fix: use human sorting on human-readable things

Author: nedbat

CHANGES.rst

@@ -23,7 +23,9 @@ This list is detailed and covers changes in each pre-release version.
 Unreleased
 ----------
 
-Nothing yet.
+- When sorting human-readable names, numeric components are sorted correctly:
+  file10.py will appear after file9.py.  This applies to file names, module
+  names, environment variables, and test contexts.
 
 
 .. _changes_602:

coverage/cmdline.py

@@ -20,6 +20,7 @@
 from coverage.debug import info_formatter, info_header, short_stack
 from coverage.exceptions import BaseCoverageException, ExceptionDuringRun, NoSource
 from coverage.execfile import PyRunner
+from coverage.misc import human_sorted
 from coverage.results import Numbers, should_fail_under
 
 
@@ -780,7 +781,7 @@ def do_debug(self, args):
                 if data:
                     print(f"has_arcs: {data.has_arcs()!r}")
                     summary = line_counts(data, fullpath=True)
-                    filenames = sorted(summary.keys())
+                    filenames = human_sorted(summary.keys())
                     print(f"\n{len(filenames)} files:")
                     for f in filenames:
                         line = f"{f}: {summary[f]} lines"

coverage/collector.py

@@ -10,7 +10,7 @@
 from coverage.debug import short_stack
 from coverage.disposition import FileDisposition
 from coverage.exceptions import CoverageException
-from coverage.misc import isolate_module
+from coverage.misc import human_sorted, isolate_module
 from coverage.pytracer import PyTracer
 
 os = isolate_module(os)
@@ -352,7 +352,7 @@ def pause(self):
             stats = tracer.get_stats()
             if stats:
                 print("\nCoverage.py tracer stats:")
-                for k in sorted(stats.keys()):
+                for k in human_sorted(stats.keys()):
                     print(f"{k:>20}: {stats[k]}")
         if self.threading:
             self.threading.settrace(None)

coverage/control.py

@@ -26,7 +26,7 @@
 from coverage.html import HtmlReporter
 from coverage.inorout import InOrOut
 from coverage.jsonreport import JsonReporter
-from coverage.misc import bool_or_none, join_regex
+from coverage.misc import bool_or_none, join_regex, human_sorted, human_sorted_items
 from coverage.misc import DefaultValue, ensure_dir_for_file, isolate_module
 from coverage.plugin import FileReporter
 from coverage.plugin_support import Plugins
@@ -309,7 +309,7 @@ def _write_startup_debug(self):
         wrote_any = False
         with self._debug.without_callers():
             if self._debug.should('config'):
-                config_info = sorted(self.config.__dict__.items())
+                config_info = human_sorted_items(self.config.__dict__.items())
                 config_info = [(k, v) for k, v in config_info if not k.startswith('_')]
                 write_formatted_info(self._debug, "config", config_info)
                 wrote_any = True
@@ -1076,7 +1076,7 @@ def plugin_info(plugins):
             ('pid', os.getpid()),
             ('cwd', os.getcwd()),
             ('path', sys.path),
-            ('environment', sorted(
+            ('environment', human_sorted(
                 f"{k} = {v}"
                 for k, v in os.environ.items()
                 if (

coverage/files.py

@@ -14,7 +14,7 @@
 
 from coverage import env
 from coverage.exceptions import CoverageException
-from coverage.misc import contract, join_regex, isolate_module
+from coverage.misc import contract, human_sorted, isolate_module, join_regex
 
 
 os = isolate_module(os)
@@ -199,7 +199,7 @@ class TreeMatcher:
 
     """
     def __init__(self, paths, name):
-        self.original_paths = sorted(paths)
+        self.original_paths = human_sorted(paths)
         self.paths = list(map(os.path.normcase, paths))
         self.name = name
 

coverage/html.py

@@ -15,6 +15,7 @@
 from coverage.exceptions import CoverageException
 from coverage.files import flat_rootname
 from coverage.misc import ensure_dir, file_be_gone, Hasher, isolate_module, format_local_datetime
+from coverage.misc import human_sorted
 from coverage.report import get_analysis_to_report
 from coverage.results import Numbers
 from coverage.templite import Templite
@@ -123,7 +124,7 @@ def data_for_file(self, fr, analysis):
             contexts = contexts_label = None
             context_list = None
             if category and self.config.show_contexts:
-                contexts = sorted(c or self.EMPTY for c in contexts_by_lineno.get(lineno, ()))
+                contexts = human_sorted(c or self.EMPTY for c in contexts_by_lineno.get(lineno, ()))
                 if contexts == [self.EMPTY]:
                     contexts_label = self.EMPTY
                 else:

