ArrayList<>();
-        snippetFtsRows.add(new Object[]{docId, 0.5d, 1L});
-        when(documentRepository.findFtsPageRaw(anyString(), anyInt(), anyInt())).thenReturn(snippetFtsRows);
-        when(documentRepository.findAllById(any())).thenReturn(List.of(doc));
+        when(documentRepository.findRankedIdsByFts("Brief")).thenReturn(List.of(docId));
+        when(documentRepository.findAll(any(org.springframework.data.jpa.domain.Specification.class)))
+                .thenReturn(List.of(doc));
        when(documentRepository.findEnrichmentData(any(), eq("Brief"))).thenReturn(rows);

        DocumentSearchResult result = documentService.searchDocuments(
@@ -2204,7 +2202,7 @@ class DocumentServiceTest {

    @Test
    void findIdsForFilter_returnsEmpty_whenFtsHasNoMatches() {
-        when(documentRepository.findAllMatchingIdsByFts("xyz")).thenReturn(List.of());
+        when(documentRepository.findRankedIdsByFts("xyz")).thenReturn(List.of());

        List<UUID> result = documentService.findIdsForFilter(
                "xyz", null, null, null, null, null, null, null, null);
@@ -2388,7 +2386,7 @@ class DocumentServiceTest {

    @Test
    void getDensity_shortCircuits_whenFtsReturnsNoMatches() {
-        when(documentRepository.findAllMatchingIdsByFts("xyz")).thenReturn(List.of());
+        when(documentRepository.findRankedIdsByFts("xyz")).thenReturn(List.of());

        DocumentDensityResult result = documentService.getDensity(
                new DensityFilters("xyz", null, null, null, null, null, null));
--- a/backend/src/test/java/org/raddatz/familienarchiv/document/comment/CommentControllerTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/document/comment/CommentControllerTest.java
@@ -44,14 +44,6 @@ class CommentControllerTest {

    // ─── Block comment endpoints ─────────────────────────────────────────────

-    @Test
-    @WithMockUser
-    void getBlockComments_returns400_when_documentId_is_not_a_UUID() throws Exception {
-        UUID blockId = UUID.randomUUID();
-        mockMvc.perform(get("/api/documents/NOT-A-UUID/transcription-blocks/" + blockId + "/comments"))
-                .andExpect(status().isBadRequest());
-    }
-
    @Test
    @WithMockUser
    void getBlockComments_returns200() throws Exception {
@@ -123,15 +115,6 @@ class CommentControllerTest {

