Use python-magic, filemagic's official python binding lib

This commit is contained in:
Daniel Martí 2013-10-16 22:50:07 +02:00
parent bd59b2f77d
commit fb5a2209fc
2 changed files with 40 additions and 37 deletions

View file

@ -139,7 +139,7 @@ Paramiko (debian package python-paramiko)
@item @item
Imaging (debian package python-imaging) Imaging (debian package python-imaging)
@item @item
Filemagic (pip install filemagic) Magic (debian package python-magic)
@end itemize @end itemize

View file

@ -23,6 +23,7 @@ import subprocess
import time import time
import operator import operator
import cgi import cgi
import magic
def getvcs(vcstype, remote, local, sdk_path): def getvcs(vcstype, remote, local, sdk_path):
if vcstype == 'git': if vcstype == 'git':
@ -1588,48 +1589,50 @@ def scan_source(build_dir, root_dir, thisbuild):
ignore = [p.strip() for p in thisbuild['scanignore'].split(';')] ignore = [p.strip() for p in thisbuild['scanignore'].split(';')]
else: else:
ignore = [] ignore = []
ms = magic.open(magic.MIME_TYPE)
ms.load()
# Iterate through all files in the source code... # Iterate through all files in the source code...
import magic for r,d,f in os.walk(build_dir):
with magic.Magic(flags=magic.MAGIC_MIME_TYPE) as m: for curfile in f:
for r,d,f in os.walk(build_dir):
for curfile in f:
if '/.hg' in r or '/.git' in r or '/.svn' in r: if '/.hg' in r or '/.git' in r or '/.svn' in r:
continue continue
# Path (relative) to the file... # Path (relative) to the file...
fp = os.path.join(r, curfile) fp = os.path.join(r, curfile)
fd = fp[len(build_dir)+1:] fd = fp[len(build_dir)+1:]
# Check if this file has been explicitly excluded from scanning... # Check if this file has been explicitly excluded from scanning...
ignorethis = False ignorethis = False
for i in ignore: for i in ignore:
if fd.startswith(i): if fd.startswith(i):
ignorethis = True ignorethis = True
break
if ignorethis:
continue
for suspect in usual_suspects:
if suspect in curfile.lower():
problems.append('Found probable non-free blob ' + fp)
mime = ms.file(fp)
if mime == 'application/x-sharedlib':
problems.append('Found shared library at %s' % fd)
elif mime == 'application/x-archive':
problems.append('Found static library at %s' % fd)
elif mime == 'application/x-executable':
problems.append('Found binary executable at %s' % fd)
elif fp.endswith('.apk'):
problems.append('Found apk archive at %s' % fd)
elif curfile.endswith('.java'):
for line in file(fp):
if 'DexClassLoader' in line:
problems.append('Found DexClassLoader in ' + fp)
break break
if ignorethis: ms.close()
continue
for suspect in usual_suspects:
if suspect in curfile.lower():
problems.append('Found probable non-free blob ' + fp)
mime = m.id_filename(fp)
if mime == 'application/x-sharedlib':
problems.append('Found shared library at %s' % fd)
elif mime == 'application/x-archive':
problems.append('Found static library at %s' % fd)
elif mime == 'application/x-executable':
problems.append('Found binary executable at %s' % fd)
elif fp.endswith('.apk'):
problems.append('Found apk archive at %s' % fd)
elif curfile.endswith('.java'):
for line in file(fp):
if 'DexClassLoader' in line:
problems.append('Found DexClassLoader in ' + fp)
break
# Presence of a jni directory without buildjni=yes might # Presence of a jni directory without buildjni=yes might
# indicate a problem... (if it's not a problem, explicitly use # indicate a problem... (if it's not a problem, explicitly use