feat(store): index-format rebuild boundary for the File-QN change

Makefile.cbm

@@ -357,7 +357,7 @@ TEST_DISCOVER_SRCS = \
 
 TEST_GRAPH_BUFFER_SRCS = tests/test_graph_buffer.c
 
-TEST_PIPELINE_SRCS = tests/test_registry.c tests/test_pipeline.c tests/test_fqn.c tests/test_route_canon.c tests/test_path_alias.c tests/test_configlink.c tests/test_infrascan.c tests/test_worker_pool.c tests/test_parallel.c tests/test_index_resilience.c
+TEST_PIPELINE_SRCS = tests/test_registry.c tests/test_pipeline.c tests/test_fqn.c tests/test_route_canon.c tests/test_path_alias.c tests/test_configlink.c tests/test_infrascan.c tests/test_worker_pool.c tests/test_parallel.c tests/test_index_resilience.c tests/test_index_format.c
 
 TEST_WATCHER_SRCS = tests/test_watcher.c
 

src/pipeline/pipeline.c

@@ -1116,17 +1116,30 @@ static int try_incremental_or_delete_db(cbm_pipeline_t *p, cbm_file_info_t *file
     if (check_store && cbm_store_check_integrity(check_store)) {
         cbm_file_hash_t *hashes = NULL;
         int hash_count = 0;
+        int fmt = 0;
+
         cbm_store_get_file_hashes(check_store, p->project_name, &hashes, &hash_count);
         cbm_store_free_file_hashes(hashes, hash_count);
+        cbm_store_get_format_version(check_store, &fmt);
+
+        bool format_stale = (fmt != CBM_INDEX_FORMAT_VERSION);
         cbm_store_close(check_store);
-        if (hash_count > 0 && file_count <= hash_count + (hash_count / PAIR_LEN)) {
+
+        if (format_stale) {
+            if (hash_count > 0) {
+                char stored_buf[CBM_SZ_32];
+                snprintf(stored_buf, sizeof(stored_buf), "%d", fmt);
+                cbm_log_info("pipeline.route", "path", "format_change_reindex", "stored_format",
+                             stored_buf);
+            }
+            /* fall through to delete + full rebuild */
+        } else if (hash_count > 0 && file_count <= hash_count + (hash_count / PAIR_LEN)) {
             cbm_log_info("pipeline.route", "path", "incremental", "stored_hashes",
                          itoa_buf(hash_count));
             int rc = cbm_pipeline_run_incremental(p, db_path, files, file_count);
             free(db_path);
             return rc;
-        }
-        if (hash_count > 0) {
+        } else if (hash_count > 0) {
             cbm_log_info("pipeline.route", "path", "mode_change_reindex", "stored_hashes",
                          itoa_buf(hash_count), "discovered", itoa_buf(file_count));
         }
@@ -1224,8 +1237,13 @@ static int dump_and_persist_hashes(cbm_pipeline_t *p, const cbm_file_info_t *fil
     CBM_PROF_START(t_reopen);
     cbm_store_t *hash_store = cbm_store_open_path(db_path);
     CBM_PROF_END("persist", "1_reopen", t_reopen);
+    bool format_stamped = true;
     if (hash_store) {
         bool hash_records_complete = true;
+        if (cbm_store_set_format_version(hash_store, CBM_INDEX_FORMAT_VERSION) != CBM_STORE_OK) {
+            cbm_log_error("pipeline.err", "phase", "persist_format_version");
+            format_stamped = false;
+        }
         CBM_PROF_START(t_delhash);
         if (cbm_store_delete_file_hashes(hash_store, p->project_name) != CBM_STORE_OK) {
             hash_records_complete = false;
@@ -1377,6 +1395,9 @@ static int dump_and_persist_hashes(cbm_pipeline_t *p, const cbm_file_info_t *fil
     free(p->saved_adr);
     p->saved_adr = NULL;
 
+    if (!format_stamped) {
+        return CBM_NOT_FOUND;
+    }
     /* Export persistent artifact if enabled */
     if (p->persistence) {
         CBM_PROF_START(t_art);

src/pipeline/pipeline_incremental.c

@@ -643,11 +643,11 @@ static void run_postpasses(cbm_pipeline_ctx_t *ctx, cbm_file_info_t *changed_fil
  * Mode-skipped hash rows are preserved across the rebuild so subsequent
  * reindexes can correctly distinguish "never indexed" from "indexed but
  * not visited this pass". */
-static void dump_and_persist(cbm_gbuf_t *gbuf, const char *db_path, const char *project,
-                             cbm_file_info_t *files, int file_count,
-                             const cbm_file_hash_t *mode_skipped, int mode_skipped_count,
-                             const char *repo_path, const cbm_coverage_row_t *cov, int cov_count,
-                             const cbm_coverage_meta_t *meta_template) {
+static int dump_and_persist(cbm_gbuf_t *gbuf, const char *db_path, const char *project,
+                            cbm_file_info_t *files, int file_count,
+                            const cbm_file_hash_t *mode_skipped, int mode_skipped_count,
+                            const char *repo_path, const cbm_coverage_row_t *cov, int cov_count,
+                            const cbm_coverage_meta_t *meta_template) {
     struct timespec t;
     cbm_clock_gettime(CLOCK_MONOTONIC, &t);
 
