gh-98692: Enable treating shebang lines as executables in py.exe launcher (GH-98732)

Author: zooba

Doc/using/windows.rst

@@ -866,7 +866,6 @@ minor version. I.e. ``/usr/bin/python3.7-32`` will request usage of the
    not provably i386/32-bit". To request a specific environment, use the new
    ``-V:<TAG>`` argument with the complete tag.
 
-
 The ``/usr/bin/env`` form of shebang line has one further special property.
 Before looking for installed Python interpreters, this form will search the
 executable :envvar:`PATH` for a Python executable. This corresponds to the
@@ -876,6 +875,13 @@ be found, it will be handled as described below. Additionally, the environment
 variable :envvar:`PYLAUNCHER_NO_SEARCH_PATH` may be set (to any value) to skip
 this additional search.
 
+Shebang lines that do not match any of these patterns are treated as **Windows**
+paths that are absolute or relative to the directory containing the script file.
+This is a convenience for Windows-only scripts, such as those generated by an
+installer, since the behavior is not compatible with Unix-style shells.
+These paths may be quoted, and may include multiple arguments, after which the
+path to the script and any additional arguments will be appended.
+
 
 Arguments in shebang lines
 --------------------------

Lib/test/test_launcher.py

@@ -516,6 +516,14 @@ def test_py_shebang(self):
         self.assertEqual("3.100", data["SearchInfo.tag"])
         self.assertEqual(f"X.Y.exe -prearg {script} -postarg", data["stdout"].strip())
 
+    def test_python_shebang(self):
+        with self.py_ini(TEST_PY_COMMANDS):
+            with self.script("#! python -prearg") as script:
+                data = self.run_py([script, "-postarg"])
+        self.assertEqual("PythonTestSuite", data["SearchInfo.company"])
+        self.assertEqual("3.100", data["SearchInfo.tag"])
+        self.assertEqual(f"X.Y.exe -prearg {script} -postarg", data["stdout"].strip())
+
     def test_py2_shebang(self):
         with self.py_ini(TEST_PY_COMMANDS):
             with self.script("#! /usr/bin/python2 -prearg") as script:
@@ -617,3 +625,42 @@ def test_install(self):
             self.assertIn("winget.exe", cmd)
         # Both command lines include the store ID
         self.assertIn("9PJPW5LDXLZ5", cmd)
+
+    def test_literal_shebang_absolute(self):
+        with self.script(f"#! C:/some_random_app -witharg") as script:
+            data = self.run_py([script])
+        self.assertEqual(
+            f"C:\\some_random_app -witharg {script}",
+            data["stdout"].strip(),
+        )
+
+    def test_literal_shebang_relative(self):
+        with self.script(f"#! ..\\some_random_app -witharg") as script:
+            data = self.run_py([script])
+        self.assertEqual(
+            f"{script.parent.parent}\\some_random_app -witharg {script}",
+            data["stdout"].strip(),
+        )
+
+    def test_literal_shebang_quoted(self):
+        with self.script(f'#! "some random app" -witharg') as script:
+            data = self.run_py([script])
+        self.assertEqual(
+            f'"{script.parent}\\some random app" -witharg {script}',
+            data["stdout"].strip(),
+        )
+
+        with self.script(f'#! some" random "app -witharg') as script:
+            data = self.run_py([script])
+        self.assertEqual(
+            f'"{script.parent}\\some random app" -witharg {script}',
+            data["stdout"].strip(),
+        )
+
+    def test_literal_shebang_quoted_escape(self):
+        with self.script(f'#! some\\" random "app -witharg') as script:
+            data = self.run_py([script])
+        self.assertEqual(
+            f'"{script.parent}\\some\\ random app" -witharg {script}',
+            data["stdout"].strip(),
+        )

Misc/NEWS.d/next/Windows/2022-10-26-17-43-09.gh-issue-98692.bOopfZ.rst

@@ -0,0 +1,2 @@
+Fix the :ref:`launcher` ignoring unrecognized shebang lines instead of
+treating them as local paths

PC/launcher2.c

@@ -871,6 +871,62 @@ _findCommand(SearchInfo *search, const wchar_t *command, int commandLength)
 }
 
 
+int
+_useShebangAsExecutable(SearchInfo *search, const wchar_t *shebang, int shebangLength)
+{
+    wchar_t buffer[MAXLEN];
+    wchar_t script[MAXLEN];
+    wchar_t command[MAXLEN];
+
+    int commandLength = 0;
+    int inQuote = 0;
+
+    if (!shebang || !shebangLength) {
+        return 0;
+    }
+
+    wchar_t *pC = command;
+    for (int i = 0; i < shebangLength; ++i) {
+        wchar_t c = shebang[i];
+        if (isspace(c) && !inQuote) {
+            commandLength = i;
+            break;
+        } else if (c == L'"') {
+            inQuote = !inQuote;
+        } else if (c == L'/' || c == L'\\') {
+            *pC++ = L'\\';
+        } else {
+            *pC++ = c;
+        }
+    }
+    *pC = L'\0';
+
+    if (!GetCurrentDirectoryW(MAXLEN, buffer) ||
+        wcsncpy_s(script, MAXLEN, search->scriptFile, search->scriptFileLength) ||
+        FAILED(PathCchCombineEx(buffer, MAXLEN, buffer, script,
+                                PATHCCH_ALLOW_LONG_PATHS)) ||
+        FAILED(PathCchRemoveFileSpec(buffer, MAXLEN)) ||
+        FAILED(PathCchCombineEx(buffer, MAXLEN, buffer, command,
+                                PATHCCH_ALLOW_LONG_PATHS))
+    ) {
+        return RC_NO_MEMORY;
+    }
+
+    int n = (int)wcsnlen(buffer, MAXLEN);
+    wchar_t *path = allocSearchInfoBuffer(search, n + 1);
+    if (!path) {
+        return RC_NO_MEMORY;
+    }
+    wcscpy_s(path, n + 1, buffer);
+    search->executablePath = path;
+    if (commandLength) {
+        search->executableArgs = &shebang[commandLength];
+        search->executableArgsLength = shebangLength - commandLength;
+    }
+    return 0;
+}
+
+
 int
 checkShebang(SearchInfo *search)
 {
@@ -963,13 +1019,19 @@ checkShebang(SearchInfo *search)
         L"/usr/bin/env ",
         L"/usr/bin/",
         L"/usr/local/bin/",
-        L"",
+        L"python",
         NULL
     };
 
     for (const wchar_t **tmpl = shebangTemplates; *tmpl; ++tmpl) {
         if (_shebangStartsWith(shebang, shebangLength, *tmpl, &command)) {
             commandLength = 0;
+            // Normally "python" is the start of the command, but we also need it
+            // as a shebang prefix for back-compat. We move the command marker back
+            // if we match on that one.
+            if (0 == wcscmp(*tmpl, L"python")) {
+                command -= 6;
+            }
             while (command[commandLength] && !isspace(command[commandLength])) {
                 commandLength += 1;
             }
@@ -1012,11 +1074,14 @@ checkShebang(SearchInfo *search)
                 debug(L"# Found shebang command but could not execute it: %.*s\n",
                     commandLength, command);
             }
-            break;
+            // search is done by this point
+            return 0;
         }
     }
 
-    return 0;
+    // Unrecognised commands are joined to the script's directory and treated
+    // as the executable path
+    return _useShebangAsExecutable(search, shebang, shebangLength);
 }