[3.11] pythonGH-109408: Move the C file whitespace check from patchcheck to pre-commit (pythonGH-109890)

Co-authored-by: Hugo van Kemenade <[email protected]>.
(cherry picked from commit f5edb56)

Co-authored-by: Adam Turner <[email protected]>

Author: AA-Turner

.pre-commit-config.yaml

@@ -16,6 +16,32 @@ repos:
       - id: trailing-whitespace
         types_or: [c, inc, python, rst]
 
+  - repo: local
+    hooks:
+      - id: python-file-whitespace
+        name: "Check Python file whitespace"
+        entry: 'python Tools/patchcheck/reindent.py --nobackup --newline LF'
+        language: 'system'
+        types: [python]
+        exclude: '^(Lib/test/tokenizedata/|Tools/c-analyzer/cpython/_parser).*$'
+
+  - repo: local
+    hooks:
+      - id: c-file-whitespace
+        name: "Check C file whitespace"
+        entry: "python Tools/patchcheck/untabify.py"
+        language: "system"
+        types_or: ['c', 'c++']
+        # Don't check the style of vendored libraries
+        exclude: |
+          (?x)^(
+            Modules/_ctypes/libffi_osx/.*
+            | Modules/_ctypes/libffi_msvc/.*
+            | Modules/_decimal/.*
+            | Modules/libmpdec/.*
+            | Modules/expat/.*
+          )$
+
   - repo: https://github.com/sphinx-contrib/sphinx-lint
     rev: v0.6.8
     hooks:

Tools/scripts/patchcheck.py

@@ -1,30 +1,15 @@
 #!/usr/bin/env python3
 """Check proposed changes for common issues."""
-import re
 import sys
-import shutil
 import os.path
 import subprocess
 import sysconfig
 
-import reindent
-import untabify
-
-
 def get_python_source_dir():
     src_dir = sysconfig.get_config_var('abs_srcdir')
     if not src_dir:
         src_dir = sysconfig.get_config_var('srcdir')
     return os.path.abspath(src_dir)
-
-
-# Excluded directories which are copies of external libraries:
-# don't check their coding style
-EXCLUDE_DIRS = [os.path.join('Modules', '_ctypes', 'libffi_osx'),
-                os.path.join('Modules', '_ctypes', 'libffi_msvc'),
-                os.path.join('Modules', '_decimal', 'libmpdec'),
-                os.path.join('Modules', 'expat'),
-                os.path.join('Modules', 'zlib')]
 SRCDIR = get_python_source_dir()
 
 
@@ -154,50 +139,8 @@ def changed_files(base_branch=None):
     else:
         sys.exit('need a git checkout to get modified files')
 
-    filenames2 = []
-    for filename in filenames:
-        # Normalize the path to be able to match using .startswith()
-        filename = os.path.normpath(filename)
-        if any(filename.startswith(path) for path in EXCLUDE_DIRS):
-            # Exclude the file
-            continue
-        filenames2.append(filename)
-
-    return filenames2
-
-
-def report_modified_files(file_paths):
-    count = len(file_paths)
-    if count == 0:
-        return n_files_str(count)
-    else:
-        lines = [f"{n_files_str(count)}:"]
-        for path in file_paths:
-            lines.append(f"  {path}")
-        return "\n".join(lines)
-
-
-@status("Fixing Python file whitespace", info=report_modified_files)
-def normalize_whitespace(file_paths):
-    """Make sure that the whitespace for .py files have been normalized."""
-    reindent.makebackup = False  # No need to create backups.
-    fixed = [path for path in file_paths if path.endswith('.py') and
-             reindent.check(os.path.join(SRCDIR, path))]
-    return fixed
-
-
-@status("Fixing C file whitespace", info=report_modified_files)
-def normalize_c_whitespace(file_paths):
-    """Report if any C files """
-    fixed = []
-    for path in file_paths:
-        abspath = os.path.join(SRCDIR, path)
-        with open(abspath, 'r') as f:
-            if '\t' not in f.read():
-                continue
-        untabify.process(abspath, 8, verbose=False)
-        fixed.append(path)
-    return fixed
+    # Normalize the path to be able to match using str.startswith()
+    return list(map(os.path.normpath, filenames))
 