@@ -662,12 +662,20 @@ static void dump_and_persist(cbm_gbuf_t *gbuf, const char *db_path, const char *
     int dump_rc = cbm_gbuf_dump_to_sqlite(gbuf, db_path);
     cbm_log_info("incremental.dump", "rc", itoa_buf(dump_rc), "elapsed_ms",
                  itoa_buf((int)elapsed_ms(t)));
-
+    bool format_stamped = true;
     cbm_store_t *hash_store = cbm_store_open_path(db_path);
     if (hash_store) {
         bool hash_records_complete = persist_hashes(hash_store, project, files, file_count,
                                                     mode_skipped, mode_skipped_count);
 
+        /* #769: cbm_writer_open truncates and rebuilds the DB file from
+         * scratch on every dump, so the format version written by the last
+         * full index is gone by the time we get here. */
+        if (cbm_store_set_format_version(hash_store, CBM_INDEX_FORMAT_VERSION) != CBM_STORE_OK) {
+            cbm_log_error("incremental.err", "msg", "persist_format_version", "project", project);
+            format_stamped = false;
+        }
+
         /* Coverage rows (#963): re-write the merged set into the rebuilt DB
          * (AFTER hashes, so the deleted-file prune sees the live file set). */
         cbm_project_t project_info = {0};
@@ -705,6 +713,8 @@ static void dump_and_persist(cbm_gbuf_t *gbuf, const char *db_path, const char *
     if (repo_path && cbm_artifact_exists(repo_path)) {
         cbm_artifact_export(db_path, repo_path, project, CBM_ARTIFACT_FAST);
     }
+
+    return format_stamped ? 0 : CBM_NOT_FOUND;
 }
 
 /* ── Incremental pipeline entry point ────────────────────────────── */
@@ -1010,13 +1020,14 @@ int cbm_pipeline_run_incremental(cbm_pipeline_t *p, const char *db_path, cbm_fil
         .ignored_files_total = run_ignored_total,
         .coverage_version = 1,
     };
-    dump_and_persist(existing, db_path, project, files, file_count, mode_skipped,
-                     mode_skipped_count, cbm_pipeline_repo_path(p), cov, cov_n, &coverage_meta);
+    int dp_rc =
+        dump_and_persist(existing, db_path, project, files, file_count, mode_skipped,
+                         mode_skipped_count, cbm_pipeline_repo_path(p), cov, cov_n, &coverage_meta);
     free(cov);
     cbm_store_free_coverage(old_cov, old_cov_count);
     free_mode_skipped(mode_skipped, mode_skipped_count);
     cbm_gbuf_free(existing);
 
     cbm_log_info("incremental.done", "elapsed_ms", itoa_buf((int)elapsed_ms(t0)));
-    return 0;
+    return dp_rc;
 }

src/store/store.c

@@ -2918,6 +2918,36 @@ int cbm_store_list_files(cbm_store_t *s, const char *project, char ***out, int *
     return CBM_STORE_OK;
 }
 
+/* ── Index format version (PRAGMA user_version) ─────────────────── */
+
+int cbm_store_get_format_version(cbm_store_t *s, int *out) {
+    if (!s || !s->db || !out) {
+        return CBM_STORE_ERR;
+    }
+    *out = 0;
+    sqlite3_stmt *stmt = NULL;
+    if (sqlite3_prepare_v2(s->db, "PRAGMA user_version;", CBM_NOT_FOUND, &stmt, NULL) !=
+        SQLITE_OK) {
+        return CBM_STORE_ERR;
+    }
+    if (sqlite3_step(stmt) == SQLITE_ROW) {
+        *out = sqlite3_column_int(stmt, 0);
+    }
+    sqlite3_finalize(stmt);
+    return CBM_STORE_OK;
+}
+
+/* PRAGMA values cannot be bound; format the integer constant in. Writes, so
+ * only call on a read-write connection (never from configure_pragmas). */
+int cbm_store_set_format_version(cbm_store_t *s, int version) {
+    if (!s || !s->db) {
+        return CBM_STORE_ERR;
+    }
+    char sql[ST_BUF_64];
+    snprintf(sql, sizeof(sql), "PRAGMA user_version = %d;", version);
+    return exec_sql(s, sql);
+}
+
 /* ── Node neighbor names ──────────────────────────────────────── */
 
 static int query_neighbor_names(sqlite3 *db, const char *sql, int64_t node_id, int limit,

src/store/store.h

@@ -23,6 +23,7 @@ typedef struct cbm_store cbm_store_t;
 #define CBM_STORE_OK 0
 #define CBM_STORE_ERR (-1)
 #define CBM_STORE_NOT_FOUND (-2)
+#define CBM_INDEX_FORMAT_VERSION 1
 
