test(ingestion): add MR + nonzero_summary tests, close bd-1au9

src/ingestion/merge_requests.rs

@@ -423,59 +423,5 @@ fn parse_timestamp(ts: &str) -> Result<i64> {
 }
 
 #[cfg(test)]
-mod tests {
+#[path = "merge_requests_tests.rs"]
-    use super::*;
+mod tests;
-
-    #[test]
-    fn result_default_has_zero_counts() {
-        let result = IngestMergeRequestsResult::default();
-        assert_eq!(result.fetched, 0);
-        assert_eq!(result.upserted, 0);
-        assert_eq!(result.labels_created, 0);
-        assert_eq!(result.assignees_linked, 0);
-        assert_eq!(result.reviewers_linked, 0);
-    }
-
-    #[test]
-    fn cursor_filter_allows_newer_mrs() {
-        let cursor = SyncCursor {
-            updated_at_cursor: Some(1705312800000),
-            tie_breaker_id: Some(100),
-        };
-
-        let later_ts = 1705399200000;
-        assert!(passes_cursor_filter_with_ts(101, later_ts, &cursor));
-    }
-
-    #[test]
-    fn cursor_filter_blocks_older_mrs() {
-        let cursor = SyncCursor {
-            updated_at_cursor: Some(1705312800000),
-            tie_breaker_id: Some(100),
-        };
-
-        let earlier_ts = 1705226400000;
-        assert!(!passes_cursor_filter_with_ts(99, earlier_ts, &cursor));
-    }
-
-    #[test]
-    fn cursor_filter_uses_tie_breaker_for_same_timestamp() {
-        let cursor = SyncCursor {
-            updated_at_cursor: Some(1705312800000),
-            tie_breaker_id: Some(100),
-        };
-
-        assert!(passes_cursor_filter_with_ts(101, 1705312800000, &cursor));
-
-        assert!(!passes_cursor_filter_with_ts(100, 1705312800000, &cursor));
-
-        assert!(!passes_cursor_filter_with_ts(99, 1705312800000, &cursor));
-    }
-
-    #[test]
-    fn cursor_filter_allows_all_when_no_cursor() {
-        let cursor = SyncCursor::default();
-        let old_ts = 1577836800000;
-        assert!(passes_cursor_filter_with_ts(1, old_ts, &cursor));
-    }
-}

src/ingestion/merge_requests_tests.rs

@@ -0,0 +1,704 @@
+use std::path::Path;
+
+use super::*;
+use crate::core::config::{
+    EmbeddingConfig, GitLabConfig, LoggingConfig, ProjectConfig, ScoringConfig, StorageConfig,
+    SyncConfig,
+};
+use crate::core::db::{create_connection, run_migrations};
+use crate::gitlab::types::{GitLabAuthor, GitLabReferences, GitLabReviewer};
+
+// ─── Test Helpers ───────────────────────────────────────────────────────────
+
+fn setup_test_db() -> Connection {
+    let conn = create_connection(Path::new(":memory:")).unwrap();
+    run_migrations(&conn).unwrap();
+
+    conn.execute(
+        "INSERT INTO projects (id, gitlab_project_id, path_with_namespace, web_url)
+         VALUES (1, 100, 'group/project', 'https://gitlab.example.com/group/project')",
+        [],
+    )
+    .unwrap();
+
+    conn
+}
+
+fn test_config() -> Config {
+    Config {
+        gitlab: GitLabConfig {
+            base_url: "https://gitlab.example.com".to_string(),
+            token_env_var: "GITLAB_TOKEN".to_string(),
+        },
+        projects: vec![ProjectConfig {
+            path: "group/project".to_string(),
+        }],
+        default_project: None,
+        sync: SyncConfig::default(),
+        storage: StorageConfig::default(),
+        embedding: EmbeddingConfig::default(),
+        logging: LoggingConfig::default(),
+        scoring: ScoringConfig::default(),
+    }
+}
+
+fn make_test_mr(id: i64, updated_at: &str) -> GitLabMergeRequest {
+    GitLabMergeRequest {
+        id,
+        iid: id,
+        project_id: 100,
