Cross session activation. Session moniker ignored - c++

There are two accounts on my windows pc: \A1 and \SYSTEM. My test project consists of three applications:
Server.exe - out-of-proc com server. implements object Q
Service.exe - out-of proc com service. implements object T
Client.exe - a client applications
My workflow is shown below:
I logged in the \A1. I launch 'Client.exe'.
Client creates an instance of the object 'T'
Client calls some method, say 'T.ActivateQ', which instantiates 'Q'.
My intention is to make Service.exe to create an instance of 'Q' in the account \A1, even though Service.exe resides under the \SYSTEM account.
As stated in this article, Windows has built in functionality to do this kind of activation called 'Session-to-Session activation'. It is implemented by means of Session Moniker ("Session:!clsid:").
Given that, I've configured 'Server.exe' to run as 'Interactive user' via dcomcnfg.
Then I've implemented the function 'CoGetClassObjectInSession':
void CoGetClassObjectInSession(DWORD sessionId, Guid const& clsid, void *pvReserved, Guid const& riid, void **ppv)
{
ATL::CComQIPtr<IBindCtx> bindCtxInst;
COMCHK(CreateBindCtx(NULL, &bindCtxInst));
CComQIPtr<IMoniker> classMonikerInst;
COMCHK(::CreateClassMoniker(clsid, &classMonikerInst));
std::wstringstream ss;
ss << L"Session:" << std::dec << sessionId;
ULONG parsed;
ATL::CComQIPtr<IMoniker> sessionMonikerInst;
COMCHK(::MkParseDisplayNameEx(bindCtxInst, ss.str().c_str(), &parsed, &sessionMonikerInst));
ATL::CComQIPtr<IMoniker> classObjectMoniker;
COMCHK(sessionMonikerInst->ComposeWith(classMonikerInst, FALSE, &classObjectMoniker));
ATL::CComPtr<IClassFactory> sessionFactoryInst;
COMCHK(classObjectMoniker->BindToObject(bindCtxInst, NULL, riid, ppv));
}
and I use it in the service:
HRESULT CMyActivator::Activate()
{
try
{
// Impersonate the client
CComQIPtr<IServerSecurity> ss;
COMCHK(::CoGetCallContext(__uuidof(IServerSecurity), reinterpret_cast<void**>(&ss)));
callctx_impersonation_handle _handle(ss);
ImpersonationScope _imp_scope(_handle);
CComQIPtr<IClassFactory> ppv;
CoGetClassObjectInSession(/*session id for \\A1*/ 1, Guid(__uuidof(Q)), nullptr, Guid(__uuidof(IClassFactory)), reinterpret_cast<void**>(&ppv));
ENSURE(NULL != ppv, "'ppv' should not be null");
return S_OK;
}
catch(std::exception const& ex)
{
return E_FAIL;
}
catch(...)
{
return E_FAIL;
}
}
But nothing happens. I mean the service always launches the server in the session 0 which corresponds to the \SYSTEM account.
Client's security settings:
COMCHK(::CoSetProxyBlanket(
activatorQ
, RPC_C_AUTHN_DEFAULT
, RPC_C_AUTHZ_DEFAULT
, NULL
, RPC_C_AUTHN_LEVEL_DEFAULT
, RPC_C_IMP_LEVEL_DELEGATE
, NULL
, EOAC_DYNAMIC_CLOAKING));
Server.rgs file:
HKLM
{
NoRemove Software
{
NoRemove Classes
{
NoRemove AppID
{
ForceRemove '%APPID%' = s 'Test Server'
{
val RunAs = s 'Interactive User'
}
}
}
}
}
Am I doing something wrong?
UPDATE
I've finally found a solution. The registration script for the class 'Q' was missing the AppID field:
HKCR
{
NoRemove CLSID
{
ForceRemove {6FDE856C-F1B2-4466-8435-20F4AEF2C2E1} = s 'SMonClass Class'
{
ForceRemove Programmable
LocalServer32 = s '%MODULE%'
{
val ServerExecutable = s '%MODULE_RAW%'
}
TypeLib = s '{4FA39B0F-5B24-43C6-A5B1-11D8F34277B7}'
Version = s '1.0'
--> val AppID = s '%APPID%'** <-- This line
}
}
}

Related

org.omg.CORBA.MARSHAL: Server-side Exception: null

I'm trying to register to a CORBA CosNotification Service. In the documentation of the service I'm trying to connect to, it says I have to a CosNotifyComm::SequencePushConsumer object, and attach it to the notification service. I've included my code, and the error I'm getting back.