    // ─── Block reply endpoints ───────────────────────────────────────────────

-    @Test
-    @WithMockUser(authorities = "ANNOTATE_ALL")
-    void replyToBlockComment_returns400_when_blockId_is_not_a_UUID() throws Exception {
-        mockMvc.perform(post("/api/documents/" + DOC_ID + "/transcription-blocks/NOT-A-UUID"
-                                + "/comments/" + COMMENT_ID + "/replies")
-                        .contentType(MediaType.APPLICATION_JSON).content(COMMENT_JSON))
-                .andExpect(status().isBadRequest());
-    }
-
    @Test
    void replyToBlockComment_returns401_whenUnauthenticated() throws Exception {
        UUID blockId = UUID.randomUUID();
--- a/backend/src/test/java/org/raddatz/familienarchiv/user/AppUserTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/user/AppUserTest.java
@@ -35,15 +35,4 @@ class AppUserTest {
            .count();
        assertThat(distinct).isGreaterThan(1);
    }
-
-    @Test
-    void computeColor_returnsValidPaletteColorForIntegerMinValueHash() {
-        // UUID "80000000-0000-0000-0000-000000000000" has hashCode() == Integer.MIN_VALUE.
-        // Math.abs(Integer.MIN_VALUE) overflows back to Integer.MIN_VALUE (negative), making
-        // Math.abs(hashCode()) % n unsafe for palette sizes that don't evenly divide MIN_VALUE.
-        // Math.floorMod eliminates this edge case entirely.
-        UUID minHashId = UUID.fromString("80000000-0000-0000-0000-000000000000");
-        assertThat(minHashId.hashCode()).isEqualTo(Integer.MIN_VALUE);
-        assertThat(EXPECTED_PALETTE).contains(AppUser.computeColor(minHashId));
-    }
 }
--- a/backend/src/test/java/org/raddatz/familienarchiv/user/UserServiceTest.java
+++ b/backend/src/test/java/org/raddatz/familienarchiv/user/UserServiceTest.java
@@ -902,18 +902,4 @@ class UserServiceTest {
        assertThat(result.getName()).isEqualTo("Familie");
        assertThat(result.getPermissions()).containsExactlyInAnyOrder("READ_ALL", "WRITE_ALL");
    }
-
-    @Test
-    void createGroup_withNullPermissions_savesGroupWithEmptyPermissionSet() {
-        org.raddatz.familienarchiv.user.GroupDTO dto = new org.raddatz.familienarchiv.user.GroupDTO();
-        dto.setName("Leser");
-        dto.setPermissions(null);
-
-        UserGroup saved = UserGroup.builder().id(UUID.randomUUID()).name("Leser").build();
-        when(groupRepository.save(any())).thenReturn(saved);
-
-        userService.createGroup(dto);
-
-        verify(groupRepository).save(argThat(g -> g.getPermissions() != null && g.getPermissions().isEmpty()));
-    }
 }
--- a/docs/adr/008-fts-sql-pagination.md
+++ b/docs/adr/008-fts-sql-pagination.md
@@ -1,68 +0,0 @@
-# ADR-008: SQL-level pagination for full-text search via window-function CTE
-
-## Status
-
-Accepted
-
-## Context
-
-`DocumentRepository.findAllMatchingIdsByFts` (formerly `findRankedIdsByFts`) returns all matching document IDs for a FTS query. `DocumentService.searchDocuments` then paginates in memory on the RELEVANCE sort path.
-
-A pre-production audit against 1,520 documents measured:
-
-```
-rows_per_call: 911 / call  (query: "walter")
-```
-
-At current scale this is acceptable — 911 UUIDs ≈ 14 KB, ms-level DB time. At 100 K+ documents two failure modes emerge:
-
-1. **Memory**: a broad query returns ~60 K UUIDs ≈ 1 MB per request, multiplied by concurrent users.
-2. **Latency**: the `LATERAL` join does work proportional to match-set size; at 60 K matches the FTS step alone exceeds 100 ms per query.
-
-Tracked as finding **F-31 (High)** in the pre-production architectural review.
-
-## Decision
-
-Push pagination and rank ordering into SQL for the RELEVANCE sort path when no non-text filters are active (pure full-text search):
-
-```sql
-WITH q AS (
-  SELECT CASE WHEN websearch_to_tsquery('german', :query)::text <> ''
-              THEN to_tsquery('simple', regexp_replace(
-                       websearch_to_tsquery('german', :query)::text,
-                       '''([^'']+)''', '''\\1'':*', 'g'))
-         END AS pq
-), matches AS (
-  SELECT d.id, ts_rank(d.search_vector, q.pq) AS rank
-  FROM documents d, q
-  WHERE d.search_vector @@ q.pq
-)
-SELECT id, rank, COUNT(*) OVER () AS total
-FROM matches
-ORDER BY rank DESC, id
-OFFSET :offset LIMIT :limit
-```
-
-`COUNT(*) OVER ()` returns the full match count alongside each page row in a single round-trip — no separate count query needed.
-
-`rows_per_call` for the FTS query drops from match-set size (911) to page size (≤ 50).
-
-When non-text filters (date range, sender, receiver, tags, status) are also active, the existing path is preserved: `findAllMatchingIdsByFts` returns all ranked IDs, which are passed as an `IN` clause to the JPA Specification, and `totalElements` comes from the JPA `Page.getTotalElements()`. This keeps the count accurate across the combined filter set.