+        title: format!("MR {}", id),
+        description: None,
+        state: "opened".to_string(),
+        draft: false,
+        work_in_progress: false,
+        source_branch: "feature".to_string(),
+        target_branch: "main".to_string(),
+        sha: Some("abc123".to_string()),
+        references: Some(GitLabReferences {
+            short: format!("!{}", id),
+            full: format!("group/project!{}", id),
+        }),
+        detailed_merge_status: Some("mergeable".to_string()),
+        merge_status_legacy: None,
+        created_at: "2024-01-01T00:00:00.000Z".to_string(),
+        updated_at: updated_at.to_string(),
+        merged_at: None,
+        closed_at: None,
+        author: GitLabAuthor {
+            id: 1,
+            username: "test".to_string(),
+            name: "Test".to_string(),
+        },
+        merge_user: None,
+        merged_by: None,
+        labels: vec![],
+        assignees: vec![],
+        reviewers: vec![],
+        web_url: "https://example.com".to_string(),
+        merge_commit_sha: None,
+        squash_commit_sha: None,
+    }
+}
+
+fn make_mr_with_labels(id: i64, labels: Vec<&str>) -> GitLabMergeRequest {
+    let mut mr = make_test_mr(id, "2024-06-01T00:00:00.000Z");
+    mr.labels = labels.into_iter().map(String::from).collect();
+    mr
+}
+
+fn make_mr_with_assignees(id: i64, assignees: Vec<(&str, &str)>) -> GitLabMergeRequest {
+    let mut mr = make_test_mr(id, "2024-06-01T00:00:00.000Z");
+    mr.assignees = assignees
+        .into_iter()
+        .enumerate()
+        .map(|(i, (username, name))| GitLabAuthor {
+            id: (i + 10) as i64,
+            username: username.to_string(),
+            name: name.to_string(),
+        })
+        .collect();
+    mr
+}
+
+fn make_mr_with_reviewers(id: i64, reviewers: Vec<(&str, &str)>) -> GitLabMergeRequest {
+    let mut mr = make_test_mr(id, "2024-06-01T00:00:00.000Z");
+    mr.reviewers = reviewers
+        .into_iter()
+        .enumerate()
+        .map(|(i, (username, name))| GitLabReviewer {
+            id: (i + 20) as i64,
+            username: username.to_string(),
+            name: name.to_string(),
+        })
+        .collect();
+    mr
+}
+
+fn count_rows(conn: &Connection, table: &str) -> i64 {
+    conn.query_row(&format!("SELECT COUNT(*) FROM {table}"), [], |row| {
+        row.get(0)
+    })
+    .unwrap()
+}
+
+// ─── Cursor Filter Tests (preserved from inline) ───────────────────────────
+
+#[test]
+fn result_default_has_zero_counts() {
+    let result = IngestMergeRequestsResult::default();
+    assert_eq!(result.fetched, 0);
+    assert_eq!(result.upserted, 0);
+    assert_eq!(result.labels_created, 0);
+    assert_eq!(result.assignees_linked, 0);
+    assert_eq!(result.reviewers_linked, 0);
+}
+
+#[test]
+fn cursor_filter_allows_newer_mrs() {
+    let cursor = SyncCursor {
+        updated_at_cursor: Some(1705312800000),
+        tie_breaker_id: Some(100),
+    };
+
+    let later_ts = 1705399200000;
+    assert!(passes_cursor_filter_with_ts(101, later_ts, &cursor));
+}
+
+#[test]
+fn cursor_filter_blocks_older_mrs() {
+    let cursor = SyncCursor {
+        updated_at_cursor: Some(1705312800000),
+        tie_breaker_id: Some(100),
+    };
+
+    let earlier_ts = 1705226400000;
+    assert!(!passes_cursor_filter_with_ts(99, earlier_ts, &cursor));
+}
+
+#[test]
+fn cursor_filter_uses_tie_breaker_for_same_timestamp() {
+    let cursor = SyncCursor {
+        updated_at_cursor: Some(1705312800000),