AlarmClient.java
import NotificationIRPSystem.*;
import org.omg.CosNotification.*;
import org.omg.CosNotifyComm.*;
import org.omg.CosNotifyChannelAdmin.*;
import org.omg.CosNotifyFilter.*;
import org.omg.CosNaming.*;
import org.omg.CosNaming.NamingContextPackage.*;
import org.omg.PortableServer.*;
public class AlarmClient
{
static _NotificationIRPOperations notiOp;
public static void main (String args [])
{
try{
org.omg.CORBA.ORB orb = org.omg.CORBA.ORB.init (args, null);
org.omg.CORBA.Object objRef =
orb.resolve_initial_references ("NameService");
System.out.println("IOR===> " + objRef);
NamingContextExt nc = NamingContextExtHelper.narrow(objRef);
String name = "com/ericsson/nms/cif/service/NMSNAConsumer";
String portal = "com/ericsson/nms/cif/service/NMSNAPortal";
org.omg.CORBA.Object notiObj = nc.resolve_str(name);
System.out.println(nc.resolve_str(portal));
System.out.println("noti---->" + notiObj);
_NotificationIRPOperations tt = _NotificationIRPOperationsHelper.narrow(notiObj);
IRPManager irpMan = new IRPManager();
POA poa = POAHelper.narrow(orb.resolve_initial_references("RootPOA"));
poa.the_POAManager().activate();
org.omg.CORBA.Object objNotiServer = poa.servant_to_reference(irpMan);
String manager_reference = orb.object_to_string(objNotiServer);
System.out.println("MANAGER REF: " + manager_reference);
System.out.println("OBJ NOTI SERVER: " + objNotiServer);
int time_tick = 15;
String filter = "";
String[] asd = {};
NotificationIRPConstDefs.EventTypesSetHolder e_list = new NotificationIRPConstDefs.EventTypesSetHolder();
NotificationIRPConstDefs.ExtendedEventTypesSetHolder ex_list = new NotificationIRPConstDefs.ExtendedEventTypesSet\
Holder();
String cats[] = tt.get_notification_categories(e_list, ex_list);
String res = tt.attach_push(manager_reference, time_tick, asd , filter);
// System.out.println("SUCCESS--->" + res);
// String[] ids = tt.get_subscription_ids(manager_reference);
System.out.println("ALARMIRPOPERATIONS_----> " + tt);
} catch (Exception e){
System.out.println ("ERROR: " + e);
e.printStackTrace (System.out);
}
}
}
IRPManager.java
import NotificationIRPSystem.*;
import org.omg.CosNotification.*;
import org.omg.CosNotifyComm.*;
import org.omg.CosNotifyChannelAdmin.*;
import org.omg.CosNotifyFilter.*;
import org.omg.CosNaming.*;
import org.omg.CosNaming.NamingContextPackage.*;
import org.omg.PortableServer.*;
public class IRPManager
extends SequencePushConsumerPOA
{
public void disconnect_sequence_push_consumer()
{
System.out.println("Disconnected!");
}
public void push_structured_events(org.omg.CosNotification.StructuredEvent[] notifications) throws org.omg.CosEvent\
Comm.Disconnected
{
System.out.println("Received Event");
}
public void offer_change(org.omg.CosNotification.EventType[] added, org.omg.CosNotification.EventType[] removed) th\
rows org.omg.CosNotifyComm.InvalidEventType
{
System.out.println("Offer Change!");
}
}
OUTPUT
14-Aug-2013 3:54:02 PM org.jacorb.orb.ORBSingleton <init>
INFO: created ORBSingleton
IOR===> IOR:000000000000001D49444C3A6F6D672E6F72672F434F5242412F4F626A6563743A312E3000000000000000010000000000000040000102000000000F3137322E33302E3132302E3135380000C06600000000000B4E616D655365727669636500000000010000000000000008000000004A414300
14-Aug-2013 3:54:03 PM org.jacorb.orb.giop.ClientConnectionManager getConnection
INFO: ClientConnectionManager: created new ClientGIOPConnection to 172.30.120.158:49254 (7bea5671)
14-Aug-2013 3:54:03 PM org.jacorb.orb.iiop.ClientIIOPConnection connect
INFO: Connected to 172.30.120.158:49254 from local port 57080
14-Aug-2013 3:54:03 PM org.jacorb.orb.iiop.ClientIIOPConnection close
INFO: Client-side TCP transport to 172.30.120.158:49254 closed.