-
-## Alternatives Considered
-
-**1. Two-query approach (separate COUNT + paged SELECT)**
-Correct, but doubles round-trips. The window function achieves the same result in one query.
-
-**2. Capped result set with a user-visible warning**
-Return at most N results (e.g. 500) and show "showing top 500 of many results". Simpler, but degrades UX for broad queries and doesn't reduce latency proportionally (still scans N rows).
-
-**3. Full SQL rewrite combining FTS + JPA Specification filters**
-Possible via a native query that embeds all filter predicates. Eliminates the in-memory SENDER/RECEIVER sort paths and the two-phase approach. High complexity, tight coupling to schema details, loses type-safe JPA Specification composition. Deferred to a future refactor if scale demands it.
-
-## Consequences
-
- **`rows_per_call` for pure-text FTS searches drops to ≤ page size** — the primary metric.
- **SENDER and RECEIVER sort paths stay in-memory** for combined text+filter queries. For pure-text queries with SENDER/RECEIVER sort, the current approach (fetch all matched IDs, build spec, load all matched entities, sort in-memory) still runs. This is acceptable while the archive stays under ~10 K documents.
- **RELEVANCE sort with text+filters still loads the full filtered entity set in-memory.** The filtered set is typically much smaller than the raw FTS match set, so the cost is bounded by filter selectivity, not total match count.
- **`findAllMatchingIdsByFts` is retained** for: (a) the bulk-edit "select all" fast path (`findIdsForFilter`), (b) the document density chart (`getDensity`), and (c) the SENDER/RECEIVER in-memory sort paths.
--- a/frontend/src/lib/document/viewer/PdfControls.svelte
+++ b/frontend/src/lib/document/viewer/PdfControls.svelte
@@ -35,7 +35,7 @@ let {
 			onclick={onPrev}
 			disabled={currentPage <= 1}
 			aria-label="Zurück"
-			class="min-h-[44px] min-w-[44px] rounded p-2 text-ink-3 transition hover:bg-surface/10 focus-visible:ring-2 focus-visible:ring-brand-navy focus-visible:ring-offset-1 disabled:opacity-40"
+			class="rounded p-1 text-ink-3 transition hover:bg-surface/10 disabled:opacity-40"
 		>
 			<svg class="h-4 w-4" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
 				<path stroke-linecap="round" stroke-linejoin="round" d="M15 19l-7-7 7-7" />
@@ -52,7 +52,7 @@ let {
 			onclick={onNext}
 			disabled={!isLoaded || currentPage >= totalPages}
 			aria-label="Weiter"
-			class="min-h-[44px] min-w-[44px] rounded p-2 text-ink-3 transition hover:bg-surface/10 focus-visible:ring-2 focus-visible:ring-brand-navy focus-visible:ring-offset-1 disabled:opacity-40"
+			class="rounded p-1 text-ink-3 transition hover:bg-surface/10 disabled:opacity-40"
 		>
 			<svg class="h-4 w-4" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
 				<path stroke-linecap="round" stroke-linejoin="round" d="M9 5l7 7-7 7" />
@@ -65,7 +65,7 @@ let {
 		<button
 			onclick={onZoomOut}
 			aria-label="Verkleinern"
-			class="min-h-[44px] min-w-[44px] rounded p-2 text-ink-3 transition hover:bg-surface/10 focus-visible:ring-2 focus-visible:ring-brand-navy focus-visible:ring-offset-1"
+			class="rounded p-1 text-ink-3 transition hover:bg-surface/10"
 		>
 			<svg class="h-4 w-4" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
 				<circle cx="11" cy="11" r="8" />
@@ -75,7 +75,7 @@ let {
 		<button
 			onclick={onZoomIn}
 			aria-label="Vergrößern"
-			class="min-h-[44px] min-w-[44px] rounded p-2 text-ink-3 transition hover:bg-surface/10 focus-visible:ring-2 focus-visible:ring-brand-navy focus-visible:ring-offset-1"
+			class="rounded p-1 text-ink-3 transition hover:bg-surface/10"
 		>
 			<svg class="h-4 w-4" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2">
 				<circle cx="11" cy="11" r="8" />
@@ -89,8 +89,7 @@ let {
 		<button
 			onclick={onToggleAnnotations}
 			aria-label={showAnnotations ? m.pdf_annotations_hide() : m.pdf_annotations_show()}
-			aria-pressed={showAnnotations}
-			class="flex min-h-[44px] min-w-[44px] items-center gap-1.5 rounded px-3 py-2 font-sans text-xs transition focus-visible:ring-2 focus-visible:ring-brand-navy focus-visible:ring-offset-1 {showAnnotations
+			class="flex items-center gap-1.5 rounded px-2 py-1 font-sans text-xs transition {showAnnotations
 				? 'text-ink-2 hover:bg-surface/10'
 				: 'bg-surface/10 text-primary'}"
 		>
--- a/frontend/src/lib/document/viewer/PdfControls.svelte.spec.ts
+++ b/frontend/src/lib/document/viewer/PdfControls.svelte.spec.ts
@@ -65,111 +65,3 @@ describe('PdfControls — annotation toggle contrast (WCAG 2.1 AA)', () => {
 		expect(annotationBtn!.className).not.toContain('text-accent');