+        tie_breaker_id: Some(100),
+    };
+
+    assert!(passes_cursor_filter_with_ts(101, 1705312800000, &cursor));
+    assert!(!passes_cursor_filter_with_ts(100, 1705312800000, &cursor));
+    assert!(!passes_cursor_filter_with_ts(99, 1705312800000, &cursor));
+}
+
+#[test]
+fn cursor_filter_allows_all_when_no_cursor() {
+    let cursor = SyncCursor::default();
+    let old_ts = 1577836800000;
+    assert!(passes_cursor_filter_with_ts(1, old_ts, &cursor));
+}
+
+// ─── parse_timestamp Tests ─────────────────────────────────────────────────
+
+#[test]
+fn parse_timestamp_valid_rfc3339() {
+    let ts = parse_timestamp("2024-06-15T12:30:00.000Z").unwrap();
+    assert_eq!(ts, 1718454600000);
+}
+
+#[test]
+fn parse_timestamp_invalid_format_returns_error() {
+    let result = parse_timestamp("not-a-date");
+    assert!(result.is_err());
+    let err_msg = result.unwrap_err().to_string();
+    assert!(err_msg.contains("not-a-date"));
+}
+
+// ─── Sync Cursor DB Tests ──────────────────────────────────────────────────
+
+#[test]
+fn get_sync_cursor_returns_default_when_no_row() {
+    let conn = setup_test_db();
+    let cursor = get_sync_cursor(&conn, 1).unwrap();
+    assert!(cursor.updated_at_cursor.is_none());
+    assert!(cursor.tie_breaker_id.is_none());
+}
+
+#[test]
+fn update_sync_cursor_creates_and_reads_back() {
+    let conn = setup_test_db();
+
+    update_sync_cursor(&conn, 1, 1705312800000, 42).unwrap();
+
+    let cursor = get_sync_cursor(&conn, 1).unwrap();
+    assert_eq!(cursor.updated_at_cursor, Some(1705312800000));
+    assert_eq!(cursor.tie_breaker_id, Some(42));
+}
+
+#[test]
+fn update_sync_cursor_upserts_on_conflict() {
+    let conn = setup_test_db();
+
+    update_sync_cursor(&conn, 1, 1000, 10).unwrap();
+    update_sync_cursor(&conn, 1, 2000, 20).unwrap();
+
+    let cursor = get_sync_cursor(&conn, 1).unwrap();
+    assert_eq!(cursor.updated_at_cursor, Some(2000));
+    assert_eq!(cursor.tie_breaker_id, Some(20));
+}
+
+#[test]
+fn reset_sync_cursor_clears_cursor() {
+    let conn = setup_test_db();
+
+    update_sync_cursor(&conn, 1, 1000, 10).unwrap();
+    reset_sync_cursor(&conn, 1).unwrap();
+
+    let cursor = get_sync_cursor(&conn, 1).unwrap();
+    assert!(cursor.updated_at_cursor.is_none());
+    assert!(cursor.tie_breaker_id.is_none());
+}
+
+#[test]
+fn sync_cursors_are_project_scoped() {
+    let conn = setup_test_db();
+    conn.execute(
+        "INSERT INTO projects (id, gitlab_project_id, path_with_namespace, web_url)
+         VALUES (2, 200, 'other/project', 'https://gitlab.example.com/other/project')",
+        [],
+    )
+    .unwrap();
+
+    update_sync_cursor(&conn, 1, 1000, 10).unwrap();
+    update_sync_cursor(&conn, 2, 2000, 20).unwrap();
+
+    let c1 = get_sync_cursor(&conn, 1).unwrap();
+    let c2 = get_sync_cursor(&conn, 2).unwrap();
+    assert_eq!(c1.updated_at_cursor, Some(1000));
+    assert_eq!(c2.updated_at_cursor, Some(2000));
+}
+
+// ─── process_single_mr Tests ───────────────────────────────────────────────
+
+#[test]
+fn process_single_mr_inserts_basic_mr() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+
+    let result = process_single_mr(&conn, &config, 1, &mr).unwrap();
+    assert_eq!(result.labels_created, 0);
+    assert_eq!(result.assignees_linked, 0);
+    assert_eq!(result.reviewers_linked, 0);
+
+    let (title, state, author, source_branch): (String, String, String, String) = conn
+        .query_row(
+            "SELECT title, state, author_username, source_branch
