Merge pull request diffblue#508 from diffblue/remove_slicer_from_stage_1

Remove taint-slicer from stage 1 [SEC-567]

Author: smowton

driver/analyser.py

@@ -92,7 +92,7 @@ def run_security_analyser(config_json_file, cmdline):
 
 
 def run_program_slicing(
-        cfgs,
+        cfg,
         results_dir,
         timeout,
         verbosity,
@@ -101,59 +101,48 @@ def run_program_slicing(
     prof = {"calling_slicer": []}
     prof_start_time = time.time()
 
-    result = []
-    for cfg in cfgs:
-        src_plain_fname = os.path.splitext(os.path.basename(cfg["goto_binary_file"]))[0]
-        src_idx_name = int(src_plain_fname[src_plain_fname.rfind("_")+1:])
-        dst_goto_program_fname = os.path.join(results_dir, "sliced_goto_program_" + str(src_idx_name) + ".gbf")
-        dst_goto_stats_fname = os.path.join(results_dir, "sliced_goto_program_" + str(src_idx_name) + ".stats.json")
+    dst_goto_program_fname = os.path.join(results_dir, "sliced_goto_program.gbf")
+    dst_goto_stats_fname = os.path.join(results_dir, "sliced_goto_program.stats.json")
 
+    command = (
+        get_goto_instrument_pathname() + " " +
+        "--reachability-slice " +
+        # "--full-slice " +
+        "--verbosity " + str(verbosity) + " " +
+        cfg["goto_binary_file"] + " " +
+        dst_goto_program_fname + " " +
+        "--save-code-statistics " + dst_goto_stats_fname
+        )
+    prof_calling_goto_instrument_start_time = time.time()
+    print("Invoking 'goto-instrument' ...")
+    with utility.PushCwd(results_dir) as cwd:
+        if verbosity >= 9:
+            print("CWD: " + cwd.get())
+            print("CMD: " + command)
+        os.system(command)
+        #shutil.copy(cfg["goto_binary_file"], dst_goto_program_fname)
+
+    prof["calling_slicer"].append({
+        "duration": time.time() - prof_calling_goto_instrument_start_time
+        })
+
+    result = cfg.copy()
+    result["goto_binary_file"] = dst_goto_program_fname
+
+    if dump_html_slice:
         command = (
-            get_goto_instrument_pathname() + " " +
-            "--reachability-slice " +
-            # "--full-slice " +
-            "--verbosity " + str(verbosity) + " " +
-            cfg["goto_binary_file"] + " " +
+            get_security_analyser_pathname() + " " +
             dst_goto_program_fname + " " +
-            "--save-code-statistics " + dst_goto_stats_fname
+            "--dump-html-program " +
+            "'" + os.path.join(results_dir, "SLICED_PROGRAM_IN_HTML") + "'"
             )
-        prof_calling_goto_instrument_start_time = time.time()
-        print("Invoking 'goto-instrument' ...")
         with utility.PushCwd(results_dir) as cwd:
             if verbosity >= 9:
                 print("CWD: " + cwd.get())
                 print("CMD: " + command)
             os.system(command)
-            #shutil.copy(cfg["goto_binary_file"], dst_goto_program_fname)
-
-        prof["calling_slicer"].append({
-            "gbf_idx": src_idx_name,
-            "duration": time.time() - prof_calling_goto_instrument_start_time
-            })
-
-        result.append(cfg.copy())
-        result[-1]["goto_binary_file"] = dst_goto_program_fname
-
-        if dump_html_slice:
-            command = (
-                get_security_analyser_pathname() + " " +
-                dst_goto_program_fname + " " +
-                "--dump-html-program " +
-                "'" +
-                os.path.join(
-                    results_dir,
-                    "HTML_dump_of_sliced_GOTO_programs",
-                    os.path.splitext(os.path.basename(dst_goto_program_fname))[0]
-                    ) +
-                "'"
-                )
-            with utility.PushCwd(results_dir) as cwd:
-                if verbosity >= 9:
-                    print("CWD: " + cwd.get())
-                    print("CMD: " + command)
-                os.system(command)
 