 	});
 });
-
-describe('PdfControls — focus rings (WCAG 2.1 §2.4.7)', () => {
-	it('annotation toggle button has focus-visible:ring-2 focus ring', async () => {
-		const { container } = render(PdfControls, {
-			...defaultProps,
-			annotationCount: 2,
-			showAnnotations: false
-		});
-		const allButtons = container.querySelectorAll('button');
-		const annotationBtn = Array.from(allButtons).find((b) =>
-			b.getAttribute('aria-label')?.toLowerCase().includes('annotierungen')
-		);
-		expect(annotationBtn).not.toBeNull();
-		expect(annotationBtn!.className).toContain('focus-visible:ring-2');
-	});
-
-	it('icon-only nav/zoom buttons each have focus-visible:ring-2 focus ring', async () => {
-		const { container } = render(PdfControls, { ...defaultProps });
-		const allButtons = container.querySelectorAll('button');
-		const iconOnlyButtons = Array.from(allButtons).filter((b) => {
-			const label = b.getAttribute('aria-label') ?? '';
-			return ['zurück', 'weiter', 'verkleinern', 'vergrößern'].includes(label.toLowerCase());
-		});
-		expect(iconOnlyButtons).toHaveLength(4);
-		for (const btn of iconOnlyButtons) {
-			expect(btn.className).toContain('focus-visible:ring-2');
-		}
-	});
-});
-
-describe('PdfControls — touch targets (WCAG 2.2 §2.5.8)', () => {
-	it('annotation toggle button has min-h-[44px] touch target', async () => {
-		const { container } = render(PdfControls, {
-			...defaultProps,
-			annotationCount: 2,
-			showAnnotations: false
-		});
-		const allButtons = container.querySelectorAll('button');
-		const annotationBtn = Array.from(allButtons).find((b) =>
-			b.getAttribute('aria-label')?.toLowerCase().includes('annotierungen')
-		);
-		expect(annotationBtn).not.toBeNull();
-		expect(annotationBtn!.className).toContain('min-h-[44px]');
-	});
-
-	it('annotation toggle button has min-w-[44px] touch target', async () => {
-		const { container } = render(PdfControls, {
-			...defaultProps,
-			annotationCount: 2,
-			showAnnotations: false
-		});
-		const allButtons = container.querySelectorAll('button');
-		const annotationBtn = Array.from(allButtons).find((b) =>
-			b.getAttribute('aria-label')?.toLowerCase().includes('annotierungen')
-		);
-		expect(annotationBtn).not.toBeNull();
-		expect(annotationBtn!.className).toContain('min-w-[44px]');
-	});
-
-	it('annotation toggle reflects pressed state via aria-pressed', async () => {
-		const { container: c1 } = render(PdfControls, {
-			...defaultProps,
-			annotationCount: 2,
-			showAnnotations: false
-		});
-		const btn1 = Array.from(c1.querySelectorAll('button')).find((b) =>
-			b.getAttribute('aria-label')?.toLowerCase().includes('annotierungen')
-		);
-		expect(btn1!.getAttribute('aria-pressed')).toBe('false');
-		cleanup();
-
-		const { container: c2 } = render(PdfControls, {
-			...defaultProps,
-			annotationCount: 2,
-			showAnnotations: true
-		});
-		const btn2 = Array.from(c2.querySelectorAll('button')).find((b) =>
-			b.getAttribute('aria-label')?.toLowerCase().includes('annotierungen')
-		);
-		expect(btn2!.getAttribute('aria-pressed')).toBe('true');
-	});
-
-	it('icon-only nav/zoom buttons each have min-h-[44px] touch target', async () => {
-		const { container } = render(PdfControls, { ...defaultProps });
-		const allButtons = container.querySelectorAll('button');
-		const iconOnlyButtons = Array.from(allButtons).filter((b) => {
-			const label = b.getAttribute('aria-label') ?? '';
-			return ['zurück', 'weiter', 'verkleinern', 'vergrößern'].includes(label.toLowerCase());
-		});
-		expect(iconOnlyButtons).toHaveLength(4);
-		for (const btn of iconOnlyButtons) {
-			expect(btn.className).toContain('min-h-[44px]');
-		}
-	});
-
-	it('icon-only nav/zoom buttons each have min-w-[44px] touch target', async () => {
-		const { container } = render(PdfControls, { ...defaultProps });
-		const allButtons = container.querySelectorAll('button');
-		const iconOnlyButtons = Array.from(allButtons).filter((b) => {
-			const label = b.getAttribute('aria-label') ?? '';
-			return ['zurück', 'weiter', 'verkleinern', 'vergrößern'].includes(label.toLowerCase());
-		});
-		expect(iconOnlyButtons).toHaveLength(4);
-		for (const btn of iconOnlyButtons) {
-			expect(btn.className).toContain('min-w-[44px]');
-		}
-	});
-});
--- a/frontend/src/lib/person/PersonTypeahead.svelte
+++ b/frontend/src/lib/person/PersonTypeahead.svelte
@@ -21,7 +21,6 @@ interface Props {
 	restrictToCorrespondentsOf?: string;
 	excludePersonId?: string;
 	badge?: 'additive' | 'replace';
-	resetKey?: number;
 	onchange?: (value: string) => void;
 	onfocused?: () => void;
 }
@@ -40,20 +39,17 @@ let {
 	restrictToCorrespondentsOf,
 	excludePersonId,
 	badge,
-	resetKey = 0,
 	onchange,
 	onfocused
 }: Props = $props();

