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.
Break out arguments to proto_register_protocol() so they can be renamed
easily. Unfortunately the "pidl_" prefix will have to stick around if/when
the old dissectors are removed because the filter and short names clash.
Add hf for function return value.
Rename struct field members to be named after the structure name
and element name rather than element name and element type.
Remember which structure or function we are processing and use it
to generate hf fields according to the naming scheme above.
I'm hoping this will allow better mapping hf fields to the structures
they are present in which isn't possible at the moment. (Line mode
allows us to use /foo/ .. /bar/ regexps and to store state during the
processing.
the backend should check for
(dce_call->state_flags & DCESRV_CALL_STATE_FLAG_MAY_ASYNC)
then it's allowed to reply async
then the backend should mark that call as async with
dce_call->state_flags |= DCESRV_CALL_STATE_FLAG_ASYNC;
later it has to manualy set r->out.result
and then send the reply by calling
status = dcesrv_reply(p->dce_call);
NOTE: that ncacn_np doesn't support async replies yet
- implement an async version of echo_TestSleep
- reenable the echo_TestSleep torture test
(this need to be more strict when we have support for async ncacn_np)
metze
This call uses a new IDL type, NTTIME_hyper. This is 8-byte aligned,
as the name suggests.
Expand the QuerySecret LSA calls in RPC-SAMLOGON and RPC-LSA, to
validate the behaviour of times, and of the old secrets.
Thanks to tridge for spotting the use of HYPER!
Andrew Bartlett
this stage does the following:
- simplifies the dcerpc_handle handling, and all the callers of it
- split out the context_id depenent state into a linked list of established contexts
- fixed some talloc handling in several rpc servers that i noticed while doing the above
dcerpc_alter_context and multiple context_ids in the dcerpc client
library.
This stage does the following:
- split "struct dcerpc_pipe" into two parts, the main part being "struct dcerpc_connection", which
contains all the parts not dependent on the context, and "struct dcerpc_pipe" which has
the context dependent part. This is similar to the layering in libcli_*() for SMB
- disable the current dcerpc_alter code. I've used a #warning until i
get the 2nd phase finished. I don't know how portable #warning is, but
it won't be long before I add full alter context support anyway, so it won't last long
- cleanup the allocation of dcerpc_pipe structures. The previous code
was quite awkward.
'declare bitmap foo1;'
'declare enum foo2;'
and also allow
typedef [public] bitmap ...
typedef [public] enum ...
you need to a forward declaration of bitmaps and enums
when you want to use them in another idl file,
and you need to make the real declaration to be public
see the next commit to samr.idl and netlogon.idl
metze
is less efficient, but I really doubt that matters.
- use enum in epmapper.idl for protocol type
- added support for "enum8bit" flag, used in epmapper.idl
based on the idea of manipulating the .c and .h files generated by
parser.pm with perl regexps and glueing it all together to make an
ethereal plugin.
I thought this was a pretty crazy idea to start off with but it has
turned out to be not as complicated as I thought and has the huge advantage
of not duplicating any of the difficult code in parser.pm.
main rpc server code. let the backends specify a ndr_push/ndr_pull function
like we already do with the dispatch() function.
this allows an interface implmentation to work as real proxy
without needing to know the idl for an interface that means
just the plain decrypted payload can be forwarded
If someone want to write such a backend, patches are wellcome
metze
hasn't given us a null pointer, otherwise we can segv when we dereference it. For example:
[size_is(*size)] *x;
uint32 *size;
if a broken server gave us x != NULL and size == NULL then we would crash. I've added
a check_null_pointer() call in pidl to catch this.
that depend on variables that come after the array in the structure or function.
This has been something that has been problematic for a while, but the
winreg QueryValue problem finally prompted me to fix it properly. We
should now go back and fix up all the ugly workarounds we have used to
avoid this problem in other calls.
Unfortunately the solution is fairly complex, and involves the use of
the internal ndr token lists (similar to the solution for relative
pointers). I wonder if anyone else will be able to follow the logic if
I get run over by a bus :-)