coverage/misc.py

@@ -389,3 +389,34 @@ def import_local_file(modname, modfile=None):
     spec.loader.exec_module(mod)
 
     return mod
+
+
+def human_key(s):
+    """Turn a string into a list of string and number chunks.
+        "z23a" -> ["z", 23, "a"]
+    """
+    def tryint(s):
+        """If `s` is a number, return an int, else `s` unchanged."""
+        try:
+            return int(s)
+        except ValueError:
+            return s
+
+    return [tryint(c) for c in re.split(r"(\d+)", s)]
+
+def human_sorted(strings):
+    """Sort the given iterable of strings the way that humans expect.
+
+    Numeric components in the strings are sorted as numbers.
+
+    Returns the sorted list.
+
+    """
+    return sorted(strings, key=human_key)
+
+def human_sorted_items(items, reverse=False):
+    """Sort the (string, value) items the way humans expect.
+
+    Returns the sorted list of items.
+    """
+    return sorted(items, key=lambda pair: (human_key(pair[0]), pair[1]), reverse=reverse)

coverage/summary.py

@@ -6,6 +6,7 @@
 import sys
 
 from coverage.exceptions import CoverageException
+from coverage.misc import human_sorted_items
 from coverage.report import get_analysis_to_report
 from coverage.results import Numbers
 
@@ -89,15 +90,17 @@ def report(self, morfs, outfile=None):
             lines.append((text, args))
 
         # Sort the lines and write them out.
-        if getattr(self.config, 'sort', None):
-            sort_option = self.config.sort.lower()
-            reverse = False
-            if sort_option[0] == '-':
-                reverse = True
-                sort_option = sort_option[1:]
-            elif sort_option[0] == '+':
-                sort_option = sort_option[1:]
-
+        sort_option = (self.config.sort or "name").lower()
+        reverse = False
+        if sort_option[0] == '-':
+            reverse = True
+            sort_option = sort_option[1:]
+        elif sort_option[0] == '+':
+            sort_option = sort_option[1:]
+
+        if sort_option == "name":
+            lines = human_sorted_items(lines, reverse=reverse)
+        else:
             position = column_order.get(sort_option)
             if position is None:
                 raise CoverageException(f"Invalid sorting option: {self.config.sort!r}")

coverage/xmlreport.py

@@ -10,7 +10,7 @@
 import xml.dom.minidom
 
 from coverage import __url__, __version__, files
-from coverage.misc import isolate_module
+from coverage.misc import isolate_module, human_sorted, human_sorted_items
 from coverage.report import get_analysis_to_report
 
 os = isolate_module(os)
@@ -77,7 +77,7 @@ def report(self, morfs, outfile=None):
         xcoverage.appendChild(xsources)
 
         # Populate the XML DOM with the source info.
-        for path in sorted(self.source_paths):
+        for path in human_sorted(self.source_paths):
             xsource = self.xml_out.createElement("source")
             xsources.appendChild(xsource)
             txt = self.xml_out.createTextNode(path)
@@ -90,13 +90,13 @@ def report(self, morfs, outfile=None):
         xcoverage.appendChild(xpackages)
 
         # Populate the XML DOM with the package info.
-        for pkg_name, pkg_data in sorted(self.packages.items()):
+        for pkg_name, pkg_data in human_sorted_items(self.packages.items()):
             class_elts, lhits, lnum, bhits, bnum = pkg_data
             xpackage = self.xml_out.createElement("package")
             xpackages.appendChild(xpackage)
             xclasses = self.xml_out.createElement("classes")
             xpackage.appendChild(xclasses)
-            for _, class_elt in sorted(class_elts.items()):
+            for _, class_elt in human_sorted_items(class_elts.items()):
                 xclasses.appendChild(class_elt)
             xpackage.setAttribute("name", pkg_name.replace(os.sep, '.'))
             xpackage.setAttribute("line-rate", rate(lhits, lnum))