 /* ── Data structures ────────────────────────────────────────────── */
 
@@ -77,6 +78,13 @@ void cbm_store_node_degree(cbm_store_t *s, int64_t node_id, int *in_deg, int *ou
  * Returns CBM_STORE_OK or CBM_STORE_ERR. */
 int cbm_store_list_files(cbm_store_t *s, const char *project, char ***out, int *count);
 
+/* Persisted index-format identity. Bump when a change alters the QN scheme
+ * or node identity of an already-written graph, so an old DB is routed
+ * through the full-reindex path instead of producing a mixed graph.
+ * 1 = File QNs keep the file extension (#769). */
+int cbm_store_get_format_version(cbm_store_t *s, int *out);
+int cbm_store_set_format_version(cbm_store_t *s, int version);
+
 /* Get caller/callee names for a node (CALLS/HTTP_CALLS/ASYNC_CALLS edges).
  * Returns 0 on success. Caller must free each out_callers[i]/out_callees[i]
  * and the arrays themselves. */

tests/test_fqn.c

@@ -89,6 +89,77 @@ TEST(fqn_compute_basic_rs) {
     PASS();
 }
 
+TEST(fqn_compute_file_sibling_distinct) {
+    char *ts = cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.ts", "__file__");
+    char *html = cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.html", "__file__");
+    char *scss = cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.scss", "__file__");
+    ASSERT_NOT_NULL(ts);
+    ASSERT_NOT_NULL(html);
+    ASSERT_NOT_NULL(scss);
+    ASSERT_STR_NEQ(ts, html);
+    ASSERT_STR_NEQ(ts, scss);
+    ASSERT_STR_NEQ(html, scss);
+    free(ts);
+    free(html);
+    free(scss);
+    PASS();
+}
+
+TEST(fqn_compute_symbol_still_strips) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.ts", "BadgeComponent"),
+               "proj.app.badge.badge.component.BadgeComponent");
+    PASS();
+}
+
+TEST(fqn_module_siblings_still_share) {
+    ASSERT_FQN(cbm_pipeline_fqn_module("proj", "app/badge/badge.component.ts"),
+               "proj.app.badge.badge.component");
+    ASSERT_FQN(cbm_pipeline_fqn_module("proj", "app/badge/badge.component.html"),
+               "proj.app.badge.badge.component");
+    PASS();
+}
+
+/* #769: exact __file__ QNs for component siblings. */
+TEST(fqn_compute_file_sibling_exact) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.ts", "__file__"),
+               "proj.app.badge.badge.component.ts.__file__");
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.html", "__file__"),
+               "proj.app.badge.badge.component.html.__file__");
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/badge/badge.component.scss", "__file__"),
+               "proj.app.badge.badge.component.scss.__file__");
+    PASS();
+}
+
+TEST(fqn_compute_file_multi_extension) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/badge/badge.spec.ts", "__file__"),
+               "proj.app.badge.badge.spec.ts.__file__");
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "archive.tar.gz", "__file__"),
+               "proj.archive.tar.gz.__file__");
+    PASS();
+}
+
+TEST(fqn_compute_file_no_extension) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "Makefile", "__file__"),
+               "proj.Makefile.__file__");
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "LICENSE", "__file__"),
+               "proj.LICENSE.__file__");
+    PASS();
+}
+
+TEST(fqn_compute_file_dotfile) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", ".gitignore", "__file__"),
+               "proj..gitignore.__file__");
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "config/.env", "__file__"),
+               "proj.config..env.__file__");
+    PASS();
+}
+
+TEST(fqn_compute_file_index_ts) {
+    ASSERT_FQN(cbm_pipeline_fqn_compute("proj", "app/index.ts", "__file__"),
+           "proj.app.index.ts.__file__");
+    PASS();
+}
+
 /* ── Nested paths ─────────────────────────────────────────────── */
 
 TEST(fqn_compute_nested_two_levels) {
@@ -595,6 +666,14 @@ SUITE(fqn) {
     RUN_TEST(fqn_file_qn_distinguishes_same_stem_header_source_issue964);
     RUN_TEST(fqn_module_qn_still_strips_extension);
     RUN_TEST(fqn_compute_basic_rs);
+    RUN_TEST(fqn_compute_file_sibling_distinct);
+    RUN_TEST(fqn_compute_symbol_still_strips);
+    RUN_TEST(fqn_module_siblings_still_share);
+    RUN_TEST(fqn_compute_file_sibling_exact);
+    RUN_TEST(fqn_compute_file_multi_extension);
+    RUN_TEST(fqn_compute_file_no_extension);
+    RUN_TEST(fqn_compute_file_dotfile);
+    RUN_TEST(fqn_compute_file_index_ts);
 
     /* fqn_compute: nested paths */
     RUN_TEST(fqn_compute_nested_two_levels);