 // searchTerm must be both prop-derived AND locally writable (user typing), so $state +
 // $effect is the correct pattern here — writable $derived is read-only and won't work.
+// eslint-disable-next-line svelte/prefer-writable-derived
 let searchTerm = $state(initialName);

-// Sync display text when initialName changes OR when resetKey increments (navigation reset).
-// resetKey is incremented by the page on every SvelteKit navigation so that a manually-typed
-// term that was never committed (no person selected) gets cleared even if initialName stays ''.
+// Sync display text when the selected person changes externally (e.g. swap, navigation).
 $effect(() => {
-	void resetKey;
 	searchTerm = initialName;
 });

--- a/frontend/src/lib/person/PersonTypeahead.svelte.spec.ts
+++ b/frontend/src/lib/person/PersonTypeahead.svelte.spec.ts
@@ -270,33 +270,6 @@ describe('PersonTypeahead – correspondent mode', () => {
 	});
 });

-// ─── resetKey ─────────────────────────────────────────────────────────────────
-
-describe('PersonTypeahead – resetKey', () => {
-	// Note: rerender() in vitest-browser-svelte causes a full re-mount, not an in-place prop
-	// update. This is a smoke test — the $effect(resetKey) path that fires during SvelteKit
-	// navigation (prop update on a live instance) cannot be isolated at this level.
-	it('clears a manually-typed term when resetKey changes even if initialName stays empty', async () => {
-		mockFetchWithPersons([]);
-		const { rerender } = render(PersonTypeahead, {
-			name: 'senderId',
-			label: 'Absender',
-			initialName: '',
-			resetKey: 0
-		});
-		const input = page.getByPlaceholder('Namen tippen...');
-
-		// User types something without selecting a person
-		await input.fill('Max');
-		await waitForDebounce();
-		await expect.element(input).toHaveValue('Max');
-
-		// Navigation resets: initialName stays '', but resetKey increments
-		await rerender({ name: 'senderId', label: 'Absender', initialName: '', resetKey: 1 });
-		await expect.element(input).toHaveValue('');
-	});
-});
-
 // ─── Click outside ────────────────────────────────────────────────────────────

 describe('PersonTypeahead – click outside', () => {
--- a/frontend/src/lib/shared/primitives/DateInput.svelte
+++ b/frontend/src/lib/shared/primitives/DateInput.svelte
@@ -1,5 +1,4 @@
 <script lang="ts">
-import { untrack } from 'svelte';
 import { isoToGerman, handleGermanDateInput, germanToIso } from '$lib/shared/utils/date';
 import { m } from '$lib/paraglide/messages.js';

@@ -25,16 +24,6 @@ let {

 let display = $state(isoToGerman(value ?? ''));

