56 lines
1.9 KiB
Python
56 lines
1.9 KiB
Python
import os
|
|
|
|
extensions = ('.uvue', '.uts', '.vue', '.json', '.js', '.ts', '.scss', '.md', '.txt', '.ps1', '.bat', '.sh')
|
|
root_dir = r'd:\骅锋\mall'
|
|
|
|
non_utf8_files = []
|
|
|
|
for root, dirs, files in os.walk(root_dir):
|
|
# Skip unpackage and node_modules for performance
|
|
if 'unpackage' in root or 'node_modules' in root or '.git' in root:
|
|
continue
|
|
|
|
for file in files:
|
|
if file.lower().endswith(extensions):
|
|
file_path = os.path.join(root, file)
|
|
try:
|
|
with open(file_path, 'rb') as f:
|
|
content = f.read()
|
|
|
|
# Try to decode as UTF-8
|
|
try:
|
|
content.decode('utf-8')
|
|
# If it succeeds, it IS UTF-8
|
|
continue
|
|
except UnicodeDecodeError:
|
|
# Not UTF-8
|
|
pass
|
|
|
|
# Try to detect encoding
|
|
encoding = "Unknown/Other"
|
|
try:
|
|
content.decode('gbk')
|
|
encoding = "GBK/GB2312"
|
|
except UnicodeDecodeError:
|
|
try:
|
|
content.decode('utf-16')
|
|
encoding = "UTF-16"
|
|
except UnicodeDecodeError:
|
|
try:
|
|
content.decode('latin-1')
|
|
encoding = "Latin-1/Windows-1252"
|
|
except UnicodeDecodeError:
|
|
pass
|
|
|
|
non_utf8_files.append((file_path, encoding))
|
|
except Exception as e:
|
|
# Permission denied or other error
|
|
continue
|
|
|
|
print(f"{'Detected Encoding':<25} | {'File Path'}")
|
|
print("-" * 100)
|
|
for file_path, encoding in non_utf8_files:
|
|
print(f"{encoding:<25} | {file_path}")
|
|
|
|
print(f"\nFound {len(non_utf8_files)} non-UTF-8 files.")
|