14-Aug-2013 3:54:03 PM org.jacorb.orb.giop.ClientConnectionManager getConnection
INFO: ClientConnectionManager: created new ClientGIOPConnection to 10.20.0.4:49254 (58a17083)
14-Aug-2013 3:54:03 PM org.jacorb.orb.iiop.ClientIIOPConnection connect
INFO: Connected to 10.20.0.4:49254 from local port 52574
IOR:000000000000001849444C3A506F7274616C2F53657276696365733A312E300000000001000000000000007C000102000000000A31302E32302E302E3800C27B0000002000504D43000000040000000C2F466163746F7279504F4100000000040000000000000003564953030000000500070801FF000000000000000000000800000000564953000000000100000018000000000001000100000001050100010001010900000000
noti---->IOR:000000000000004449444C3A336770707361352E6F72672F4E6F74696669636174696F6E49525053797374656D2F4E6F74696669636174696F6E4952504F7065726174696F6E733A312E3000000000010000000000000088000102000000000A31302E32302E302E3800C27B0000002C00504D43000000040000000C2F466163746F7279504F4100000000104E4D534E41436F6E73756D6572322E3300000003564953030000000500070801FF000000000000000000000800000000564953000000000100000018000000000001000100000001050100010001010900000000
14-Aug-2013 3:54:03 PM org.jacorb.poa.AOM add
INFO: oid: 00 40 46 20 4B 4D 29 05 2A 07 10 06 30 46 38 14 14 1B 48 4C .#F KM).*...0F8...HL1B .object is activated
14-Aug-2013 3:54:03 PM org.jacorb.poa.POA getImplName
INFO: Using server ID (8624227886) for transient POA
MANAGER REF: IOR:000000000000003349444C3A6F6D672E6F72672F436F734E6F74696679436F6D6D2F53657175656E636550757368436F6E73756D65723A312E300000000000010000000000000050000102000000000E3137322E31362E32342E31353200B91900000020383632343232373838362F004046204B4D29052A07100630463814141B484C1B000000010000000000000008000000004A414300
OBJ NOTI SERVER: IOR:000000000000003349444C3A6F6D672E6F72672F436F734E6F74696679436F6D6D2F53657175656E636550757368436F6E73756D65723A312E300000000000010000000000000050000102000000000E3137322E31362E32342E31353200B91900000020383632343232373838362F004046204B4D29052A07100630463814141B484C1B000000010000000000000008000000004A414300
14-Aug-2013 3:54:03 PM org.jacorb.orb.giop.ClientConnectionManager getConnection
INFO: ClientConnectionManager: created new ClientGIOPConnection to 10.20.0.8:49787 (68e6ff0d)
14-Aug-2013 3:54:03 PM org.jacorb.orb.iiop.ClientIIOPConnection connect
INFO: Connected to 10.20.0.8:49787 from local port 35910
**************************************************************************org.jacorb.orb.giop.RequestOutputStream#1be1a408
14-Aug-2013 3:54:04 PM org.jacorb.orb.iiop.ClientIIOPConnection close
INFO: Client-side TCP transport to 10.20.0.8:49787 closed.
org.omg.CORBA.MARSHAL: Server-side Exception: null
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:532)
at org.jacorb.orb.SystemExceptionHelper.read(SystemExceptionHelper.java:222)
at org.jacorb.orb.ReplyReceiver.getReply(ReplyReceiver.java:456)
at org.jacorb.orb.Delegate._invoke_internal(Delegate.java:1418)
at org.jacorb.orb.Delegate.invoke_internal(Delegate.java:1187)
at org.jacorb.orb.Delegate.invoke(Delegate.java:1175)
at org.omg.CORBA.portable.ObjectImpl._invoke(ObjectImpl.java:80)
at NotificationIRPSystem.__NotificationIRPOperationsStub.attach_push(__NotificationIRPOperationsStub.java:288)
at AlarmClient.main(AlarmClient.java:43)
I'm not really sure what's happening. I think maybe that the object I'm trying to marshal is different from what the server expects? I hope someone can help, let me know if you need more information or if I missed something.
UPDATE
This is the _NotificationIRPOperationsStub code which contains attach_push
public java.lang.String attach_push(java.lang.String manager_reference, int time_tick, java.lang.String[] notification_category_set, java.lang.String filter) throws NotificationIRPSystem.AlreadySubscribed,NotificationIRPSystem.Attach,NotificationIRPSystem.AtLeast\
OneNotificationCategoryNotSupported,NotificationIRPSystem.ParameterNotSupported,NotificationIRPSystem.InvalidParameter
{
while(true)
{
if(! this._is_local())
{
org.omg.CORBA.portable.InputStream _is = null;
org.omg.CORBA.portable.OutputStream _os = null;
try
{
_os = _request( "attach_push", true);
java.lang.String tmpResult10 = manager_reference;
_os.write_string( tmpResult10 );
_os.write_ulong(time_tick);
NotificationIRPConstDefs.NotificationCategorySetHelper.write(_os,notification_category_set);
java.lang.String tmpResult11 = filter;
_os.write_string( tmpResult11 );
System.out.println("**************************************************************************" + _os.toString());
_is = _invoke(_os);
java.lang.String _result = _is.read_string();
return _result;
}
catch( org.omg.CORBA.MARSHAL m){
m.printStackTrace();
return m.getStackTrace().getClassName();
}
catch( org.omg.CORBA.portable.RemarshalException _rx )
{
continue;
}
catch( org.omg.CORBA.portable.ApplicationException _ax )
{
String _id = _ax.getId();
try
{
if( _id.equals("IDL:3gppsa5.org/NotificationIRPSystem/AlreadySubscribed:1.0"))
{
throw NotificationIRPSystem.AlreadySubscribedHelper.read(_ax.getInputStream());
}
else