+             FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| Ok((row.get(0)?, row.get(1)?, row.get(2)?, row.get(3)?)),
+        )
+        .unwrap();
+    assert_eq!(title, "MR 1001");
+    assert_eq!(state, "opened");
+    assert_eq!(author, "test");
+    assert_eq!(source_branch, "feature");
+}
+
+#[test]
+fn process_single_mr_upserts_on_conflict() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr_v1 = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    process_single_mr(&conn, &config, 1, &mr_v1).unwrap();
+
+    let mut mr_v2 = make_test_mr(1001, "2024-06-16T12:00:00.000Z");
+    mr_v2.title = "Updated MR title".to_string();
+    mr_v2.state = "merged".to_string();
+    process_single_mr(&conn, &config, 1, &mr_v2).unwrap();
+
+    assert_eq!(count_rows(&conn, "merge_requests"), 1);
+
+    let (title, state): (String, String) = conn
+        .query_row(
+            "SELECT title, state FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| Ok((row.get(0)?, row.get(1)?)),
+        )
+        .unwrap();
+    assert_eq!(title, "Updated MR title");
+    assert_eq!(state, "merged");
+}
+
+#[test]
+fn process_single_mr_creates_labels() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_mr_with_labels(1001, vec!["bug", "critical"]);
+
+    let result = process_single_mr(&conn, &config, 1, &mr).unwrap();
+    assert_eq!(result.labels_created, 2);
+
+    assert_eq!(count_rows(&conn, "labels"), 2);
+
+    let label_count: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM mr_labels ml
+             JOIN merge_requests m ON ml.merge_request_id = m.id
+             WHERE m.gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert_eq!(label_count, 2);
+}
+
+#[test]
+fn process_single_mr_label_upsert_idempotent() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr1 = make_mr_with_labels(1001, vec!["bug"]);
+    let created1 = process_single_mr(&conn, &config, 1, &mr1).unwrap();
+    assert_eq!(created1.labels_created, 1);
+
+    // Second MR with same label
+    let mr2 = make_mr_with_labels(1002, vec!["bug"]);
+    let created2 = process_single_mr(&conn, &config, 1, &mr2).unwrap();
+    assert_eq!(created2.labels_created, 0); // label already exists
+
+    // Only 1 label row, but 2 junction rows
+    assert_eq!(count_rows(&conn, "labels"), 1);
+    assert_eq!(count_rows(&conn, "mr_labels"), 2);
+}
+
+#[test]
+fn process_single_mr_replaces_labels_on_update() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr_v1 = make_mr_with_labels(1001, vec!["bug", "critical"]);
+    process_single_mr(&conn, &config, 1, &mr_v1).unwrap();
+
+    let mut mr_v2 = make_mr_with_labels(1001, vec!["bug", "fixed"]);
+    mr_v2.updated_at = "2024-06-02T00:00:00.000Z".to_string();
+    process_single_mr(&conn, &config, 1, &mr_v2).unwrap();
+
+    let labels: Vec<String> = {
+        let mut stmt = conn
+            .prepare(
+                "SELECT l.name FROM labels l
+                 JOIN mr_labels ml ON l.id = ml.label_id
+                 JOIN merge_requests m ON ml.merge_request_id = m.id
+                 WHERE m.gitlab_id = 1001
+                 ORDER BY l.name",
+            )
+            .unwrap();
+        stmt.query_map([], |row| row.get(0))
+            .unwrap()
+            .collect::<std::result::Result<Vec<_>, _>>()
+            .unwrap()
+    };
+    assert_eq!(labels, vec!["bug", "fixed"]);
+}
+
+#[test]
+fn process_single_mr_creates_assignees() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_mr_with_assignees(1001, vec![("alice", "Alice"), ("bob", "Bob")]);