tests/test_misc.py

@@ -10,6 +10,7 @@
 from coverage.exceptions import CoverageException
 from coverage.misc import contract, dummy_decorator_with_args, file_be_gone
 from coverage.misc import Hasher, one_of, substitute_variables, import_third_party
+from coverage.misc import human_sorted, human_sorted_items
 from coverage.misc import USE_CONTRACTS
 
 from tests.coveragetest import CoverageTest
@@ -180,3 +181,23 @@ def test_failure(self):
         mod = import_third_party("xyzzy")
         assert mod is None
         assert "xyzzy" not in sys.modules
+
+
+HUMAN_DATA = [
+    ("z1 a2z a2a a3 a1", "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"),
+]
+
+@pytest.mark.parametrize("words, ordered", HUMAN_DATA)
+def test_human_sorted(words, ordered):
+    assert " ".join(human_sorted(words.split())) == ordered
+
+@pytest.mark.parametrize("words, ordered", HUMAN_DATA)
+def test_human_sorted_items(words, ordered):
+    keys = words.split()
+    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]]
+    assert human_sorted_items(items) == oitems
+    assert human_sorted_items(items, reverse=True) == oitems[::-1]

tests/test_summary.py

@@ -847,8 +847,8 @@ def get_summary_text(self, *options):
         """
         self.make_rigged_file("file1.py", 339, 155)
         self.make_rigged_file("file2.py", 13, 3)
-        self.make_rigged_file("file3.py", 234, 228)
-        self.make_file("doit.py", "import file1, file2, file3")
+        self.make_rigged_file("file10.py", 234, 228)
+        self.make_file("doit.py", "import file1, file2, file10")
 
         cov = Coverage(source=["."], omit=["doit.py"])
         cov.start()
@@ -871,7 +871,7 @@ def test_test_data(self):
         # ------------------------------
         # file1.py     339    155    54%
         # file2.py      13      3    77%
-        # file3.py     234    228     3%
+        # file10.py    234    228     3%
         # ------------------------------
         # TOTAL        586    386    34%
 
@@ -906,30 +906,40 @@ def assert_ordering(self, text, *words):
         msg = f"The words {words!r} don't appear in order in {text!r}"
         assert indexes == sorted(indexes), msg
 
+    def test_default_sort_report(self):
+        # Sort the text report by the default (Name) column.
+        report = self.get_summary_text()
+        self.assert_ordering(report, "file1.py", "file2.py", "file10.py")
+
+    def test_sort_report_by_name(self):
+        # Sort the text report explicitly by the Name column.
+        report = self.get_summary_text(('report:sort', 'Name'))
+        self.assert_ordering(report, "file1.py", "file2.py", "file10.py")
+
     def test_sort_report_by_stmts(self):
         # Sort the text report by the Stmts column.
         report = self.get_summary_text(('report:sort', 'Stmts'))
-        self.assert_ordering(report, "file2.py", "file3.py", "file1.py")
+        self.assert_ordering(report, "file2.py", "file10.py", "file1.py")
 
     def test_sort_report_by_missing(self):
         # Sort the text report by the Missing column.
         report = self.get_summary_text(('report:sort', 'Miss'))
-        self.assert_ordering(report, "file2.py", "file1.py", "file3.py")
+        self.assert_ordering(report, "file2.py", "file1.py", "file10.py")
 
     def test_sort_report_by_cover(self):
         # Sort the text report by the Cover column.
         report = self.get_summary_text(('report:sort', 'Cover'))
-        self.assert_ordering(report, "file3.py", "file1.py", "file2.py")
+        self.assert_ordering(report, "file10.py", "file1.py", "file2.py")
 
     def test_sort_report_by_cover_plus(self):
         # Sort the text report by the Cover column, including the explicit + sign.
         report = self.get_summary_text(('report:sort', '+Cover'))
-        self.assert_ordering(report, "file3.py", "file1.py", "file2.py")
+        self.assert_ordering(report, "file10.py", "file1.py", "file2.py")
 
     def test_sort_report_by_cover_reversed(self):
         # Sort the text report by the Cover column reversed.
         report = self.get_summary_text(('report:sort', '-Cover'))
-        self.assert_ordering(report, "file2.py", "file1.py", "file3.py")
+        self.assert_ordering(report, "file2.py", "file1.py", "file10.py")
 
     def test_sort_report_by_invalid_option(self):
         # Sort the text report by a nonsense column.