if( _id.equals("IDL:3gppsa5.org/NotificationIRPSystem/Attach:1.0"))
{
throw NotificationIRPSystem.AttachHelper.read(_ax.getInputStream());
}
else
if( _id.equals("IDL:3gppsa5.org/NotificationIRPSystem/AtLeastOneNotificationCategoryNotSupported:1.0"))
{
throw NotificationIRPSystem.AtLeastOneNotificationCategoryNotSupportedHelper.read(_ax.getInputStream());
}
else
if( _id.equals("IDL:3gppsa5.org/NotificationIRPSystem/ParameterNotSupported:1.0"))
{
throw NotificationIRPSystem.ParameterNotSupportedHelper.read(_ax.getInputStream());
}
else
if( _id.equals("IDL:3gppsa5.org/NotificationIRPSystem/InvalidParameter:1.0"))
{
throw NotificationIRPSystem.InvalidParameterHelper.read(_ax.getInputStream());
}
else{
throw new RuntimeException("Unexpected exception " + _id );
}
}
finally
{
try
{
_ax.getInputStream().close();
}
catch (java.io.IOException e)
{
throw new RuntimeException("Unexpected exception " + e.toString() );
}
}
}
finally
{
if (_os != null)
{
try
{
_os.close();
}
catch (java.io.IOException e)
{
throw new RuntimeException("Unexpected exception " + e.toString() );
}
}
this._releaseReply(_is);
}
}
else
{
org.omg.CORBA.portable.ServantObject _so = _servant_preinvoke( "attach_push", _opsClass );
if( _so == null )
continue;
_NotificationIRPOperationsOperations _localServant = (_NotificationIRPOperationsOperations)_so.servant;
java.lang.String _result;
try
{
_result = _localServant.attach_push(manager_reference,time_tick,notification_category_set,filter);
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).normalCompletion();
return _result;
}
catch (NotificationIRPSystem.AlreadySubscribed ex)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(ex);
throw ex;
}
catch (NotificationIRPSystem.Attach ex)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(ex);
throw ex;
}
catch (NotificationIRPSystem.AtLeastOneNotificationCategoryNotSupported ex)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(ex);
throw ex;
}
catch (NotificationIRPSystem.ParameterNotSupported ex)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(ex);
throw ex;
}
catch (NotificationIRPSystem.InvalidParameter ex)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(ex);
throw ex;
}
catch (RuntimeException re)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(re);
throw re;
}
catch (java.lang.Error err)
{
if ( _so instanceof org.omg.CORBA.portable.ServantObjectExt)
((org.omg.CORBA.portable.ServantObjectExt)_so).exceptionalCompletion(err);
throw err;
}
finally
{
_servant_postinvoke(_so);
}
}
}
}
Thanks.
UPDATE # 2
Thanks to Brian for posting that version of the document. I took the IDL from there, and with a minor adjustment I was able to call attach_push with my objNotiServer. My next question is about how I can set my AlarmClient up to receive these notifications. My assumption is that I can call orb.run() and then when notifications come in, I will receive them in my IRPManager object, is that right?
CORBA::MARSHAL exceptions usually occur because of one of these reasons:
IDL mismatches between client and server (causing unexpected payload differences)
Crappy ORB products that don't know how to unmarshal complex but valid payloads correctly
Reason #2 is really unlikely nowadays, as ORBs have matured enough that this rarely happens. As long as you're using ORB products on both the client and server side that were built in this century then you're probably ok. That leaves reason #1.
The IDL for the method you're calling is (I believe) from this document, and looks like this:
interface NotificationIRPOperations {
NotificationIRPConstDefs::SubscriptionId attach_push (
in Object manager_reference,
in long time_tick,
in NotificationCategorySet notification_category_set,
in string filter
)
raises (Attach, ParameterNotSupported, InvalidParameter, AlreadySubscribed,
AtLeastOneNotificationCategoryNotSupported);
Note the first parameter, it's of type Object. However in your code, you're passing a string. However the IDL type Object maps to org.omg.CORBA.Object in Java, not to a String. I would have expected to see you pass objNotiServer instead for this parameter.
So either your IDL is mismatched with what your server is expecting (likely), or your IDL compiler isn't following the basic IDL-to-Java mapping rules (unlikely).
Either way, something smells bad there. I don't believe that the ORB will implicitly call ORB.string_to_object() for you, so that would lead to a regular string being sent on the network when the server is actually expecting a stringified Object reference. The difference is subtle but it might be the reason why the server is throwing a MARSHAL back at you.
So your first port of call should be a check on the IDL file that you're using to build your client. Make sure that you've got the exact and correct IDL for the server you're trying to call. If there's any mismatch at all then MARSHAL exceptions will happen, and happen a lot.
The server side has a problem, therefor you receive the server side exception.
Your IDL is different on Server/Client which causes this error.
