IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Apart from the leak fix, this is faster and stricter, not accepting
SID string buffers with trailing garbage ("S-1-2-3qwerty" would have
been accepted, but not now).
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
These are just as readable with `less` as they were with `zless`.
This file has been slightly manually edited to add line-breaks. There
is not an easy setting in Python's json module to get good formatting.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Nov 27 02:10:12 UTC 2023 on atb-devel-224
We had two sets of test vectors (Windows ground-truth for SDDL
compilation) that got mixed up.
The "oversized ACLs" set is ACLs that contain repeated ACEs, like
"D:P(D;;;;;MP)(D;;;;;MP)" -- Windows will assign a size to the ACL
that is greater than the sum of the ACEs, while Samba will not (in
part because we don't actually store a size for the ACL, instead
calculating it on the fly from the size of the ACEs).
The "TX integers" set is for resource attribute ACEs with octet-string
data that contains pure integers (lacking '#' characters) in their
SDDL, like «(RA;;;;;WD;("bar",TX,0x0,0077,00,0077,00))». We used to
think that was weird, and that RA-TX ACEs should contain octet-strings
in the conditional ACE style. But now we have realised it's not weird,
it's normal, and we have fixed our handling of these ACEs.
As a result of this mix-up, some of the tests labelled as "oversized
ACLs" started passing when we fixed the TX integer problem, and that
was confusing. All of the removed tests are already on the TX integer
set -- the removed ones were duplicates.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This is unneeded, as now all the checks are done in the relevant
parse_* functions.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We were reusing parse_literal() because it almost does what we need,
but it is different enough that check_resource_attr_type() is large
and complicated, and can't handle all the cases (in particular octet-
strings and SIDs are different in resource ACEs).
This way is better because we know the type in advance, so we can use
that to choose the parser, which will help with octet-strings that are
only digits.
In this commit we're leaving the check there, but it soon won't do
anything that the parse_* functions don't, and we will remove it.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We are going to parse octet strings like Windows (as opposed to like
Windows docs), so the tests need changing.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
the message offset is largely calculated using the differences
between pointers in many places scattered throughout the code.
If we got one of these wrong, we could easily have a SIZE_MAX-ish
offset, which would be unfortunate if we came decided to display
the offset using spaces.
We can sanely limit the offset to the length of the SDDL.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
the way we parse things, we can't really distinguish between complete
nonsense and an ACL that seems to end early because of bad flags. That
is, "D:ZZ(A;;;;;WD)" looks the same as "ZZ" to the parser. But at least
we can point to the right place in the string.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The wrong number of semicolons is usually one less than count (which
counts sections separated by semicolons), except when count is zero.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This allows the messages to be more reliably presented by client tools
in a useful way.
The messages lose the trailing \n, and some were slightly tweaked (e.g.
s/Resource ACE/Resource Attribute ACE/).
They will still show up in logs for callers of sddl_decode(), but at
NOTICE level rather than WARNING.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Having it optionally NULL just complicates the code, and Coverity
rightly complained.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
These have been unsupported since commit
3b6c1f1a9c47d8d76a7cd946468c1c42e4fb097a.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Usually the conditions are embedded in part of some SDDL, and the
offset from the beginning of the condtions is a bit useless and
confusing. Callers of sddl_decode_err_msg get the offset from the
beginning of the SDDL which is a different and more useful number.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
As it stands, ace_conditions_compile_sddl() won't produce a message when
it succeeds (i.e. return non-NULL), so this debug is just clutter.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This will return an error message, if it can, along with an indicative
position.
For conditional ACEs the message might be accurate, and the position
fine-grained. For example, you might be able to construct the message
like this:
D:(XA;;CC;;;S-1-2-3;(@User.Title == !(@User.Title)))
^
16: unexpected operator
For non-conditional ACEs, the position typically points to the beginning
of the ACE, like this:
D:(D;OICI;GA;;;BG)(D;OICI;GA;;;AN)(A; OICI; GRGWGX;;;AU)
^
unknown error
Here the error is in the spaces either side of " OICI; ", but the pointer
points to the beginning of the ACE.
The old sddl_decode() function becomes a wrapper around the new function,
which inherits the guts of the old function.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
It is simpler for the message to have consistent parentage; it
is easier to drop one message we'll never see than to talloc it.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This is what Windows does, and it removes a couple of knownfails.
We can change it here cheaply without affecting the core dom_sid code,
which is good because there seem to be other places where we need the
uppercase S (for example in ldap search <SID=> queries).
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The LIBNDR_FLAG_ namespace is getting dangerously full, with only a
single flag value (1 << 9) remaining for use. After that flag is put
into use, we won’t be able to add any new flags without increasing the
flag width to 64‐bit.
Up to now we’ve used a haphazard mix of int, unsigned, and uint32_t to
store these flags. Introduce a new type, ‘libndr_flags’, to be used
consistently to hold LIBNDR flags. If in the future we find we need to
move to 64‐bit flags, this type gives us an opportunity to do that.
Bump the NDR version to 4.0.0 — an major version increment, for we’re
changing the function ABI and adding the new symbol
ndr_print_libndr_flags.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
