fix: Use json output for severities

src/codechecker.bzl

@@ -274,7 +274,7 @@ _codechecker_test = rule(
         ),
         "severities": attr.string_list(
             default = ["HIGH"],
-            doc = "List of defect severities: HIGH, MEDIUM, LOW, STYLE etc",
+            doc = "List of defect severities or checker names: HIGH, MEDIUM, LOW, STYLE etc",
         ),
         "skip": attr.string_list(
             default = [],

src/common.py

@@ -9,7 +9,8 @@
 import subprocess
 import sys
 import os
-import re
+import json
+
 
 def setup_logging(verbosity, codechecker_log):
     """Setup logging parameters for execution session"""
@@ -53,9 +54,7 @@ def build_env(env, log, clang, clang_tidy):
     elif new_env.get("CC_ANALYZER_BIN"):
         logging.debug("CC_ANALYZER_BIN is set by the configuration")
     else:
-        new_env["CC_ANALYZER_BIN"] = (
-            f"clangsa:{clang};clang-tidy:{clang_tidy}"
-        )
+        new_env["CC_ANALYZER_BIN"] = f"clangsa:{clang};clang-tidy:{clang_tidy}"
     logging.debug("env: %s", str(new_env))
     return new_env
 
@@ -153,10 +152,7 @@ def parse(
     stage("CodeChecker parse:")
     env = build_env(env, log, clang, clang_tidy)
     logging.info("CodeChecker parse -e json")
-    codechecker_parse = (
-        f"{codechecker} parse --config "
-        f"{config} {input_dir}"
-    )
+    codechecker_parse = f"{codechecker} parse --config " f"{config} {input_dir}"
     # Save results to JSON file
     command = (
         f"{codechecker_parse} --export=json > " f"{input_dir}/result.json"
@@ -183,25 +179,27 @@ def check_results(output_dir, log, severities):
     Args:
         output_dir (str): Path to the output directory
         log (str): Path to the log file
-        severities (str): Comma separated list of defect severities to fail on
+        severities (str): Comma separated list of defect severities 
+                          or checker names to fail on.
     """
     stage("Checking result:")
     # Get results file and read it
-    result_file = output_dir + "/result.txt"
     logging.info("Find CodeChecker results in bazel-bin")
     logging.info("      all artifacts: %s/", output_dir)
     logging.info("      HTML report:   %s/report/index.html", output_dir)
-    logging.info("      result file:   %s", result_file)
-    results = read_file(log, result_file)
-    logging.info("Results: \n\n%s\n", results)
+    logging.info("      result file:   %s/result.txt", output_dir)
+    report_data = json.loads(read_file(log, output_dir + "/result.json"))
+    logging.info(
+        "Results: \n\n%s\n", read_file(log, output_dir + "/result.txt")
+    )
+    reports = report_data.get("reports", [])
     # Collect defect severities to detect
     if severities is None:
         fail(
             log,
-            "CodeChecker defect severities are invalid: "
-            f"{str(severities)}",
+            "CodeChecker defect severities are invalid: " f"{str(severities)}",
         )
-    severities = shlex.split(severities) # pyright: ignore[reportArgumentType]
+    severities = shlex.split(severities)  # pyright: ignore[reportArgumentType]
     # Add HIGH severity by default
     if not severities:
         severities.append("HIGH")
@@ -211,12 +209,20 @@ def check_results(output_dir, log, severities):
     logging.debug("Severities: %s", str(severities))
     issues = dict.fromkeys(severities, 0)
     logging.debug("Issues: %s", str(issues))
-    # Grep results for defects according to severities
-    for issue in issues:
-        found = re.findall(rf"^{issue} .* (\d+)", results, re.M)
-        defects = sum(int(number) for number in found)
-        logging.debug("   %s : %s = %d", issue, str(found), defects)
-        issues[issue] = defects
+    for report in reports:
+        severity = report.get("severity")
+        checker_name = report.get("checker_name", "unknown")
+        if severity in issues or checker_name in issues:
+            # We need to increment the key that was actually found.
+            matched_key = severity if severity in issues else checker_name
+            issues[matched_key] += 1
+            logging.debug(
+                "   %s : %s = %d:%d",
+                severity,
+                checker_name,
+                report.get("file", {}).get("path", "unknown"),
+                report.get("line", "?"),
+            )
     logging.info("Defects: %s", str(issues))
     # Check collected defects
     passed = True

test/unit/severities/severities_check.py

@@ -13,11 +13,11 @@
 # limitations under the License.
 
 """
-Drives common.check_results() against a real CodeChecker result.txt fixture.
+Drives common.check_results() against a real CodeChecker result.json fixture.
 
 Intended to be used as the main of a py_test Bazel target, one target per
 case, generated by the severities_test() macro. check_results() reads
-"<output_dir>/result.txt" and calls sys.exit(1) when it finds defects of the
+"<output_dir>/result.json" and calls sys.exit(1) when it finds defects of the
 requested severities or checker/bug names, so this driver copies the fixture
 into a temp dir, runs check_results(), and compares the outcome against the
 expected pass/fail.
@@ -36,14 +36,14 @@ def parse_args() -> argparse.Namespace:
     """Parse command-line arguments"""
     parser = argparse.ArgumentParser(
         description=(
-            "Run common.check_results() on a result.txt fixture and assert "
+            "Run common.check_results() on a result.json fixture and assert "
             "whether it passes or fails for the given severities."
         )
     )
     parser.add_argument(
         "--file",
         required=True,
-        help="Path to the result.txt fixture to check.",
+        help="Path to the result.json fixture to check.",
     )
     parser.add_argument(
         "--severities",
@@ -70,7 +70,14 @@ def run(fixture: str, severities) -> bool:
     """
     output_dir = tempfile.mkdtemp()
     try:
-        shutil.copyfile(fixture, os.path.join(output_dir, "result.txt"))
+        shutil.copyfile(fixture, os.path.join(output_dir, "result.json"))
+        # The function logs the content of result.txt
+        # Since it is not used in any meaningful capacity,
+        # an empty file is sufficient
+        with open(
+            os.path.join(output_dir, "result.txt"), "w", encoding="utf-8"
+        ) as handle:
+            handle.write("")
         try:
             check_results(
                 output_dir=output_dir, log=None, severities=severities

test/unit/severities/severities_test.bzl

@@ -67,12 +67,12 @@ def severities_test(
     # The producer label may be written relative to this package,
     # e.g. "producers:high", but we need an absolute label.
     if producer.startswith("//") or producer.startswith("@"):
-        result_txt = "%s/codechecker-files/result.txt" % producer
+        result_json = "%s/codechecker-files/result.json" % producer
     else:
-        result_txt = "//%s/%s/codechecker-files/result.txt" % (native.package_name(), producer)
+        result_json = "//%s/%s/codechecker-files/result.json" % (native.package_name(), producer)
 
     args = [
-        "--file=$(location {})".format(result_txt),
+        "--file=$(location {})".format(result_json),
         "--expect={}".format(expect),
     ]
 
@@ -85,7 +85,7 @@ def severities_test(
         srcs = ["severities_check.py"],
         main = "severities_check.py",
         args = args,
-        data = [result_txt],
+        data = [result_json],
         local = True,
         tags = ["unit"] + tags,
         size = size,