You have to look at the server's logfile. If you do not have access to the server, implement your own dummy-server and find out whether or not it works.
Regarding the manager_reference, test it in your own code via: narrow and a call to _non_existent(). I.e.
org.omg.CORBA.Object managerObj = orb.string_to_object(manager_reference);
IRPManager managerImpl = IRPManagerHelper.narrow(managerObj);
managerImpl._non_existent()
If this works, your IRPManager servant works.
By the way, you have a lot of unused imports and should refactor them. Your IDE would help you.
For Update2
Yout Alarmclient has to implement StructuredPushConsumer (see Initializing Corba notification service in java successfully but cannot get any events in linux but it's done in windows) and subscribe to the notifications.
But I recomment to open a new question.
Your code looks suspicious, you never activate irpMan in the POA, but you do use servant_to_reference. Try the code below, handle it as pseudo code, I haven't tested it and I normally program C++.
IRPManager irpMan = new IRPManager();
POA poa = POAHelper.narrow(orb.resolve_initial_references("RootPOA"));
poa.the_POAManager().activate();
ObjectId id = poa.activate_object (irpMan);
org.omg.CORBA.Object objNotiServer = poa.id_to_reference(id);
String manager_reference = orb.object_to_string(objNotiServer);
System.out.println("MANAGER REF: " + manager_reference);
System.out.println("OBJ NOTI SERVER: " + objNotiServer);
int time_tick = 15;
String filter = "";
String[] asd = {};
NotificationIRPConstDefs.EventTypesSetHolder e_list = new NotificationIRPConstDefs.EventTypesSetHolder();
NotificationIRPConstDefs.ExtendedEventTypesSetHolder ex_list = new NotificationIRPConstDefs.ExtendedEventTypesSet\
Holder();
String cats[] = tt.get_notification_categories(e_list, ex_list);
String res = tt.attach_push(objNotiServer, time_tick, asd , filter);

ModuleLoadException when Thread.CurrentPrincipal is set with a principal containing a custom IIdentity

We have an application hosted in ASP.NET MVC WebAPI, which uses an external managed library. This managed library again uses a COM library. The COM library is a general purpose library used by many applications and in different versions, so we want to be able to to run it side-by-side by loading it directly from the bin folder and not not use the traditional COM registration. We have achieved this by adding a manifest file and making a call to CreateActCtx in kernel32.dll to register a custom activation context (code is included below).
This is working fine by itself, however when we set Thread.CurrentPrincipal to a GenericPrincipal instance with a custom IIdentity implementation before the COM library is loaded, we get a <CrtImplementationDetails>.ModuleLoadException with the message "The C++ module failed to load while attempting to initialize the default appdomain.". It has an inner System.Runtime.Serialization.SerializationException saying it cannot load the assembly which actually initiated the load of the COM assembly in the first place. The stacktrace is:
Server stack trace:
at System.Runtime.Serialization.Formatters.Binary.BinaryAssemblyInfo.GetAssembly()
at System.Runtime.Serialization.Formatters.Binary.ObjectReader.GetType(BinaryAssemblyInfo assemblyInfo, String name)
at System.Runtime.Serialization.Formatters.Binary.ObjectMap..ctor(String objectName, String[] memberNames, BinaryTypeEnum[] binaryTypeEnumA, Object[] typeInformationA, Int32[] memberAssemIds, ObjectReader objectReader, Int32 objectId, BinaryAssemblyInfo assemblyInfo, SizedArray assemIdToAssemblyTable)
at System.Runtime.Serialization.Formatters.Binary.__BinaryParser.ReadObjectWithMapTyped(BinaryObjectWithMapTyped record)
at System.Runtime.Serialization.Formatters.Binary.__BinaryParser.Run()
at System.Runtime.Serialization.Formatters.Binary.ObjectReader.Deserialize(HeaderHandler handler, __BinaryParser serParser, Boolean fCheck, Boolean isCrossAppDomain, IMethodCallMessage methodCallMessage)
at System.Runtime.Serialization.Formatters.Binary.BinaryFormatter.Deserialize(Stream serializationStream, HeaderHandler handler, Boolean fCheck, Boolean isCrossAppDomain, IMethodCallMessage methodCallMessage)
at System.Runtime.Remoting.Channels.CrossAppDomainSerializer.DeserializeObject(MemoryStream stm)
at System.Runtime.Remoting.Messaging.SmuggledMethodCallMessage.FixupForNewAppDomain()
at System.Runtime.Remoting.Channels.CrossAppDomainSink.DoDispatch(Byte[] reqStmBuff, SmuggledMethodCallMessage smuggledMcm, SmuggledMethodReturnMessage& smuggledMrm)
at System.Runtime.Remoting.Channels.CrossAppDomainSink.DoTransitionDispatchCallback(Object[] args)
Exception rethrown at [0]:
at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)
at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(MessageData& msgData, Int32 type)
at System.AppDomain.get_Id()
at <CrtImplementationDetails>.DoCallBackInDefaultDomain(IntPtr function, Void* cookie)
at <CrtImplementationDetails>.LanguageSupport._Initialize(LanguageSupport* )
at <CrtImplementationDetails>.LanguageSupport.Initialize(LanguageSupport* )
--- End of inner exception stack trace ---
at <CrtImplementationDetails>.LanguageSupport.Initialize(LanguageSupport* )
at .cctor()
--- End of inner exception stack trace ---
When we do not set Thread.CurrentPrincipal or use a GenericPrincipal with a GenericIdentity inside it works fine. For now the solution with a GenericIdentity is the workaround for us, because we are mainly setting Thread.CurrentPrincipal to take use of ASP.NET MVC authentication.
