From f83e754ae61046fd441cbaf09a825378c3ab2ff2 Mon Sep 17 00:00:00 2001 From: Sreepathi Pai Date: Mon, 5 Jan 2009 08:49:53 +1100 Subject: Create schema.ldif at runtime directly from ad-schema files Here's a first attempt at moving the minschema_wspp code into a library as Andrew requested. Since this script no longer has to generate CN=aggregate, I've simplified it quite a bit to a level where it almost does a line-by-line translation. This is faster and simpler, but it may not catch as many errors in the ad-schema files as the previous versions did. Signed-off-by: Andrew Bartlett --- source4/scripting/python/samba/ms_schema.py | 253 ++++++++++++++++++++++++++++ source4/scripting/python/samba/provision.py | 30 +++- 2 files changed, 278 insertions(+), 5 deletions(-) create mode 100644 source4/scripting/python/samba/ms_schema.py (limited to 'source4/scripting') diff --git a/source4/scripting/python/samba/ms_schema.py b/source4/scripting/python/samba/ms_schema.py new file mode 100644 index 0000000000..636cc3e34b --- /dev/null +++ b/source4/scripting/python/samba/ms_schema.py @@ -0,0 +1,253 @@ +# +# create schema.ldif (as a string) from WSPP documentation +# +# based on minschema.py and minschema_wspp +# + +import re +import base64 + +bitFields = {} + +# ADTS: 2.2.9 +# bit positions as labeled in the docs +bitFields["searchflags"] = { + 'fATTINDEX': 31, # IX + 'fPDNTATTINDEX': 30, # PI + 'fANR': 29, #AR + 'fPRESERVEONDELETE': 28, # PR + 'fCOPY': 27, # CP + 'fTUPLEINDEX': 26, # TP + 'fSUBTREEATTINDEX': 25, # ST + 'fCONFIDENTIAL': 24, # CF + 'fNEVERVALUEAUDIT': 23, # NV + 'fRODCAttribute': 22, # RO + + + # missing in ADTS but required by LDIF + 'fRODCFilteredAttribute': 22, # RO ? + 'fCONFIDENTAIL': 24, # typo + 'fRODCFILTEREDATTRIBUTE': 22 # case + } + +# ADTS: 2.2.10 +bitFields["systemflags"] = { + 'FLAG_ATTR_NOT_REPLICATED': 31, 'FLAG_CR_NTDS_NC': 31, # NR + 'FLAG_ATTR_REQ_PARTIAL_SET_MEMBER': 30, 'FLAG_CR_NTDS_DOMAIN': 30, # PS + 'FLAG_ATTR_IS_CONSTRUCTED': 29, 'FLAG_CR_NTDS_NOT_GC_REPLICATED': 29, # CS + 'FLAG_ATTR_IS_OPERATIONAL': 28, # OP + 'FLAG_SCHEMA_BASE_OBJECT': 27, # BS + 'FLAG_ATTR_IS_RDN': 26, # RD + 'FLAG_DISALLOW_MOVE_ON_DELETE': 6, # DE + 'FLAG_DOMAIN_DISALLOW_MOVE': 5, # DM + 'FLAG_DOMAIN_DISALLOW_RENAME': 4, # DR + 'FLAG_CONFIG_ALLOW_LIMITED_MOVE': 3, # AL + 'FLAG_CONFIG_ALLOW_MOVE': 2, # AM + 'FLAG_CONFIG_ALLOW_RENAME': 1, # AR + 'FLAG_DISALLOW_DELETE': 0 # DD + } + +# ADTS: 2.2.11 +bitFields["schemaflagsex"] = { + 'FLAG_ATTR_IS_CRITICAL': 31 + } + +# ADTS: 3.1.1.2.2.2 +oMObjectClassBER = { + '1.3.12.2.1011.28.0.702' : base64.b64encode('\x2B\x0C\x02\x87\x73\x1C\x00\x85\x3E'), + '1.2.840.113556.1.1.1.12': base64.b64encode('\x2A\x86\x48\x86\xF7\x14\x01\x01\x01\x0C'), + '2.6.6.1.2.5.11.29' : base64.b64encode('\x56\x06\x01\x02\x05\x0B\x1D'), + '1.2.840.113556.1.1.1.11': base64.b64encode('\x2A\x86\x48\x86\xF7\x14\x01\x01\x01\x0B'), + '1.3.12.2.1011.28.0.714' : base64.b64encode('\x2B\x0C\x02\x87\x73\x1C\x00\x85\x4A'), + '1.3.12.2.1011.28.0.732' : base64.b64encode('\x2B\x0C\x02\x87\x73\x1C\x00\x85\x5C'), + '1.2.840.113556.1.1.1.6' : base64.b64encode('\x2A\x86\x48\x86\xF7\x14\x01\x01\x01\x06') +} + +# separated by commas in docs, and must be broken up +multivalued_attrs = set(["auxiliaryclass","maycontain","mustcontain","posssuperiors", + "systemauxiliaryclass","systemmaycontain","systemmustcontain", + "systemposssuperiors"]) + +def __read_folded_line(f, buffer): + """ reads a line from an LDIF file, unfolding it""" + line = buffer + + while True: + l = f.readline() + + if l[:1] == " ": + # continued line + + # cannot fold an empty line + assert(line != "" and line != "\n") + + # preserves '\n ' + line = line + l + else: + # non-continued line + if line == "": + line = l + + if l == "": + # eof, definitely won't be folded + break + else: + # marks end of a folded line + # line contains the now unfolded line + # buffer contains the start of the next possibly folded line + buffer = l + break + + return (line, buffer) + + +def __read_raw_entries(f): + """reads an LDIF entry, only unfolding lines""" + + # will not match options after the attribute type + attr_type_re = re.compile("^([A-Za-z]+[A-Za-z0-9-]*):") + + buffer = "" + + while True: + entry = [] + + while True: + (l, buffer) = __read_folded_line(f, buffer) + + if l[:1] == "#": + continue + + if l == "\n" or l == "": + break + + m = attr_type_re.match(l) + + if m: + if l[-1:] == "\n": + l = l[:-1] + + entry.append(l) + else: + print >>sys.stderr, "Invalid line: %s" % l, + sys.exit(1) + + if len(entry): + yield entry + + if l == "": + break + + +def fix_dn(dn): + """fix a string DN to use ${SCHEMADN}""" + + # folding? + if dn.find("") != -1: + dn = dn.replace("\n ", "") + dn = dn.replace(" ", "") + return dn.replace("CN=Schema,CN=Configuration,", "${SCHEMADN}") + else: + return dn + +def __convert_bitfield(key, value): + """Evaluate the OR expression in 'value'""" + assert(isinstance(value, str)) + + value = value.replace("\n ", "") + value = value.replace(" ", "") + + try: + # some attributes already have numeric values + o = int(value) + except ValueError: + o = 0 + flags = value.split("|") + for f in flags: + bitpos = bitFields[key][f] + o = o | (1 << (31 - bitpos)) + + return str(o) + +def __write_ldif_one(entry): + """Write out entry as LDIF""" + out = [] + + for l in entry: + if isinstance(l[1], str): + vl = [l[1]] + else: + vl = l[1] + + if l[0].lower() == 'omobjectclass': + out.append("%s:: %s" % (l[0], l[1])) + continue + + for v in vl: + out.append("%s: %s" % (l[0], v)) + + + return "\n".join(out) + +def __transform_entry(entry, objectClass): + """Perform transformations required to convert the LDIF-like schema + file entries to LDIF, including Samba-specific stuff.""" + + entry = [l.split(":", 1) for l in entry] + + cn = "" + + for l in entry: + key = l[0].lower() + l[1] = l[1].lstrip() + + if not cn and key == "cn": + cn = l[1] + + if key in multivalued_attrs: + # unlike LDIF, these are comma-separated + l[1] = l[1].replace("\n ", "") + l[1] = l[1].replace(" ", "") + + l[1] = l[1].split(",") + + if key in bitFields: + l[1] = __convert_bitfield(key, l[1]) + + if key == "omobjectclass": + l[1] = oMObjectClassBER[l[1].strip()] + + if isinstance(l[1], str): + l[1] = fix_dn(l[1]) + + + assert(cn) + entry.insert(0, ["dn", "CN=%s,${SCHEMADN}" % cn]) + entry.insert(1, ["objectClass", ["top", objectClass]]) + + return entry + +def __parse_schema_file(filename, objectClass): + """Load and transform a schema file.""" + + out = [] + + f = open(filename, "rU") + for entry in __read_raw_entries(f): + out.append(__write_ldif_one(__transform_entry(entry, objectClass))) + + return "\n\n".join(out) + + +def read_ms_schema(attr_file, classes_file, dump_attributes = True, dump_classes = True, debug = False): + """Read WSPP documentation-derived schema files.""" + + attr_ldif = "" + classes_ldif = "" + + if dump_attributes: + attr_ldif = __parse_schema_file(attr_file, "attributeSchema") + if dump_classes: + classes_ldif = __parse_schema_file(classes_file, "classSchema") + + return attr_ldif + "\n\n" + classes_ldif + "\n\n" diff --git a/source4/scripting/python/samba/provision.py b/source4/scripting/python/samba/provision.py index c817bffbdd..4da8079e09 100644 --- a/source4/scripting/python/samba/provision.py +++ b/source4/scripting/python/samba/provision.py @@ -44,6 +44,7 @@ from samba.dcerpc import security import urllib from ldb import SCOPE_SUBTREE, SCOPE_ONELEVEL, SCOPE_BASE, LdbError, \ timestring, CHANGETYPE_MODIFY, CHANGETYPE_NONE +from ms_schema import read_ms_schema __docformat__ = "restructuredText" @@ -854,9 +855,10 @@ def setup_samdb(path, setup_path, session_info, credentials, lp, message("Setting up sam.ldb Samba4 schema") setup_add_ldif(samdb, setup_path("schema_samba4.ldif"), {"SCHEMADN": names.schemadn }) + message("Setting up sam.ldb AD schema") - setup_add_ldif(samdb, setup_path("schema.ldif"), - {"SCHEMADN": names.schemadn}) + data = get_schema_data(setup_path, {"SCHEMADN": names.schemadn}) + samdb.add_ldif(data) setup_add_ldif(samdb, setup_path("aggregate_schema.ldif"), {"SCHEMADN": names.schemadn}) @@ -1268,8 +1270,9 @@ def provision_backend(setup_dir=None, message=None, setup_add_ldif(schemadb, setup_path("schema_samba4.ldif"), {"SCHEMADN": names.schemadn }) - setup_add_ldif(schemadb, setup_path("schema.ldif"), - {"SCHEMADN": names.schemadn}) + + data = get_schema_data(setup_path, {"SCHEMADN": names.schemadn}) + schemadb.add_ldif(data) if ldap_backend_type == "fedora-ds": if ldap_backend_port is not None: @@ -1659,7 +1662,7 @@ def load_schema(setup_path, samdb, schemadn, netbiosname, configdn, sitename, :param serverdn: DN of the server :param servername: Host name of the server """ - schema_data = open(setup_path("schema.ldif"), 'r').read() + schema_data = get_schema_data(setup_path, {"SCHEMADN": schemadn}) schema_data += open(setup_path("schema_samba4.ldif"), 'r').read() schema_data = substitute_var(schema_data, {"SCHEMADN": schemadn}) check_all_substituted(schema_data) @@ -1679,3 +1682,20 @@ def load_schema(setup_path, samdb, schemadn, netbiosname, configdn, sitename, check_all_substituted(head_data) samdb.attach_schema_from_ldif(head_data, schema_data) + +def get_schema_data(setup_path, subst_vars = None): + """Get schema data from the AD schema files instead of schema.ldif. + + :param setup_path: Setup path function. + :param subst_vars: Optional variables to substitute in the file. + """ + + # this data used to be read from schema.ldif + + data = read_ms_schema(setup_path('ad-schema/MS-AD_Schema_Attributes_v20080618.txt'), + setup_path('ad-schema/MS-AD_Schema_Classes_v20080618.txt')) + + if subst_vars is not None: + data = substitute_var(data, subst_vars) + check_all_substituted(data) + return data -- cgit