fixup! perf: hash data files during combining to avoid unneeded work. #…

…1483

coverage/data.py

@@ -112,7 +112,8 @@ def combine_parallel_data(
         raise NoDataError("No data to combine")
 
     file_hashes = set()
-    files_combined = 0
+    combined_any = False
+
     for f in files_to_combine:
         if f == data.data_filename():
             # Sometimes we are combining into a file which is one of the
@@ -133,33 +134,37 @@ def combine_parallel_data(
             hasher = hashlib.new("sha3_256")
             hasher.update(fobj.read())
             sha = hasher.digest()
-            if sha in file_hashes:
-                if message:
-                    message(f"Skipping duplicate data {rel_file_name}")
-                continue
-            file_hashes.add(sha)
+            combine_this_one = sha not in file_hashes
 
-        if data._debug.should('dataio'):
-            data._debug.write(f"Combining data file {f!r}")
-        try:
-            new_data = CoverageData(f, debug=data._debug)
-            new_data.read()
-        except CoverageException as exc:
-            if data._warn:
-                # The CoverageException has the file name in it, so just
-                # use the message as the warning.
-                data._warn(str(exc))
+        delete_this_one = not keep
+        if combine_this_one:
+            if data._debug.should('dataio'):
+                data._debug.write(f"Combining data file {f!r}")
+            file_hashes.add(sha)
+            try:
+                new_data = CoverageData(f, debug=data._debug)
+                new_data.read()
+            except CoverageException as exc:
+                if data._warn:
+                    # The CoverageException has the file name in it, so just
+                    # use the message as the warning.
+                    data._warn(str(exc))
+                delete_this_one = False
+            else:
+                data.update(new_data, aliases=aliases)
+                combined_any = True
+                if message:
+                    message(f"Combined data file {rel_file_name}")
         else:
-            data.update(new_data, aliases=aliases)
-            files_combined += 1
             if message:
-                message(f"Combined data file {rel_file_name}")
-            if not keep:
-                if data._debug.should('dataio'):
-                    data._debug.write(f"Deleting combined data file {f!r}")
-                file_be_gone(f)
+                message(f"Skipping duplicate data {rel_file_name}")
+
+        if delete_this_one:
+            if data._debug.should('dataio'):
+                data._debug.write(f"Deleting data file {f!r}")
+            file_be_gone(f)
 
-    if strict and not files_combined:
+    if strict and not combined_any:
         raise NoDataError("No usable data files")
 
 

tests/test_concurrency.py

@@ -490,6 +490,7 @@ def try_multiprocessing_code(
                 )
                 for line in out_lines
             )
+            assert len(glob.glob(".coverage.*")) == 0
             out = self.run_command("coverage report -m")
 
             last_line = self.squeezed_lines(out)[-1]