The following code works:
Thread.CurrentPrincipal = new GenericPrincipal(new GenericIdentity("bla"), new string[0]);
...while the following doesn't:
Thread.CurrentPrincipal = new GenericPrincipal(new CustomIdentity("bla"), new string[0]);
The following is the global.asax code being used to register the custom activation context referencing the bin folder and the manifest file to do reg-free COM:
public class Global : System.Web.HttpApplication
{
private const UInt32 ERROR_SXS_PROCESS_DEFAULT_ALREADY_SET = 14011;
IntPtr hActCtx;
private static readonly ILogger s_logger = LogManager.GetCurrentClassLogger();
protected void Application_Start(object sender, EventArgs e)
{
// Required files are copied in host projects post-build event
// NB! you can only set the default activation context like this once per app domain,
// so if you need multiple apps/services using side-by-side they cannot share app pool.
UInt32 dwError = 0;
string binDirectory = Path.Combine(AppDomain.CurrentDomain.BaseDirectory,"bin");
string manifestPath = Path.Combine(binDirectory, "Iit.OpenApi.Services.ReferenceData.manifest");
var activationContext = new ACTCTX();
activationContext.cbSize = Marshal.SizeOf(typeof(ACTCTX));
activationContext.dwFlags = ACTCTX_FLAG_SET_PROCESS_DEFAULT | ACTCTX_FLAG_ASSEMBLY_DIRECTORY_VALID;
activationContext.lpSource = manifestPath;
activationContext.lpAssemblyDirectory = binDirectory;
hActCtx = CreateActCtx(ref activationContext);
if (hActCtx.ToInt32() == -1)
{
dwError = (UInt32)Marshal.GetLastWin32Error();
}
if ((hActCtx.ToInt32() == -1) && (dwError != Global.ERROR_SXS_PROCESS_DEFAULT_ALREADY_SET))
{
var err = string.Format("Cannot create process-default win32 sxs context, error={0} manifest={1}", dwError, manifestPath);
throw new ApplicationException(err);
}
// TODO: add eventid to enum
if ((hActCtx.ToInt32() == -1) && (dwError == Global.ERROR_SXS_PROCESS_DEFAULT_ALREADY_SET))
{
GetCurrentActCtx(out hActCtx);
s_logger.Info(9998, string.Format("Re-using exising ActivationContext [{0}].", hActCtx.ToInt64()));
}
else
{
s_logger.Info(9998, string.Format("ActivationContext [{0}] created successfully.", hActCtx.ToInt64()));
}
}
protected void Application_End()
{
if (hActCtx.ToInt64() != INVALID_HANDLE_VALUE)
{
ReleaseActCtx(hActCtx);
// TODO: add eventid to enum
s_logger.Info(9999, string.Format("ActivationContext [{0}] released successfully.", hActCtx.ToInt64()));
}
}
[StructLayout(LayoutKind.Sequential)]
private struct ACTCTX
{
public int cbSize;
public uint dwFlags;
public string lpSource;
public ushort wProcessorArchitecture;
public ushort wLangId;
public string lpAssemblyDirectory;
public string lpResourceName;
public string lpApplicationName;
}
private const uint ACTCTX_FLAG_ASSEMBLY_DIRECTORY_VALID = 0x004;
private const uint ACTCTX_FLAG_SET_PROCESS_DEFAULT = 0x010;
private const Int64 INVALID_HANDLE_VALUE = -1;
[DllImport("kernel32.dll", SetLastError = true)]
private static extern IntPtr CreateActCtx(ref ACTCTX actctx);
[DllImport("kernel32.dll", SetLastError = true)]
private static extern bool GetCurrentActCtx(out IntPtr hActCtx);
[DllImport("kernel32.dll", SetLastError = true)]
private static extern void ReleaseActCtx(IntPtr hActCtx);
}
Any input to what is going wrong when using a custom IIdentity, and how I can make it work with reg-free COM?

How do I get the Service Display name in C++?

I am trying to get the display name of the running service using c++. I was trying to use the GetServiceDisplayName function but it does not seem to be working, not sure why.
