From cf5fa3959cf33425a6495d137c0f793fb21dfcd8 Mon Sep 17 00:00:00 2001
From: Richard Berger <richard.berger@temple.edu>
Date: Fri, 12 Jun 2020 21:42:17 -0400
Subject: [PATCH] Add first version of whitespace utility

---
 cmake/CMakeLists.txt                |   1 +
 cmake/Modules/CodingStandard.cmake  |  21 +++++
 tools/coding_standard/whitespace.py | 132 ++++++++++++++++++++++++++++
 3 files changed, 154 insertions(+)
 create mode 100644 cmake/Modules/CodingStandard.cmake
 create mode 100644 tools/coding_standard/whitespace.py

diff --git a/cmake/CMakeLists.txt b/cmake/CMakeLists.txt
index 2407d03a80..c0563565e0 100644
--- a/cmake/CMakeLists.txt
+++ b/cmake/CMakeLists.txt
@@ -676,6 +676,7 @@ endif()
 
 include(Testing)
 include(CodeCoverage)
+include(CodingStandard)
 
 ###############################################################################
 # Print package summary
diff --git a/cmake/Modules/CodingStandard.cmake b/cmake/Modules/CodingStandard.cmake
new file mode 100644
index 0000000000..d7440fba9d
--- /dev/null
+++ b/cmake/Modules/CodingStandard.cmake
@@ -0,0 +1,21 @@
+if(CMAKE_VERSION VERSION_LESS 3.12)
+    find_package(PythonInterp) # Deprecated since version 3.12
+    if(PYTHONINTERP_FOUND)
+        set(Python_EXECUTABLE ${PYTHON_EXECUTABLE})
+    endif()
+else()
+    find_package(Python COMPONENTS Interpreter)
+endif()
+
+if (Python_EXECUTABLE)
+    add_custom_target(
+      check-whitespace
+      ${Python_EXECUTABLE} ${LAMMPS_TOOLS_DIR}/coding_standard/whitespace.py .
+      WORKING_DIRECTORY  ${LAMMPS_DIR}
+      COMMENT "Check for whitespace errors")
+    add_custom_target(
+      fix-whitespace
+      ${Python_EXECUTABLE} ${LAMMPS_TOOLS_DIR}/coding_standard/whitespace.py -f .
+      WORKING_DIRECTORY  ${LAMMPS_DIR}
+      COMMENT "Fix whitespace errors")
+endif()
diff --git a/tools/coding_standard/whitespace.py b/tools/coding_standard/whitespace.py
new file mode 100644
index 0000000000..f9923893a0
--- /dev/null
+++ b/tools/coding_standard/whitespace.py
@@ -0,0 +1,132 @@
+#!/usr/bin/env python3
+# Utility for detecting and fixing whitespace issues in LAMMPS
+#
+# Written by Richard Berger (Temple University)
+import os
+import glob
+import re
+import yaml
+import argparse
+import shutil
+
+DEFAULT_CONFIG = """
+recursive: true
+include:
+    - cmake/**
+    - doc
+    - doc/src/**
+    - python
+    - src/**
+patterns:
+    - ".gitignore"
+    - "README"
+    - "requirements.txt"
+    - "*.c"
+    - "*.cpp"
+    - "*.h"
+    - "*.sh"
+    - "*.py"
+    - "*.md"
+    - "*.rst"
+    - "*.cmake"
+"""
+
+def check_trailing_whitespace(f):
+    pattern = re.compile(r'\s+\n$')
+    last_line = "\n"
+    lineno = 1
+    errors = set()
+
+    for line in f:
+        if pattern.match(line):
+            errors.add(lineno)
+        last_line = line
+        lineno += 1
+
+    return errors, last_line
+
+def check_file(path):
+    encoding = 'UTF-8'
+    last_line = "\n"
+    whitespace_errors = set()
+    try:
+        with open(path, 'r') as f:
+            whitespace_errors, last_line = check_trailing_whitespace(f)
+    except UnicodeDecodeError:
+        encoding = 'ISO-8859-1'
+        try:
+            with open(path, 'r', encoding=encoding) as f:
+                whitespace_errors, last_line = check_trailing_whitespace(f)
+        except Exception:
+            encoding = 'unknown'
+
+    return {
+        'whitespace_errors': whitespace_errors,
+        'encoding': encoding,
+        'eof_error': not last_line.endswith('\n')
+    }
+
+def fix_file(path, check_result):
+    newfile = path + ".modified"
+    with open(newfile, 'w', encoding='UTF-8') as out:
+        with open(path, 'r', encoding=check_result['encoding']) as src:
+            for line in src:
+                print(line.rstrip(), file=out)
+    shutil.move(newfile, path)
+
+def check_folder(directory, config, fix=False, verbose=False):
+    files = []
+
+    for base_path in config['include']:
+        for pattern in config['patterns']:
+            path = os.path.join(directory, base_path, pattern)
+            files += glob.glob(path, recursive=config['recursive'])
+
+    for f in files:
+        path = os.path.normpath(f)
+
+        if verbose:
+            print("Checking file:", path)
+
+        result = check_file(path)
+
+        has_resolvable_errors = False
+
+        for lineno in result['whitespace_errors']:
+            print("[Error] Trailing whitespace @ {}:{}".format(path, lineno))
+            has_resolvable_errors = True
+
+        if result['eof_error']:
+            print("[Error] Missing newline at end of file @ {}".format(path))
+            has_resolvable_errors = True
+
+        if result['encoding'] == 'unknown':
+            print("[Error] Unknown text encoding @ {}".format(path))
+            has_resolvable_errors = False
+        elif result['encoding'] == 'ISO-8859-1':
+            print("[Error] Found ISO-8859-1 encoding instead of UTF-8 @ {}".format(path))
+            has_resolvable_errors = True
+
+        if has_resolvable_errors and fix:
+            print("Applying automatic fixes to file:", path)
+            fix_file(path, result)
+
+
+def main():
+    parser = argparse.ArgumentParser(description='Utility for detecting and fixing whitespace issues in LAMMPS')
+    parser.add_argument('-c', '--config', metavar='CONFIG_FILE', help='location of a optional configuration file')
+    parser.add_argument('-f', '--fix', action='store_true', help='automatically fix common issues')
+    parser.add_argument('-v', '--verbose', action='store_true', help='verbose output')
+    parser.add_argument('DIRECTORY', help='directory that should be checked')
+    args = parser.parse_args()
+
+    if args.config:
+        with open(args.config, 'r') as cfile:
+            config = yaml.load(cfile, Loader=yaml.FullLoader)
+    else:
+        config = yaml.load(DEFAULT_CONFIG, Loader=yaml.FullLoader)
+
+    check_folder(args.DIRECTORY, config, args.fix, args.verbose)
+
+if __name__ == "__main__":
+    main()