+
+    let result = process_single_mr(&conn, &config, 1, &mr).unwrap();
+    assert_eq!(result.assignees_linked, 2);
+
+    let count: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM mr_assignees ma
+             JOIN merge_requests m ON ma.merge_request_id = m.id
+             WHERE m.gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert_eq!(count, 2);
+}
+
+#[test]
+fn process_single_mr_replaces_assignees_on_update() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr_v1 = make_mr_with_assignees(1001, vec![("alice", "Alice"), ("bob", "Bob")]);
+    process_single_mr(&conn, &config, 1, &mr_v1).unwrap();
+
+    let mut mr_v2 = make_mr_with_assignees(1001, vec![("alice", "Alice"), ("charlie", "Charlie")]);
+    mr_v2.updated_at = "2024-06-02T00:00:00.000Z".to_string();
+    process_single_mr(&conn, &config, 1, &mr_v2).unwrap();
+
+    let assignees: Vec<String> = {
+        let mut stmt = conn
+            .prepare(
+                "SELECT ma.username FROM mr_assignees ma
+                 JOIN merge_requests m ON ma.merge_request_id = m.id
+                 WHERE m.gitlab_id = 1001
+                 ORDER BY ma.username",
+            )
+            .unwrap();
+        stmt.query_map([], |row| row.get(0))
+            .unwrap()
+            .collect::<std::result::Result<Vec<_>, _>>()
+            .unwrap()
+    };
+    assert_eq!(assignees, vec!["alice", "charlie"]);
+}
+
+#[test]
+fn process_single_mr_creates_reviewers() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_mr_with_reviewers(1001, vec![("reviewer1", "Reviewer One")]);
+
+    let result = process_single_mr(&conn, &config, 1, &mr).unwrap();
+    assert_eq!(result.reviewers_linked, 1);
+
+    let count: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM mr_reviewers mr
+             JOIN merge_requests m ON mr.merge_request_id = m.id
+             WHERE m.gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert_eq!(count, 1);
+}
+
+#[test]
+fn process_single_mr_replaces_reviewers_on_update() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr_v1 = make_mr_with_reviewers(1001, vec![("alice", "Alice"), ("bob", "Bob")]);
+    process_single_mr(&conn, &config, 1, &mr_v1).unwrap();
+
+    let mut mr_v2 = make_mr_with_reviewers(1001, vec![("charlie", "Charlie")]);
+    mr_v2.updated_at = "2024-06-02T00:00:00.000Z".to_string();
+    process_single_mr(&conn, &config, 1, &mr_v2).unwrap();
+
+    let reviewers: Vec<String> = {
+        let mut stmt = conn
+            .prepare(
+                "SELECT mr.username FROM mr_reviewers mr
+                 JOIN merge_requests m ON mr.merge_request_id = m.id
+                 WHERE m.gitlab_id = 1001
+                 ORDER BY mr.username",
+            )
+            .unwrap();
+        stmt.query_map([], |row| row.get(0))
+            .unwrap()
+            .collect::<std::result::Result<Vec<_>, _>>()
+            .unwrap()
+    };
+    assert_eq!(reviewers, vec!["charlie"]);
+}
+
+#[test]
+fn process_single_mr_marks_dirty() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    let local_id: i64 = conn
+        .query_row(
+            "SELECT id FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+
+    let dirty_count: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM dirty_sources
+             WHERE source_type = 'merge_request' AND source_id = ?",
+            [local_id],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert_eq!(dirty_count, 1);