-    with open(os.path.join(results_dir, "sliced_goto_programs.json"), "w") as results_json:
+    with open(os.path.join(results_dir, "sliced_goto_program.json"), "w") as results_json:
         results_json.write(json.dumps(result, sort_keys=False, indent=4))
 
     prof["duration"] = time.time() - prof_start_time
@@ -207,44 +196,43 @@ def run_search_for_error_traces(
     traces_list = []
 
     with open(json_cfg_fname) as data_file:
-        cfgs = json.load(data_file)
-    for cfg in cfgs:
-        old_cwd = os.getcwd()
-        os.chdir(results_dir)
-        # verbosity must be at least 4 for error traces to get printed!
-        command = (
-            get_jbmc_pathname() + " " +
-            cfg["goto_binary_file"] + " " +
-            " --trace --trace-json-extended --no-refine-strings --json-ui --unwind 10 " +
-            "--verbosity " + str(max(verbosity, 4)) + " "
-            "> \"jbmc_results.json\""
-            )
-        prof["calling_jbmc"] = {}
-        prof_calling_jbmc_start_time = time.time()
-        print("Invoking 'jbmc' ...")
-        if verbosity >= 9:
-            print("CWD: " + results_dir)
-            print("CMD: " + command)
-        os.system(command)
-        prof["calling_jbmc"]["duration"] = time.time() - prof_calling_jbmc_start_time
-        os.chdir(old_cwd)
-
-        with open(os.path.join(results_dir,"jbmc_results.json")) as data_file:
-            jbmc_results = json.load(data_file)
-        for outer_element in jbmc_results:
-            if "result" in outer_element:
-                for inner_element in outer_element["result"]:
-                    if "trace" in inner_element and len(inner_element["trace"]) > 0:
-                        trace=inner_element["trace"]
-                        was_trace_saved = append_jbmc_error_trace_if_it_is_relevant(
-                                                trace,
-                                                trace_id,
-                                                results_dir,
-                                                cfg,
-                                                traces_list
-                                                )
-                        if was_trace_saved:
-                            trace_id += 1
+        cfg = json.load(data_file)
+    old_cwd = os.getcwd()
+    os.chdir(results_dir)
+    # verbosity must be at least 4 for error traces to get printed!
+    command = (
+        get_jbmc_pathname() + " " +
+        cfg["goto_binary_file"] + " " +
+        " --trace --trace-json-extended --no-refine-strings --json-ui --unwind 10 " +
+        "--verbosity " + str(max(verbosity, 4)) + " "
+        "> \"jbmc_results.json\""
+        )
+    prof["calling_jbmc"] = {}
+    prof_calling_jbmc_start_time = time.time()
+    print("Invoking 'jbmc' ...")
+    if verbosity >= 9:
+        print("CWD: " + results_dir)
+        print("CMD: " + command)
+    os.system(command)
+    prof["calling_jbmc"]["duration"] = time.time() - prof_calling_jbmc_start_time
+    os.chdir(old_cwd)
+
+    with open(os.path.join(results_dir,"jbmc_results.json")) as data_file:
+        jbmc_results = json.load(data_file)
+    for outer_element in jbmc_results:
+        if "result" in outer_element:
+            for inner_element in outer_element["result"]:
+                if "trace" in inner_element and len(inner_element["trace"]) > 0:
+                    trace=inner_element["trace"]
+                    was_trace_saved = append_jbmc_error_trace_if_it_is_relevant(
+                                            trace,
+                                            trace_id,
+                                            results_dir,
+                                            cfg,
+                                            traces_list
+                                            )
+                    if was_trace_saved:
+                        trace_id += 1
 
     with open(os.path.join(results_dir, "error_traces.json"), "w") as traces_file:
         traces_file.write(json.dumps(traces_list, sort_keys=True, indent=4))

driver/mkbench.py

@@ -213,7 +213,7 @@ def collect_java_binaries(cmdline):
             "verbosity": cmdline.verbosity,
             "dump_html_summaries": cmdline.dump_html_summaries,
             "dump_html_statistics": cmdline.dump_html_statistics,
-            "dump_html_slice": cmdline.dump_html_slice,
+            "dump_html_instrumented_goto": cmdline.dump_html_instrumented_goto,
             "dump_html_program": cmdline.dump_html_program,
             "do_not_use_precise_access_paths": cmdline.do_not_use_precise_access_paths,
             "output-dir": "./",