-// Re-derive display when value changes externally (e.g. timeline drag, reset nav).
-// Guard prevents overwriting while the user is mid-typing a partial date:
-// germanToIso returns '' for partial input, matching value '' → no re-derive.
-$effect(() => {
-	const externalIso = value ?? '';
-	if (germanToIso(untrack(() => display)) !== externalIso) {
-		display = isoToGerman(externalIso);
-	}
-});
-
 // ─── Validation helper ────────────────────────────────────────────────────
 function isCalendarValid(iso: string): boolean {
 	if (!iso) return false;
--- a/frontend/src/lib/shared/primitives/DateInput.svelte.spec.ts
+++ b/frontend/src/lib/shared/primitives/DateInput.svelte.spec.ts
@@ -183,26 +183,6 @@ describe('DateInput – clearing the date', () => {
 	});
 });

-// ─── External value changes ───────────────────────────────────────────────────
-
-describe('DateInput – external value changes', () => {
-	it('clears display when value prop is reset to empty externally', async () => {
-		const { rerender } = render(DateInput, { value: '1920-01-01' });
-		const input = page.getByRole('textbox');
-		await expect.element(input).toHaveValue('01.01.1920');
-		await rerender({ value: '' });
-		await expect.element(input).toHaveValue('');
-	});
-
-	it('updates display when value prop changes to a new date externally', async () => {
-		const { rerender } = render(DateInput, { value: '1920-01-01' });
-		const input = page.getByRole('textbox');
-		await expect.element(input).toHaveValue('01.01.1920');
-		await rerender({ value: '1945-05-08' });
-		await expect.element(input).toHaveValue('08.05.1945');
-	});
-});
-
 // ─── Hidden input ─────────────────────────────────────────────────────────────

 describe('DateInput – hidden input for form submission', () => {
--- a/frontend/src/routes/SearchFilterBar.svelte
+++ b/frontend/src/routes/SearchFilterBar.svelte
@@ -20,7 +20,6 @@ let {
 	showAdvanced = $bindable(false),
 	initialSenderName = '',
 	initialReceiverName = '',
-	navKey = 0,
 	isLoading = false,
 	onSearch,
 	onSearchImmediate,
@@ -40,7 +39,6 @@ let {
 	showAdvanced?: boolean;
 	initialSenderName?: string;
 	initialReceiverName?: string;
-	navKey?: number;
 	isLoading?: boolean;
 	onSearch: () => void;
 	onSearchImmediate?: () => void;
@@ -199,7 +197,6 @@ $effect(() => {
 						label={m.docs_filter_label_sender()}
 						bind:value={senderId}
 						initialName={initialSenderName}
-						resetKey={navKey}
 						onchange={onSearch}
 					/>
 				</div>
@@ -215,7 +212,6 @@ $effect(() => {
 						label={m.docs_filter_label_receivers()}
 						bind:value={receiverId}
 						initialName={initialReceiverName}
-						resetKey={navKey}
 						onchange={onSearch}
 					/>
 				</div>
--- a/frontend/src/routes/documents/+page.server.ts
+++ b/frontend/src/routes/documents/+page.server.ts
@@ -3,23 +3,6 @@ import { createApiClient } from '$lib/shared/api.server';
 import { getErrorMessage } from '$lib/shared/errors';
 import type { components } from '$lib/generated/api';

-const UUID_RE = /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/i;
-
-async function resolvePersonName(
-	id: string,
-	api: ReturnType<typeof createApiClient>
-): Promise<string> {
-	if (!UUID_RE.test(id)) return '';
-	try {
-		const result = await api.GET('/api/persons/{id}', { params: { path: { id } } });
-		if (!result.response.ok) return '';
-		return result.data?.displayName ?? '';
-	} catch (e) {
-		console.error('[resolvePersonName] failed for id', id, e);
-		return '';
-	}
-}
-
 type DocumentSearchItem = components['schemas']['DocumentSearchItem'];

 const VALID_SORTS = ['DATE', 'TITLE', 'SENDER', 'RECEIVER', 'UPLOAD_DATE', 'RELEVANCE'] as const;
@@ -51,30 +34,25 @@ export async function load({ url, fetch }) {
 	const api = createApiClient(fetch);

