Repair coverage tests to make them work with coverage 6.1, which changed the...

Repair coverage tests to make them work with coverage 6.1, which changed the HTML output format that the tests parse.

tests/run/coverage_cmd.srctree

@@ -176,23 +176,33 @@ def run_xml_report():
 
 
 def run_html_report():
+    from collections import defaultdict
+
     stdout = run_coverage_command('html', '-d', 'html')
-    _parse_lines = re.compile(
-        r'<p[^>]* id=["\'][^0-9"\']*(?P<id>[0-9]+)[^0-9"\']*["\'][^>]*'
-        r' class=["\'][^"\']*(?P<run>mis|run)[^"\']*["\']').findall
+    # coverage 6.1+ changed the order of the attributes => need to parse them separately
+    _parse_id = re.compile(r'id=["\'][^0-9"\']*(?P<id>[0-9]+)[^0-9"\']*["\']').search
+    _parse_state = re.compile(r'class=["\'][^"\']*(?P<state>mis|run|exc)[^"\']*["\']').search
 
     files = {}
     for file_path in iglob('html/*.html'):
         with open(file_path) as f:
             page = f.read()
-        executed = set()
-        missing = set()
-        for line, has_run in _parse_lines(page):
-            (executed if has_run == 'run' else missing).add(int(line))
-        files[file_path] = (executed, missing)
-
-    executed, missing = [data for path, data in files.items() if 'coverage_test_pyx' in path][0]
-    assert executed
+        report = defaultdict(set)
+        for line in re.split(r'id=["\']source["\']', page)[-1].splitlines():
+            lineno = _parse_id(line)
+            state = _parse_state(line)
+            if not lineno or not state:
+                continue
+            report[state.group('state')].add(int(lineno.group('id')))
+        files[file_path] = report
+
+    for filename, report in files.items():
+        if "coverage_test_pyx" not in filename:
+            continue
+        executed = report["run"]
+        missing = report["mis"]
+        excluded = report["exc"]
+        assert executed, (filename, report)
         assert 5 in executed, executed
         assert 6 in executed, executed
         assert 7 in executed, executed

tests/run/coverage_cmd_src_layout.srctree

@@ -133,20 +133,25 @@ def run_xml_report():
 
 
 def run_html_report():
+    from collections import defaultdict
+
     stdout = run_coverage_command('html', '-d', 'html')
-    _parse_lines = re.compile(
-        r'<p[^>]* id=["\'][^0-9"\']*(?P<id>[0-9]+)[^0-9"\']*["\'][^>]*'
-        r' class=["\'][^"\']*(?P<run>mis|run)[^"\']*["\']').findall
+    # coverage 6.1+ changed the order of the attributes => need to parse them separately
+    _parse_id = re.compile(r'id=["\'][^0-9"\']*(?P<id>[0-9]+)[^0-9"\']*["\']').search
+    _parse_state = re.compile(r'class=["\'][^"\']*(?P<state>mis|run|exc)[^"\']*["\']').search
 
     files = {}
     for file_path in iglob('html/*.html'):
         with open(file_path) as f:
             page = f.read()
-        executed = set()
-        missing = set()
-        for line, has_run in _parse_lines(page):
-            (executed if has_run == 'run' else missing).add(int(line))
-        files[file_path] = (executed, missing)
+        report = defaultdict(set)
+        for line in re.split(r'id=["\']source["\']', page)[-1].splitlines():
+            lineno = _parse_id(line)
+            state = _parse_state(line)
+            if not lineno or not state:
+                continue
+            report[state.group('state')].add(int(lineno.group('id')))
+        files[file_path] = (report['run'], report['mis'])
 
     executed, missing = [data for path, data in files.items() if 'trivial_module' in path][0]
     assert executed