TTServiceBegin( const char *svcName, PFNSERVICE pfnService, bool *svc, PFNTERMINATE pfnTerm,
int flags, int argc, char *argv[], DWORD dynamiteThreadWaitTime )
{
SC_HANDLE serviceStatusHandle;
DWORD dwSizeNeeded = 0 ;
TCHAR* szKeyName = NULL ;
serviceStatusHandle=OpenSCManager(NULL, SERVICES_ACTIVE_DATABASE ,SC_MANAGER_ALL_ACCESS);
GetServiceDisplayName(serviceStatusHandle,svcName, NULL, &dwSizeNeeded);
if(dwSizeNeeded)
{
szKeyName = new char[dwSizeNeeded+1];
ZeroMemory(szKeyName,dwSizeNeeded+1);
if(GetServiceDisplayName(serviceStatusHandle ,svcName,szKeyName,&dwSizeNeeded)!=0)
{
MessageBox(0,szKeyName,"Got the key name",0);
}
}
When i run this code, i can never see the value of szKeyName in my debugger and it goes into the if block for the message box but never displays the message box. Not sure why?
Anyway to get this to work to get the display name of the service or any other/easier way to accomplish that task?
You need to use the WTSSendMessage instead of the MessageBox to interact with the active session.
WTS_SESSION_INFO* pSessionInfo = NULL;
DWORD dwSessionsCount = 0;
if(WTSEnumerateSessions(WTS_CURRENT_SERVER_HANDLE, 0, 1, &pSessionInfo, &dwSessionsCount))
{
for(int i=0; i<(int)dwSessionsCount; i++)
{
WTS_SESSION_INFO &si = pSessionInfo[i];
if(si.State == WTSActive)
{
DWORD dwIdCurrentSession = si.SessionId;
std::string strTitle = "Hello";
std::string strMessage = "This is a message from the service";
DWORD dwMsgBoxRetValue = 0;
if(WTSSendMessage(
WTS_CURRENT_SERVER_HANDLE,
dwIdCurrentSession,
(char*)strTitle.c_str(),
strTitle.size(),
(char*)strMessage.c_str(),
strMessage.size(),
MB_RETRYCANCEL | MB_ICONINFORMATION | MB_TOPMOST,
60000,
&dwMsgBoxRetValue,
TRUE))
{
switch(dwMsgBoxRetValue)
{
case IDTIMEOUT:
// Deal with TimeOut...
break;
case IDCANCEL:
// Deal With Cancel....
break;
}
}
else
{
// Deal With Error
}
break;
}
}
WTSFreeMemory(pSessionInfo);
}
The message box will not be visible on Windows Vista and later due to a change that has services running in a separate session (Session 0 Isolation) that does not have access to a desktop so the message box would not be visible to you, the logged on user.
On Window XP and earlier, you need to tick the Allow service to interact with desktop checkbox under the Log On tab in the service's properties dialog for your service to make message box appear.
Instead, you could write the service name out to a file or run a user application that accepts the name of the service to query and have it query and display the service name (I just tried with the posted code and it works correctly, displaying the message box).

How do I invoke Multiple Startup Projects when running a unit tests in Debug Mode

This seems like a simple thing to do but I can't seem to find any info anywhere! I've got a solution that has a service that we run in 'Console Mode' when debugging. I want it to be started and 'attached' when I run my unit test from Visual Studio.
I'm using Resharper as the unit test runner.
Not a direct answer to your question, BUT
We faced a similar problem recently and eventually settled on a solution using AppDomain
As your solution is already running as a Console project it would be little work to make it boot in a new AppDomain. Furthermore, you could run Assertions on this project as well as part of unit testing. (if required)
Consider the following static class Sandbox which you can use to boot multiple app domains.
The Execute method requires a Type which is-a SandboxAction. (class definition also included below)
You would first extend this class and provide any bootup actions for running your console project.
public class ConsoleRunnerProjectSandbox : SandboxAction
{
protected override void OnRun()
{
Bootstrapper.Start(); //this code will be run on the newly create app domain
}
}
Now to get your app domain running you simply call
Sandbox.Execute<ConsoleRunnerProjectSandbox>("AppDomainName", configFile)
Note you can pass this call a config file so you can bootup your project in the same fashion as if you were running it via the console
Any more questions please ask.