+}
+
+#[test]
+fn process_single_mr_stores_raw_payload() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    let payload_id: Option<i64> = conn
+        .query_row(
+            "SELECT raw_payload_id FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert!(payload_id.is_some());
+
+    let payload_count: i64 = conn
+        .query_row(
+            "SELECT COUNT(*) FROM raw_payloads WHERE id = ?",
+            [payload_id.unwrap()],
+            |row| row.get(0),
+        )
+        .unwrap();
+    assert_eq!(payload_count, 1);
+}
+
+#[test]
+fn process_single_mr_stores_merge_commit_sha() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mut mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    mr.state = "merged".to_string();
+    mr.merge_commit_sha = Some("deadbeef1234".to_string());
+    mr.squash_commit_sha = Some("cafebabe5678".to_string());
+
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    let (merge_sha, squash_sha): (Option<String>, Option<String>) = conn
+        .query_row(
+            "SELECT merge_commit_sha, squash_commit_sha
+             FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| Ok((row.get(0)?, row.get(1)?)),
+        )
+        .unwrap();
+    assert_eq!(merge_sha.as_deref(), Some("deadbeef1234"));
+    assert_eq!(squash_sha.as_deref(), Some("cafebabe5678"));
+}
+
+// ─── Discussion Sync Queue Tests ───────────────────────────────────────────
+
+#[test]
+fn get_mrs_needing_discussion_sync_detects_updated() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    // MR was just upserted, discussions_synced_for_updated_at is NULL
+    let needing_sync = get_mrs_needing_discussion_sync(&conn, 1).unwrap();
+    assert_eq!(needing_sync.len(), 1);
+    assert_eq!(needing_sync[0].iid, 1001);
+}
+
+#[test]
+fn get_mrs_needing_discussion_sync_skips_already_synced() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    // Simulate discussion sync
+    let updated_at: i64 = conn
+        .query_row(
+            "SELECT updated_at FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| row.get(0),
+        )
+        .unwrap();
+    conn.execute(
+        "UPDATE merge_requests SET discussions_synced_for_updated_at = ? WHERE gitlab_id = 1001",
+        [updated_at],
+    )
+    .unwrap();
+
+    let needing_sync = get_mrs_needing_discussion_sync(&conn, 1).unwrap();
+    assert!(needing_sync.is_empty());
+}
+
+#[test]
+fn get_mrs_needing_discussion_sync_is_project_scoped() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    conn.execute(
+        "INSERT INTO projects (id, gitlab_project_id, path_with_namespace, web_url)
+         VALUES (2, 200, 'other/project', 'https://gitlab.example.com/other/project')",
+        [],
+    )
+    .unwrap();
+
+    let mr1 = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    process_single_mr(&conn, &config, 1, &mr1).unwrap();
+
+    let mut mr2 = make_test_mr(1002, "2024-06-15T12:00:00.000Z");
+    mr2.project_id = 200;
+    process_single_mr(&conn, &config, 2, &mr2).unwrap();
+
+    // Only project 1's MR should appear
+    let needing_sync = get_mrs_needing_discussion_sync(&conn, 1).unwrap();
+    assert_eq!(needing_sync.len(), 1);
+    assert_eq!(needing_sync[0].iid, 1001);
+}
+
+// ─── Reset / Full Sync Tests ──────────────────────────────────────────────
+
+#[test]
+fn reset_discussion_watermarks_clears_all_for_project() {
+    let conn = setup_test_db();
+    let config = test_config();
+
+    let mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    // Set some watermarks