 	let result;
-	let initialSenderName = '';
-	let initialReceiverName = '';
 	try {
-		[result, [initialSenderName, initialReceiverName]] = await Promise.all([
-			api.GET('/api/documents/search', {
-				params: {
-					query: {
-						q: q || undefined,
-						from: from || undefined,
-						to: to || undefined,
-						senderId: senderId || undefined,
-						receiverId: receiverId || undefined,
-						tag: tags.length ? tags : undefined,
-						tagQ: tagQ && !tags.length ? tagQ : undefined,
-						tagOp: tagOp === 'OR' ? 'OR' : undefined,
-						sort,
-						dir: dir || undefined,
-						page,
-						size: PAGE_SIZE
-					}
+		result = await api.GET('/api/documents/search', {
+			params: {
+				query: {
+					q: q || undefined,
+					from: from || undefined,
+					to: to || undefined,
+					senderId: senderId || undefined,
+					receiverId: receiverId || undefined,
+					tag: tags.length ? tags : undefined,
+					tagQ: tagQ && !tags.length ? tagQ : undefined,
+					tagOp: tagOp === 'OR' ? 'OR' : undefined,
+					sort,
+					dir: dir || undefined,
+					page,
+					size: PAGE_SIZE
 				}
-			}),
-			Promise.all([resolvePersonName(senderId, api), resolvePersonName(receiverId, api)])
-		]);
+			}
+		});
 	} catch {
 		return {
 			items: [] as DocumentSearchItem[],
@@ -87,8 +65,6 @@ export async function load({ url, fetch }) {
 			to,
 			senderId,
 			receiverId,
-			initialSenderName: '',
-			initialReceiverName: '',
 			tags,
 			sort,
 			dir,
@@ -118,8 +94,6 @@ export async function load({ url, fetch }) {
 		to,
 		senderId,
 		receiverId,
-		initialSenderName,
-		initialReceiverName,
 		tags,
 		sort,
 		dir,
--- a/frontend/src/routes/documents/+page.svelte
+++ b/frontend/src/routes/documents/+page.svelte
@@ -22,9 +22,6 @@ let from = $state(untrack(() => data.from || ''));
 let to = $state(untrack(() => data.to || ''));
 let senderId = $state(untrack(() => data.senderId || ''));
 let receiverId = $state(untrack(() => data.receiverId || ''));
-let initialSenderName = $state(untrack(() => data.initialSenderName ?? ''));
-let initialReceiverName = $state(untrack(() => data.initialReceiverName ?? ''));
-let navKey = $state(0);
 let tagNames = $state<{ name: string; id?: string; color?: string; parentId?: string }[]>(
 	untrack(() => (data.tags || []).map((name: string) => ({ name })))
 );
@@ -210,17 +207,12 @@ async function editAllMatching() {