public static class Sandbox
{
private static readonly List<Tuple<AppDomain, SandboxAction>> _sandboxes = new List<Tuple<AppDomain, SandboxAction>>();
public static T Execute<T>(string friendlyName, string configFile, params object[] args)
where T : SandboxAction
{
Trace.WriteLine(string.Format("Sandboxing {0}: {1}", typeof (T).Name, configFile));
AppDomain sandbox = CreateDomain(friendlyName, configFile);
var objectHandle = sandbox.CreateInstance(typeof(T).Assembly.FullName, typeof(T).FullName, true, BindingFlags.Default, null, args, null, null, null);
T sandBoxAction = objectHandle.Unwrap() as T;
sandBoxAction.Run();
Tuple<AppDomain, SandboxAction> box = new Tuple<AppDomain, SandboxAction>(sandbox, sandBoxAction);
_sandboxes.Add(box);
return sandBoxAction;
}
private static AppDomain CreateDomain(string name, string customConfigFile)
{
FileInfo info = customConfigFile != null ? new FileInfo(customConfigFile) : null;
if (!string.IsNullOrEmpty(customConfigFile) && !info.Exists)
throw new ArgumentException("customConfigFile not found using " + customConfigFile + " at " + info.FullName);
var appsetup = new AppDomainSetup();
//appsetup.ApplicationBase = Path.GetDirectoryName(typeof(Sandbox).Assembly.Location);
appsetup.ApplicationBase = AppDomain.CurrentDomain.SetupInformation.ApplicationBase;
if (customConfigFile==null)
customConfigFile = AppDomain.CurrentDomain.SetupInformation.ConfigurationFile;
appsetup.ConfigurationFile = customConfigFile;
var sandbox = AppDomain.CreateDomain(
name,
AppDomain.CurrentDomain.Evidence,
appsetup);
return sandbox;
}
public static void DestroyAppDomainForSandbox(SandboxAction action)
{
foreach(var tuple in _sandboxes)
{
if(tuple.Second == action)
{
AppDomain.Unload(tuple.First);
Console.WriteLine("Unloaded sandbox ");
_sandboxes.Remove(tuple);
return;
}
}
}
}
[Serializable]
public abstract class SandboxAction : MarshalByRefObject
{
public override object InitializeLifetimeService()
{
return null;
}
public void Run()
{
string name = AppDomain.CurrentDomain.FriendlyName;
Log.Info("Executing {0} in AppDomain:{1} thread:{2}", name, AppDomain.CurrentDomain.Id, Thread.CurrentThread.ManagedThreadId);
try
{
OnRun();
}
catch (Exception ex)
{
Log.Error(ex, "Exception in app domain {0}", name);
throw;
}
}
protected abstract void OnRun();
public virtual void Stop()
{
}
}

Why does SQL Server CLR procedure hang in GetResponse() call to web service

Environment: C#, .Net 3.5, Sql Server 2005
I have a method that works in a stand-alone C# console application project. It creates an XMLElement from data in the database and uses a private method to send it to a web service on our local network. When run from VS in this test project, it runs in < 5 seconds.
I copied the class into a CLR project, built it, and installed it in SQL Server (WITH PERMISSION_SET = EXTERNAL_ACCESS). The only difference is the SqlContext.Pipe.Send() calls that I added for debugging.
I am testing it by using an EXECUTE command one stored procedure (in the CLR) from an SSMS query window. It never returns. When I stop execution of the call after a minute, the last thing displayed is "Calling GetResponse() using http://servername:53694/odata.svc/Customers/". Any ideas as to why the GetResponse() call doesn't return when executing within SQL Server?
private static string SendPost(XElement entry, SqlString url, SqlString entityName)
{
// Send the HTTP request
string serviceURL = url.ToString() + entityName.ToString() + "/";
HttpWebRequest request = (HttpWebRequest)HttpWebRequest.Create(serviceURL);
request.Method = "POST";
request.Accept = "application/atom+xml,application/xml";
request.ContentType = "application/atom+xml";
request.Timeout = 20000;
request.Proxy = null;
using (var writer = XmlWriter.Create(request.GetRequestStream()))
{
entry.WriteTo(writer);
}
try
{
SqlContext.Pipe.Send("Calling GetResponse() using " + request.RequestUri);
WebResponse response = request.GetResponse();
SqlContext.Pipe.Send("Back from GetResponse()");
/*
string feedData = string.Empty;
Stream stream = response.GetResponseStream();
using (StreamReader streamReader = new StreamReader(stream))
{
feedData = streamReader.ReadToEnd();
}
*/
HttpStatusCode StatusCode = ((HttpWebResponse)response).StatusCode;
response.Close();
if (StatusCode == HttpStatusCode.Created /* 201 */ )
{
return "Created # Location= " + response.Headers["Location"];
}
return "Creation failed; StatusCode=" + StatusCode.ToString();
}
catch (WebException ex)
{
return ex.Message.ToString();
}
finally
{
if (request != null)
request.Abort();
}
}
The problem turned out to be the creation of the request content from the XML. The original:
using (var writer = XmlWriter.Create(request.GetRequestStream()))
{
entry.WriteTo(writer);
}
The working replacement:
using (Stream requestStream = request.GetRequestStream())
{
using (var writer = XmlWriter.Create(requestStream))
{
entry.WriteTo(writer);
}
}
You need to dispose the WebResponse. Otherwise, after a few calls it goes to timeout.
You are asking for trouble doing this in the CLR. And you say you are calling this from a trigger? This belongs in the application tier.
Stuff like this is why when the CLR functionality came out, DBAs were very concerned about how it would be misused.