fix for symbols in read-csv

Author: ser-vasilich

core/symbols.c

@@ -147,7 +147,7 @@ i64_t symbols_intern(lit_p str, i64_t len) {
         raw_p node = __atomic_load_n(&symbols->strings_node, __ATOMIC_ACQUIRE);
         i64_t rounds2 = 0;
 
-        while (slot_end > node) {
+        while ((i64_t)node == NULL_I64 || slot_end > node) {
             if ((i64_t)node == NULL_I64) {
                 backoff_spin(&rounds2);
                 node = __atomic_load_n(&symbols->strings_node, __ATOMIC_ACQUIRE);

tests/lang.c

@@ -5280,3 +5280,23 @@ test_result_t test_lang_safety() {
 
     PASS();
 }
+
+// ==================== READ-CSV LARGE SYMBOLS ====================
+test_result_t test_lang_read_csv() {
+    i64_t i;
+    FILE *f;
+
+    f = fopen("/tmp/rf_test_syms.csv", "w");
+    TEST_ASSERT(f != NULL, "failed to create temp CSV");
+    fprintf(f, "id,sym\n");
+    for (i = 0; i < 20000; i++)
+        fprintf(f, "%lld,s%lld\n", (long long)i, (long long)i);
+    fclose(f);
+
+    TEST_ASSERT_EQ(
+        "(count (read-csv [I64 SYMBOL] \"/tmp/rf_test_syms.csv\"))",
+        "20000");
+
+    remove("/tmp/rf_test_syms.csv");
+    PASS();
+}

tests/main.c

@@ -261,6 +261,7 @@ test_entry_t tests[] = {
     {"test_lang_do_let", test_lang_do_let},
     {"test_lang_error", test_lang_error},
     {"test_lang_safety", test_lang_safety},
+    {"test_lang_read_csv", test_lang_read_csv},
     // Parted table tests
     {"test_parted_load", test_parted_load},
     {"test_parted_select_where_date", test_parted_select_where_date},
@@ -455,6 +456,7 @@ test_entry_t tests[] = {
     {"test_string_list_ops", test_string_list_ops},
     {"test_string_where_in", test_string_where_in},
     {"test_string_combined", test_string_combined},
+    {"test_symbols_intern_large", test_symbols_intern_large},
     // Vector operation tests
     {"test_vec_til", test_vec_til},
     {"test_vec_take", test_vec_take},

tests/string_ops.c

@@ -432,3 +432,28 @@ test_result_t test_string_combined() {
 
     PASS();
 }
+
+// ==================== SYMBOL INTERN LARGE ====================
+test_result_t test_symbols_intern_large() {
+    symbols_p symbols = runtime_get()->symbols;
+    i64_t i, id;
+    char buf[32];
+
+    for (i = 0; i < 20000; i++) {
+        i64_t len = snprintf(buf, sizeof(buf), "sym_%lld", (long long)i);
+        id = symbols_intern(buf, len);
+        TEST_ASSERT(id != NULL_I64, "symbols_intern returned NULL_I64");
+    }
+
+    TEST_ASSERT(symbols_count(symbols) >= 20000, "expected >= 20000 symbols");
+
+    for (i = 0; i < 20000; i++) {
+        i64_t len = snprintf(buf, sizeof(buf), "sym_%lld", (long long)i);
+        id = symbols_intern(buf, len);
+        str_p s = str_from_symbol(id);
+        TEST_ASSERT(s != NULL, "str_from_symbol returned NULL");
+        TEST_ASSERT(str_cmp(s, len, buf, len) == 0, "symbol string mismatch");
+    }
+
+    PASS();
+}