Merge pull request #158 from jeromekelleher/c_0.99.2_fixes

C 0.99.2 fixes
tskit-dev · Mar 27, 2019 · ed2cc1f · ed2cc1f
2 parents 02e2421 + 5d68315
commit ed2cc1f
Show file tree

Hide file tree

Showing 6 changed files with 209 additions and 4 deletions.
diff --git a/c/CHANGELOG.rst b/c/CHANGELOG.rst
@@ -1,3 +1,12 @@
+---------------------
+[0.99.2] - 2019-03-27
+---------------------
+
+Bugfix release. Changes:
+
+- Fix incorrect errors on tbl_collection_dump (#132)
+- Catch table overflows (#157)
+
 ---------------------
 [0.99.1] - 2019-01-24
 ---------------------

diff --git a/c/tests/test_tables.c b/c/tests/test_tables.c
@@ -274,12 +274,17 @@ test_table_collection_dump_errors(void)
 
     ret = tsk_table_collection_init(&tables, 0);
     CU_ASSERT_EQUAL_FATAL(ret, 0);
+    tables.sequence_length = 1.0;
+
     ret = tsk_table_collection_dump(&tables, "/", 0);
     CU_ASSERT_TRUE(tsk_is_kas_error(ret));
     CU_ASSERT_EQUAL_FATAL(ret ^ (1 << TSK_KAS_ERR_BIT), KAS_ERR_IO);
     str = tsk_strerror(ret);
     CU_ASSERT_TRUE(strlen(str) > 0);
 
+    /* We'd like to catch close errors also, but it's hard to provoke them
+     * without intercepting calls to fclose() */
+
     tsk_table_collection_free(&tables);
 }
 static void
@@ -2098,6 +2103,99 @@ test_load_reindex(void)
     tsk_treeseq_free(&ts);
 }
 
+static void
+test_table_overflow(void)
+{
+    int ret;
+    tsk_table_collection_t tables;
+    tsk_size_t max_rows = ((tsk_size_t) INT32_MAX) + 1;
+
+    ret = tsk_table_collection_init(&tables, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, 0);
+
+    /* Simulate overflows */
+    tables.individuals.max_rows = max_rows;
+    tables.individuals.num_rows = max_rows;
+    ret = tsk_individual_table_add_row(&tables.individuals, 0, 0, 0, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.nodes.max_rows = max_rows;
+    tables.nodes.num_rows = max_rows;
+    ret = tsk_node_table_add_row(&tables.nodes, 0, 0, 0, 0, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.edges.max_rows = max_rows;
+    tables.edges.num_rows = max_rows;
+    ret = tsk_edge_table_add_row(&tables.edges, 0, 0, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.migrations.max_rows = max_rows;
+    tables.migrations.num_rows = max_rows;
+    ret = tsk_migration_table_add_row(&tables.migrations, 0, 0, 0, 0, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.sites.max_rows = max_rows;
+    tables.sites.num_rows = max_rows;
+    ret = tsk_site_table_add_row(&tables.sites, 0, 0, 0, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.mutations.max_rows = max_rows;
+    tables.mutations.num_rows = max_rows;
+    ret = tsk_mutation_table_add_row(&tables.mutations, 0, 0, 0, 0, 0, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.provenances.max_rows = max_rows;
+    tables.provenances.num_rows = max_rows;
+    ret = tsk_provenance_table_add_row(&tables.provenances, 0, 0, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tables.populations.max_rows = max_rows;
+    tables.populations.num_rows = max_rows;
+    ret = tsk_population_table_add_row(&tables.populations, 0, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_TABLE_OVERFLOW);
+
+    tsk_table_collection_free(&tables);
+}
+
+static void
+test_column_overflow(void)
+{
+    int ret;
+    tsk_table_collection_t tables;
+    tsk_size_t too_big = ((tsk_size_t) INT32_MAX) + 2;
+
+    ret = tsk_table_collection_init(&tables, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, 0);
+
+    ret = tsk_individual_table_add_row(&tables.individuals, 0, NULL, too_big, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+    ret = tsk_individual_table_add_row(&tables.individuals, 0, NULL, 0, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    ret = tsk_node_table_add_row(&tables.nodes, 0, 0, 0, 0, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    ret = tsk_site_table_add_row(&tables.sites, 0, NULL, too_big, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+    ret = tsk_site_table_add_row(&tables.sites, 0, NULL, 0, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    ret = tsk_mutation_table_add_row(&tables.mutations, 0, 0, 0, NULL, 0, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+    ret = tsk_mutation_table_add_row(&tables.mutations, 0, 0, 0, NULL, too_big, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    ret = tsk_provenance_table_add_row(&tables.provenances, NULL, too_big, NULL, 0);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+    ret = tsk_provenance_table_add_row(&tables.provenances, NULL, 0, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    ret = tsk_population_table_add_row(&tables.populations, NULL, too_big);
+    CU_ASSERT_EQUAL_FATAL(ret, TSK_ERR_COLUMN_OVERFLOW);
+
+    tsk_table_collection_free(&tables);
+}
+
 int
 main(int argc, char **argv)
 {
@@ -2124,8 +2222,11 @@ main(int argc, char **argv)
         {"test_dump_load_empty", test_dump_load_empty},
         {"test_dump_load_unsorted", test_dump_load_unsorted},
         {"test_load_reindex", test_load_reindex},
+        {"test_table_overflow", test_table_overflow},
+        {"test_column_overflow", test_column_overflow},
         {NULL, NULL},
     };
 
+
     return test_main(tests, argc, argv);
 }
diff --git a/c/tskit/core.c b/c/tskit/core.c
@@ -293,6 +293,12 @@ tsk_strerror_internal(int err)
         case TSK_ERR_TABLES_NOT_INDEXED:
             ret = "Table collection must be indexed";
             break;
+        case TSK_ERR_TABLE_OVERFLOW:
+            ret = "Table too large; cannot allocate more than 2**31 rows.";
+            break;
+        case TSK_ERR_COLUMN_OVERFLOW:
+            ret = "Table column too large; cannot be more than 2**31 bytes.";
+            break;
 
         /* Limitations */
         case TSK_ERR_ONLY_INFINITE_SITES:

diff --git a/c/tskit/core.h b/c/tskit/core.h
@@ -80,7 +80,7 @@ to the API or ABI are introduced, i.e., the addition of a new function.
 The library patch version. Incremented when any changes not relevant to the
 to the API or ABI are introduced, i.e., internal refactors of bugfixes.
 */
-#define TSK_VERSION_PATCH   1
+#define TSK_VERSION_PATCH   2
 /** @} */
 
 /* Node flags */
@@ -189,6 +189,8 @@ of tskit.
 #define TSK_ERR_BAD_TABLE_POSITION                                  -700
 #define TSK_ERR_BAD_SEQUENCE_LENGTH                                 -701
 #define TSK_ERR_TABLES_NOT_INDEXED                                  -702
+#define TSK_ERR_TABLE_OVERFLOW                                      -703
+#define TSK_ERR_COLUMN_OVERFLOW                                     -704
 
 /* Limitations */
 #define TSK_ERR_ONLY_INFINITE_SITES                                 -800