 // Keep local filter state in sync with server data after navigation completes.
 // Guard q: skip overwrite while the user is actively typing.
-// navKey increments on every navigation so PersonTypeahead clears manually-typed
-// terms even when initialSenderName/initialReceiverName stays '' across navigations.
 $effect(() => {
 	if (!qFocused) q = data.q || '';
 	from = data.from || '';
 	to = data.to || '';
 	senderId = data.senderId || '';
 	receiverId = data.receiverId || '';
-	initialSenderName = data.initialSenderName ?? '';
-	initialReceiverName = data.initialReceiverName ?? '';
-	untrack(() => navKey++);
 	tagNames = (data.tags || []).map((name: string) => ({ name }));
 	sort = data.sort || 'DATE';
 	dir = data.dir || 'desc';
@@ -255,9 +247,6 @@ $effect(() => {
 		bind:dir={dir}
 		bind:tagQ={tagQ}
 		bind:tagOperator={tagOperator}
-		initialSenderName={initialSenderName}
-		initialReceiverName={initialReceiverName}
-		navKey={navKey}
 		isLoading={navigating.to !== null}
 		onSearch={handleTextSearch}
 		onSearchImmediate={handleImmediateSearch}
--- a/frontend/src/routes/documents/page.server.spec.ts
+++ b/frontend/src/routes/documents/page.server.spec.ts
@@ -167,76 +167,3 @@ describe('documents page load — network error fallback', () => {
 		expect(result.items).toEqual([]);
 	});
 });
-
-// ─── person name resolution ───────────────────────────────────────────────────
-
-describe('documents page load — person name resolution', () => {
-	function makeSearchMock(personResult?: { ok: boolean; displayName?: string }) {
-		const mockGet = vi.fn().mockImplementation((path: string) => {
-			if (path === '/api/documents/search') {
-				return Promise.resolve({
-					response: { ok: true, status: 200 },
-					data: { items: [], totalElements: 0, pageNumber: 0, pageSize: 50, totalPages: 0 }
-				});
-			}
-			// person lookup via api.GET('/api/persons/{id}', ...)
-			if (!personResult?.ok) {
-				return Promise.resolve({ response: { ok: false, status: 404 }, data: undefined });
-			}
-			return Promise.resolve({
-				response: { ok: true, status: 200 },
-				data: { displayName: personResult.displayName ?? '' }
-			});
-		});
-		vi.mocked(createApiClient).mockReturnValue({ GET: mockGet } as ReturnType<
-			typeof createApiClient
-		>);
-		return mockGet;
-	}
-
-	it('returns initialSenderName from person lookup when senderId is a valid UUID', async () => {
-		makeSearchMock({ ok: true, displayName: 'Max Mustermann' });
-
-		const result = await load({
-			url: makeUrl({ senderId: '11111111-1111-1111-1111-111111111111' }),
-			fetch: vi.fn() as unknown as typeof fetch
-		});
-
-		expect(result.initialSenderName).toBe('Max Mustermann');
-	});
-
-	it('returns initialReceiverName from person lookup when receiverId is a valid UUID', async () => {
-		makeSearchMock({ ok: true, displayName: 'Anna Musterfrau' });
-
-		const result = await load({
-			url: makeUrl({ receiverId: '22222222-2222-2222-2222-222222222222' }),
-			fetch: vi.fn() as unknown as typeof fetch
-		});
-
-		expect(result.initialReceiverName).toBe('Anna Musterfrau');
-	});
-
-	it('returns empty string when senderId is not a valid UUID', async () => {
-		const mockGet = makeSearchMock();
-
-		const result = await load({
-			url: makeUrl({ senderId: 'not-a-uuid' }),
-			fetch: vi.fn() as unknown as typeof fetch
-		});
-
-		expect(result.initialSenderName).toBe('');
-		// UUID guard fires before any api.GET call — only document search is called
-		expect(mockGet).toHaveBeenCalledTimes(1);
-	});
-
-	it('returns empty string when person api returns 404', async () => {
-		makeSearchMock({ ok: false });
-
-		const result = await load({
-			url: makeUrl({ senderId: '11111111-1111-1111-1111-111111111111' }),
-			fetch: vi.fn() as unknown as typeof fetch
-		});
-
-		expect(result.initialSenderName).toBe('');
-	});
-});
--- a/frontend/src/routes/documents/page.svelte.spec.ts
+++ b/frontend/src/routes/documents/page.svelte.spec.ts
@@ -23,8 +23,6 @@ function makeData(overrides: Record<string, unknown> = {}) {
 		to: '',
 		senderId: '',
 		receiverId: '',
-		initialSenderName: '',
-		initialReceiverName: '',
 		tags: [],
 		sort: 'DATE',
 		dir: 'desc',
@@ -138,22 +136,6 @@ describe('documents page — URL building', () => {
 	});
 });

-// ─── Sender / receiver name display ──────────────────────────────────────────
-
-describe('documents page — sender/receiver display', () => {
-	it('pre-fills sender typeahead from initialSenderName when senderId filter is active', async () => {
-		render(Page, {
-			data: makeData({
-				senderId: '11111111-1111-1111-1111-111111111111',
-				initialSenderName: 'Max Mustermann'
-			})
-		});
-		// Advanced filters are auto-shown when senderId is set
-		const inputs = page.getByPlaceholder('Namen tippen...');
-		await expect.element(inputs.first()).toHaveValue('Max Mustermann');
-	});
-});
-
 // ─── Timeline density widget wiring (#385) ────────────────────────────────────

 describe('documents page — timeline density widget', () => {
--- a/runner-config.yaml
+++ b/runner-config.yaml
@@ -1,16 +0,0 @@
-# runner-config.yaml — only the relevant section
-container:
-  # passed as DOCKER_HOST inside the job container
-  docker_host: "unix:///var/run/docker.sock"
-  # whitelists the socket path so workflows can mount it
-  valid_volumes:
-    - "/var/run/docker.sock"
-  # appended to `docker run` when the runner spawns a job container
-  # SECURITY: Mounting the Docker socket grants job containers root-equivalent
-  # access to the host Docker daemon. Acceptable here because only trusted code
-  # from this private repo runs on this runner. Do NOT use on a runner that
-  # accepts untrusted PRs from external contributors.
-  options: "-v /var/run/docker.sock:/var/run/docker.sock"
-  # keep network mode default (bridge) — Testcontainers handles its own networking
-  force_pull: false
-