 
 @status("Docs modified", modal=True)
@@ -237,38 +180,12 @@ def regenerated_pyconfig_h_in(file_paths):
         return "not needed"
 
 
-def ci(pull_request):
-    if pull_request == 'false':
-        print('Not a pull request; skipping')
-        return
-    base_branch = get_base_branch()
-    file_paths = changed_files(base_branch)
-    python_files = [fn for fn in file_paths if fn.endswith('.py')]
-    c_files = [fn for fn in file_paths if fn.endswith(('.c', '.h'))]
-    fixed = []
-    fixed.extend(normalize_whitespace(python_files))
-    fixed.extend(normalize_c_whitespace(c_files))
-    if not fixed:
-        print('No whitespace issues found')
-    else:
-        count = len(fixed)
-        print(f'Please fix the {n_files_str(count)} with whitespace issues')
-        print('(on Unix you can run `make patchcheck` to make the fixes)')
-        sys.exit(1)
-
-
 def main():
     base_branch = get_base_branch()
     file_paths = changed_files(base_branch)
-    python_files = [fn for fn in file_paths if fn.endswith('.py')]
-    c_files = [fn for fn in file_paths if fn.endswith(('.c', '.h'))]
     doc_files = [fn for fn in file_paths if fn.startswith('Doc') and
                  fn.endswith(('.rst', '.inc'))]
     misc_files = {p for p in file_paths if p.startswith('Misc')}
-    # PEP 8 whitespace rules enforcement.
-    normalize_whitespace(python_files)
-    # C rules enforcement.
-    normalize_c_whitespace(c_files)
     # Docs updated.
     docs_modified(doc_files)
     # Misc/ACKS changed.
@@ -281,19 +198,14 @@ def main():
     regenerated_pyconfig_h_in(file_paths)
 
     # Test suite run and passed.
-    if python_files or c_files:
-        end = " and check for refleaks?" if c_files else "?"
-        print()
-        print("Did you run the test suite" + end)
+    has_c_files = any(fn for fn in file_paths if fn.endswith(('.c', '.h')))
+    has_python_files = any(fn for fn in file_paths if fn.endswith('.py'))
+    print()
+    if has_c_files:
+        print("Did you run the test suite and check for refleaks?")
+    elif has_python_files:
+        print("Did you run the test suite?")
 
 
 if __name__ == '__main__':
-    import argparse
-    parser = argparse.ArgumentParser(description=__doc__)
-    parser.add_argument('--ci',
-                        help='Perform pass/fail checks')
-    args = parser.parse_args()
-    if args.ci:
-        ci(args.ci)
-    else:
-        main()
+    main()

Tools/scripts/untabify.py

@@ -21,8 +21,7 @@ def main():
         if optname == '-t':
             tabsize = int(optvalue)
 
-    for filename in args:
-        process(filename, tabsize)
+    return max(process(filename, tabsize) for filename in args)
 
 
 def process(filename, tabsize, verbose=True):
@@ -32,10 +31,10 @@ def process(filename, tabsize, verbose=True):
             encoding = f.encoding
     except IOError as msg:
         print("%r: I/O error: %s" % (filename, msg))
-        return
+        return 2
     newtext = text.expandtabs(tabsize)
     if newtext == text:
-        return
+        return 0
     backup = filename + "~"
     try:
         os.unlink(backup)
@@ -49,7 +48,8 @@ def process(filename, tabsize, verbose=True):
         f.write(newtext)
     if verbose:
         print(filename)
+    return 1
 
 
 if __name__ == '__main__':
-    main()
+    raise SystemExit(main())