fix: leading zeros can confuse human sorting. #1709

Author: nedbat

CHANGES.rst

@@ -24,6 +24,10 @@ Unreleased
   matching any of the lines, closing `issue 684`_.  Thanks, `Jan Rusak,
   Maciej Kowalczyk and Joanna Ejzel <pull 1705_>`_.
 
+- Fix: XML reports could fail with a TypeError if files had numeric components
+  that were duplicates except for leading zeroes, like ``file1.py`` and
+  ``file001.py``.  Fixes `issue 1709`_.
+
 - The ``coverage annotate`` command used to announce that it would be removed
   in a future version. Enough people got in touch to say that they use it, so
   it will stay.  Don't expect it to keep up with other new features though.
@@ -36,6 +40,7 @@ Unreleased
 
 .. _issue 684: https://github.com/nedbat/coveragepy/issues/684
 .. _pull 1705: https://github.com/nedbat/coveragepy/pull/1705
+.. _issue 1709: https://github.com/nedbat/coveragepy/issues/1709
 
 
 .. scriv-start-here

coverage/misc.py

@@ -338,9 +338,13 @@ def import_local_file(modname: str, modfile: Optional[str] = None) -> ModuleType
     return mod
 
 
-def _human_key(s: str) -> List[Union[str, int]]:
+def _human_key(s: str) -> Tuple[List[Union[str, int]], str]:
     """Turn a string into a list of string and number chunks.
-        "z23a" -> ["z", 23, "a"]
+
+    "z23a" -> (["z", 23, "a"], "z23a")
+
+    The original string is appended as a last value to ensure the
+    key is unique enough so that "x1y" and "x001y" can be distinguished.
     """
     def tryint(s: str) -> Union[str, int]:
         """If `s` is a number, return an int, else `s` unchanged."""
@@ -349,7 +353,7 @@ def tryint(s: str) -> Union[str, int]:
         except ValueError:
             return s
 
-    return [tryint(c) for c in re.split(r"(\d+)", s)]
+    return ([tryint(c) for c in re.split(r"(\d+)", s)], s)
 
 def human_sorted(strings: Iterable[str]) -> List[str]:
     """Sort the given iterable of strings the way that humans expect.

tests/test_misc.py

@@ -137,7 +137,7 @@ def test_failure(self) -> None:
 
 
 HUMAN_DATA = [
-    ("z1 a2z a2a a3 a1", "a1 a2a a2z a3 z1"),
+    ("z1 a2z a01 a2a a3 a1", "a01 a1 a2a a2z a3 z1"),
     ("a10 a9 a100 a1", "a1 a9 a10 a100"),
     ("4.0 3.10-win 3.10-mac 3.9-mac 3.9-win", "3.9-mac 3.9-win 3.10-mac 3.10-win 4.0"),
 ]
@@ -149,6 +149,8 @@ def test_human_sorted(words: str, ordered: str) -> None:
 @pytest.mark.parametrize("words, ordered", HUMAN_DATA)
 def test_human_sorted_items(words: str, ordered: str) -> None:
     keys = words.split()
+    # Check that we never try to compare the values in the items
+    human_sorted_items([(k, object()) for k in keys])
     items = [(k, 1) for k in keys] + [(k, 2) for k in keys]
     okeys = ordered.split()
     oitems = [(k, v) for k in okeys for v in [1, 2]]

tests/test_xml.py

@@ -350,6 +350,20 @@ def test_no_duplicate_packages(self) -> None:
         named_sub_package = dom.findall(".//package[@name='namespace.package.subpackage']")
         assert len(named_sub_package) == 1
 
+    def test_bug_1709(self) -> None:
+        # https://github.com/nedbat/coveragepy/issues/1709
+        self.make_file("main.py", "import x1y, x01y, x001y")
+        self.make_file("x1y.py", "print('x1y')")
+        self.make_file("x01y.py", "print('x01y')")
+        self.make_file("x001y.py", "print('x001y')")
+
+        cov = coverage.Coverage()
+        self.start_import_stop(cov, "main")
+        assert self.stdout() == "x1y\nx01y\nx001y\n"
+        # This used to raise:
+        # TypeError: '<' not supported between instances of 'Element' and 'Element'
+        cov.xml_report()
+
 
 def unbackslash(v: Any) -> Any:
     """Find strings in `v`, and replace backslashes with slashes throughout."""