[Python-checkins] bpo-37064: Add option -k to Tools/scripts/pathfix.py (GH-15548)

[Victor Stinner]

https://github.com/python/cpython/commit/50254ac4c179cb412e90682098c97db786143929
commit: 50254ac4c179cb412e90682098c97db786143929
branch: master
author: PatrikKopkan <kopkanpatrik at gmail.com>
committer: Victor Stinner <vstinner at redhat.com>
date: 2019-09-05T16:54:54+02:00
summary:

bpo-37064: Add option -k to Tools/scripts/pathfix.py (GH-15548)

Add flag -k to pathscript.py script: preserve shebang flags.

files:
A Lib/test/test_tools/test_pathfix.py
A Misc/NEWS.d/next/Tools-Demos/2019-05-27-15-26-12.bpo-37064.k_SPW2.rst
M Tools/scripts/pathfix.py

diff --git a/Lib/test/test_tools/test_pathfix.py b/Lib/test/test_tools/test_pathfix.py
new file mode 100644
index 000000000000..5c0dd1d5a60c
--- /dev/null
+++ b/Lib/test/test_tools/test_pathfix.py
@@ -0,0 +1,64 @@
+import os
+import subprocess
+import sys
+import unittest
+from test import support
+from test.test_tools import import_tool, scriptsdir
+
+
+class TestPathfixFunctional(unittest.TestCase):
+    script = os.path.join(scriptsdir, 'pathfix.py')
+
+    def setUp(self):
+        self.temp_file = support.TESTFN
+        self.addCleanup(support.unlink, support.TESTFN)
+
+    def pathfix(self, shebang, pathfix_flags):
+        with open(self.temp_file, 'w', encoding='utf8') as f:
+            f.write(f'{shebang}\n' + 'print("Hello world")\n')
+
+        proc = subprocess.run(
+            [sys.executable, self.script,
+             *pathfix_flags, '-n', self.temp_file],
+            capture_output=True)
+        self.assertEqual(proc.returncode, 0, proc)
+
+        with open(self.temp_file, 'r', encoding='utf8') as f:
+            output = f.read()
+
+        lines = output.split('\n')
+        self.assertEqual(lines[1:], ['print("Hello world")', ''])
+        shebang = lines[0]
+        return shebang
+
+    def test_pathfix(self):
+        self.assertEqual(
+            self.pathfix(
+                '#! /usr/bin/env python',
+                ['-i', '/usr/bin/python3',]),
+            '#! /usr/bin/python3',
+        )
+        self.assertEqual(
+            self.pathfix(
+                '#! /usr/bin/env python -R',
+                ['-i', '/usr/bin/python3', ]),
+            '#! /usr/bin/python3',
+        )
+
+    def test_pathfix_keeping_flags(self):
+        self.assertEqual(
+            self.pathfix(
+                '#! /usr/bin/env python -R',
+                ['-i', '/usr/bin/python3', '-k',]),
+            '#! /usr/bin/python3 -R',
+        )
+        self.assertEqual(
+            self.pathfix(
+                '#! /usr/bin/env python',
+                ['-i', '/usr/bin/python3', '-k',]),
+            '#! /usr/bin/python3',
+        )
+
+
+if __name__ == '__main__':
+    unittest.main()
diff --git a/Misc/NEWS.d/next/Tools-Demos/2019-05-27-15-26-12.bpo-37064.k_SPW2.rst b/Misc/NEWS.d/next/Tools-Demos/2019-05-27-15-26-12.bpo-37064.k_SPW2.rst
new file mode 100644
index 000000000000..02599077c9d5
--- /dev/null
+++ b/Misc/NEWS.d/next/Tools-Demos/2019-05-27-15-26-12.bpo-37064.k_SPW2.rst
@@ -0,0 +1 @@
+Add flag -k to pathscript.py script: preserve shebang flags.
\ No newline at end of file
diff --git a/Tools/scripts/pathfix.py b/Tools/scripts/pathfix.py
index 1a0cf1c9e69c..eee684e3502d 100755
--- a/Tools/scripts/pathfix.py
+++ b/Tools/scripts/pathfix.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 
-# Change the #! line occurring in Python scripts.  The new interpreter
+# Change the #! line (shebang) occurring in Python scripts.  The new interpreter
 # pathname must be given with a -i option.
 #
 # Command line arguments are files or directories to be processed.
@@ -10,7 +10,12 @@
 # arguments).
 # The original file is kept as a back-up (with a "~" attached to its name),
 # -n flag can be used to disable this.
-#
+
+# Sometimes you may find shebangs with flags such as `#! /usr/bin/env python -si`.
+# Normally, pathfix overwrites the entire line, including the flags.
+# To change interpreter and keep flags from the original shebang line, use -k.
+
+
 # Undoubtedly you can do this using find and sed or perl, but this is
 # a nice example of Python code that recurses down a directory tree
 # and uses regular expressions.  Also note several subtleties like
@@ -33,16 +38,19 @@
 new_interpreter = None
 preserve_timestamps = False
 create_backup = True
+keep_flags = False
 
 
 def main():
     global new_interpreter
     global preserve_timestamps
     global create_backup
-    usage = ('usage: %s -i /interpreter -p -n file-or-directory ...\n' %
+    global keep_flags
+
+    usage = ('usage: %s -i /interpreter -p -n -k file-or-directory ...\n' %
              sys.argv[0])
     try:
-        opts, args = getopt.getopt(sys.argv[1:], 'i:pn')
+        opts, args = getopt.getopt(sys.argv[1:], 'i:kpn')
     except getopt.error as msg:
         err(str(msg) + '\n')
         err(usage)
@@ -54,6 +62,8 @@ def main():
             preserve_timestamps = True
         if o == '-n':
             create_backup = False
+        if o == '-k':
+            keep_flags = True
     if not new_interpreter or not new_interpreter.startswith(b'/') or \
            not args:
         err('-i option or file-or-directory missing\n')
@@ -70,10 +80,14 @@ def main():
             if fix(arg): bad = 1
     sys.exit(bad)
 
+
 ispythonprog = re.compile(r'^[a-zA-Z0-9_]+\.py$')
+
+
 def ispython(name):
     return bool(ispythonprog.match(name))
 
+
 def recursedown(dirname):
     dbg('recursedown(%r)\n' % (dirname,))
     bad = 0
@@ -96,6 +110,7 @@ def recursedown(dirname):
         if recursedown(fullname): bad = 1
     return bad
 
+
 def fix(filename):
 ##  dbg('fix(%r)\n' % (filename,))
     try:
@@ -164,12 +179,26 @@ def fix(filename):
     # Return success
     return 0
 
+
+def parse_shebang(shebangline):
+    shebangline = shebangline.rstrip(b'\n')
+    start = shebangline.find(b' -')
+    if start == -1:
+        return b''
+    return shebangline[start:]
+
+
 def fixline(line):
     if not line.startswith(b'#!'):
         return line
+
     if b"python" not in line:
         return line
-    return b'#! ' + new_interpreter + b'\n'
+    flags = b''
+    if keep_flags:
+        flags = parse_shebang(line)
+    return b'#! ' + new_interpreter + flags + b'\n'
+
 
 if __name__ == '__main__':
     main()