+    conn.execute(
+        "UPDATE merge_requests SET
+            discussions_synced_for_updated_at = 1000,
+            resource_events_synced_for_updated_at = 2000,
+            closes_issues_synced_for_updated_at = 3000,
+            diffs_synced_for_updated_at = 4000
+         WHERE gitlab_id = 1001",
+        [],
+    )
+    .unwrap();
+
+    reset_discussion_watermarks(&conn, 1).unwrap();
+
+    let (disc_wm, events_wm, closes_wm, diffs_wm): (
+        Option<i64>,
+        Option<i64>,
+        Option<i64>,
+        Option<i64>,
+    ) = conn
+        .query_row(
+            "SELECT discussions_synced_for_updated_at,
+                    resource_events_synced_for_updated_at,
+                    closes_issues_synced_for_updated_at,
+                    diffs_synced_for_updated_at
+             FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| Ok((row.get(0)?, row.get(1)?, row.get(2)?, row.get(3)?)),
+        )
+        .unwrap();
+    assert!(disc_wm.is_none());
+    assert!(events_wm.is_none());
+    assert!(closes_wm.is_none());
+    assert!(diffs_wm.is_none());
+}
+
+#[test]
+fn process_single_mr_draft_and_references_stored() {
+    let conn = setup_test_db();
+    let config = test_config();
+    let mut mr = make_test_mr(1001, "2024-06-15T12:00:00.000Z");
+    mr.draft = true;
+    mr.references = Some(GitLabReferences {
+        short: "!1001".to_string(),
+        full: "group/project!1001".to_string(),
+    });
+    mr.detailed_merge_status = Some("checking".to_string());
+
+    process_single_mr(&conn, &config, 1, &mr).unwrap();
+
+    let (draft, refs_short, refs_full, merge_status): (
+        bool,
+        Option<String>,
+        Option<String>,
+        Option<String>,
+    ) = conn
+        .query_row(
+            "SELECT draft, references_short, references_full, detailed_merge_status
+             FROM merge_requests WHERE gitlab_id = 1001",
+            [],
+            |row| Ok((row.get(0)?, row.get(1)?, row.get(2)?, row.get(3)?)),
+        )
+        .unwrap();
+    assert!(draft);
+    assert_eq!(refs_short.as_deref(), Some("!1001"));
+    assert_eq!(refs_full.as_deref(), Some("group/project!1001"));
+    assert_eq!(merge_status.as_deref(), Some("checking"));
+}

src/ingestion/mod.rs

@@ -37,3 +37,38 @@ pub use orchestrator::{
     ingest_project_issues, ingest_project_issues_with_progress, ingest_project_merge_requests,
     ingest_project_merge_requests_with_progress,
 };
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn nonzero_summary_all_zero_returns_nothing() {
+        let result = nonzero_summary(&[("fetched", 0), ("upserted", 0)]);
+        assert_eq!(result, "nothing to update");
+    }
+
+    #[test]
+    fn nonzero_summary_empty_input_returns_nothing() {
+        let result = nonzero_summary(&[]);
+        assert_eq!(result, "nothing to update");
+    }
+
+    #[test]
+    fn nonzero_summary_single_nonzero() {
+        let result = nonzero_summary(&[("fetched", 5), ("skipped", 0)]);
+        assert_eq!(result, "5 fetched");
+    }
+
+    #[test]
+    fn nonzero_summary_multiple_nonzero_joined_with_middot() {
+        let result = nonzero_summary(&[("fetched", 10), ("upserted", 3), ("skipped", 0)]);
+        assert_eq!(result, "10 fetched \u{b7} 3 upserted");
+    }
+
+    #[test]
+    fn nonzero_summary_all_nonzero() {
+        let result = nonzero_summary(&[("a", 1), ("b", 2), ("c", 3)]);
+        assert_eq!(result, "1 a \u{b7} 2 b \u{b7} 3 c");
+    }
+}