Improve error handling on signature mismatch cases

and cleanup old todos that have been addressed

bin/sarif-extract-scans

@@ -43,7 +43,6 @@ def load(fname):
         try:
             content = json.load(fp)
         except json.decoder.JSONDecodeError as err:
-            # TODO knewbury error handling
             logging.error('Error reading from {}: {}: line {}, column {}'
                           .format(fname, err.msg, err.lineno, err.colno))
             status_writer.file_load_error["file"] = fname
@@ -69,18 +68,14 @@ sarif_struct = signature.fillsig(args, sarif_struct, context)
 
 #
 # Use reference type graph (signature) to traverse sarif and attach values to tables
-#
-# try:
-#     tgraph = typegraph.Typegraph(signature_single.struct_graph_2022_02_01)
-#     typegraph.destructure(tgraph, signature_single.start_node_2022_02_01, sarif_struct)
-# except json.decoder.JSONDecodeError as err:
-#     logging.error('Error reading from {}: {}: line {}, column {}'
-#                           .format(args.file, err.msg, err.lineno, err.colno))
-#     sys.exit(1)
-tgraph = typegraph.Typegraph(signature_single.struct_graph_2022_02_01)
-typegraph.destructure(tgraph, signature_single.start_node_2022_02_01, sarif_struct)
-# may have gathered warnings below, if not does nothing
-status_writer.csv_write_warnings()
+try:
+    tgraph = typegraph.Typegraph(signature_single.struct_graph_2022_02_01)
+    typegraph.destructure(tgraph, signature_single.start_node_2022_02_01, sarif_struct)
+except Exception:
+    # will have gathered errors/warnings
+    status_writer.csv_write_warnings()
+    #pass the exception up to be put into log by runner
+    raise(Exception)
 
 #
 # Form output tables

sarif_cli/scan_tables.py

@@ -126,7 +126,7 @@ def joins_for_scans(basetables, external_info, scantables):
         "id"                   : e.scan_id,
         "commit_id"            : b.project.revisionId[0],
         "project_id"           : e.project_id,
-        # TODO extract real date information from somewhere
+        # TODO extract real date information from somewhere external
         "db_create_start"      : pd.Timestamp(0.0, unit='s'),
         "db_create_stop"       : pd.Timestamp(0.0, unit='s'),
         "scan_start_date"      : pd.Timestamp(0.0, unit='s'),
@@ -165,7 +165,6 @@ def joins_for_results(basetables, external_info):
         res = pd.concat(stack)
     else:
         if stack == []:
-            # TODO knewbury to error handling
             logging.warning("Zero problem/path_problem results found in sarif "
                             "file but processing anyway.")
             status_writer.csv_write(status_writer.zero_results)

sarif_cli/typegraph.py

@@ -113,7 +113,8 @@ def destructure(typegraph: Typegraph, node: NodeId, tree: Tree):
     elif t in [str, int, bool]:
         pass
     else:
-        # TODO knewbury error handling
+        status_writer.unknown_sarif_parsing_shape["extra_info"] = "Unhandled type: %s" % t
+        status_writer.csv_write(status_writer.unknown_sarif_parsing_shape)
         raise Exception("Unhandled type: %s" % t)
 
 def _destructure_dict_1(typegraph, node, tree):
@@ -139,7 +140,7 @@ def _destructure_dict_1(typegraph, node, tree):
     # Sanity check
     sig = typegraph.signature_graph[node]
     if type(sig) != tuple:
-        # TODO knewbury error handling
+        # TODO add error handling?
         raise SignatureMismatch()
 
     # Destructure this dictionary
@@ -160,10 +161,8 @@ def _destructure_dict(typegraph: Typegraph, node, tree):
     type_fields = typegraph.fields[node]
     if tree_fields == type_fields:
         _destructure_dict_1(typegraph, node, tree)
-        # TODO knewbury error handling here
     elif set(tree_fields).issuperset(set(type_fields)):
         # Log a warning
-        # log.warning("XX: Tree has unrecognized fields")
         logging.warning('Input tree has unrecognized fields, collecting only '
                         'known entries: {}'.format(tree))
         logging.warning('tree fields: {}'.format(sorted(tree_fields)))
@@ -189,7 +188,6 @@ def _destructure_dict(typegraph: Typegraph, node, tree):
         )
 
     else:
-        # TODO knewbury error handling
         status_writer.unknown_sarif_parsing_shape["extra_info"] = "type fields {} do not match tree fields {}.".format(type_fields, tree_fields)
         status_writer.csv_write(status_writer.unknown_sarif_parsing_shape)
         raise Exception("typegraph: unhandled case reached: cannot match type "