aboutsummaryrefslogtreecommitdiffstats
path: root/main/ca-certificates/certdata2pem.py
diff options
context:
space:
mode:
Diffstat (limited to 'main/ca-certificates/certdata2pem.py')
-rw-r--r--main/ca-certificates/certdata2pem.py155
1 files changed, 0 insertions, 155 deletions
diff --git a/main/ca-certificates/certdata2pem.py b/main/ca-certificates/certdata2pem.py
deleted file mode 100644
index f91422bf18..0000000000
--- a/main/ca-certificates/certdata2pem.py
+++ /dev/null
@@ -1,155 +0,0 @@
-#!/usr/bin/python
-# vim:set et sw=4:
-#
-# certdata2pem.py - splits certdata.txt into multiple files
-#
-# Copyright (C) 2009 Philipp Kern <pkern@debian.org>
-#
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation; either version 2 of the License, or
-# (at your option) any later version.
-#
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-# GNU General Public License for more details.
-#
-# You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301,
-# USA.
-
-import base64
-import os.path
-import re
-import sys
-import textwrap
-import io
-
-objects = []
-
-# Dirty file parser.
-in_data, in_multiline, in_obj = False, False, False
-field, type, value, obj = None, None, None, dict()
-
-# Python 3 will not let us decode non-ascii characters if we
-# have not specified an encoding, but Python 2's open does not
-# have an option to set the encoding. Python 3's open is io.open
-# and io.open has been backported to Python 2.6 and 2.7, so use io.open.
-for line in io.open('certdata.txt', 'rt', encoding='utf8'):
- # Ignore the file header.
- if not in_data:
- if line.startswith('BEGINDATA'):
- in_data = True
- continue
- # Ignore comment lines.
- if line.startswith('#'):
- continue
- # Empty lines are significant if we are inside an object.
- if in_obj and len(line.strip()) == 0:
- objects.append(obj)
- obj = dict()
- in_obj = False
- continue
- if len(line.strip()) == 0:
- continue
- if in_multiline:
- if not line.startswith('END'):
- if type == 'MULTILINE_OCTAL':
- line = line.strip()
- for i in re.finditer(r'\\([0-3][0-7][0-7])', line):
- value.append(int(i.group(1), 8))
- else:
- value += line
- continue
- obj[field] = value
- in_multiline = False
- continue
- if line.startswith('CKA_CLASS'):
- in_obj = True
- line_parts = line.strip().split(' ', 2)
- if len(line_parts) > 2:
- field, type = line_parts[0:2]
- value = ' '.join(line_parts[2:])
- elif len(line_parts) == 2:
- field, type = line_parts
- value = None
- else:
- raise NotImplementedError('line_parts < 2 not supported.')
- if type == 'MULTILINE_OCTAL':
- in_multiline = True
- value = bytearray()
- continue
- obj[field] = value
-if len(obj) > 0:
- objects.append(obj)
-
-# Read blacklist.
-blacklist = []
-if os.path.exists('blacklist.txt'):
- for line in open('blacklist.txt', 'r'):
- line = line.strip()
- if line.startswith('#') or len(line) == 0:
- continue
- item = line.split('#', 1)[0].strip()
- blacklist.append(item)
-
-# Build up trust database.
-trust = dict()
-for obj in objects:
- if obj['CKA_CLASS'] != 'CKO_NSS_TRUST':
- continue
- if obj['CKA_LABEL'] in blacklist:
- print("Certificate %s blacklisted, ignoring." % obj['CKA_LABEL'])
- elif obj['CKA_TRUST_SERVER_AUTH'] == 'CKT_NSS_TRUSTED_DELEGATOR':
- trust[obj['CKA_LABEL']] = True
- elif obj['CKA_TRUST_EMAIL_PROTECTION'] == 'CKT_NSS_TRUSTED_DELEGATOR':
- trust[obj['CKA_LABEL']] = True
- elif obj['CKA_TRUST_SERVER_AUTH'] == 'CKT_NSS_NOT_TRUSTED':
- print('!'*74)
- print("UNTRUSTED BUT NOT BLACKLISTED CERTIFICATE FOUND: %s" % obj['CKA_LABEL'])
- print('!'*74)
- else:
- print("Ignoring certificate %s. SAUTH=%s, EPROT=%s" % \
- (obj['CKA_LABEL'], obj['CKA_TRUST_SERVER_AUTH'],
- obj['CKA_TRUST_EMAIL_PROTECTION']))
-
-for obj in objects:
- if obj['CKA_CLASS'] == 'CKO_CERTIFICATE':
- if not obj['CKA_LABEL'] in trust or not trust[obj['CKA_LABEL']]:
- continue
- bname = obj['CKA_LABEL'][1:-1].replace('/', '_')\
- .replace(' ', '_')\
- .replace('(', '=')\
- .replace(')', '=')\
- .replace(',', '_')
-
- # this is the only way to decode the way NSS stores multi-byte UTF-8
- # and we need an escaped string for checking existence of things
- # otherwise we're dependant on the user's current locale.
- if bytes != str:
- # We're in python 3, convert the utf-8 string to a
- # sequence of bytes that represents this utf-8 string
- # then encode the byte-sequence as an escaped string that
- # can be passed to open() and os.path.exists()
- bname = bname.encode('utf-8').decode('unicode_escape').encode('latin-1')
- else:
- # Python 2
- # Convert the unicode string back to its original byte form
- # (contents of files returned by io.open are returned as
- # unicode strings)
- # then to an escaped string that can be passed to open()
- # and os.path.exists()
- bname = bname.encode('utf-8').decode('string_escape')
-
- fname = bname + b'.crt'
- if os.path.exists(fname):
- print("Found duplicate certificate name %s, renaming." % bname)
- fname = bname + b'_2.crt'
- f = open(fname, 'w')
- f.write("-----BEGIN CERTIFICATE-----\n")
- encoded = base64.b64encode(obj['CKA_VALUE']).decode('utf-8')
- f.write("\n".join(textwrap.wrap(encoded, 64)))
- f.write("\n-----END CERTIFICATE-----\n")
-