d642e11438
winapi_test: Regenerate tests.
...
Remove structs from test list which have preprocessor directives in
their definitions (winapi_test can't handle that).
2006-10-11 12:12:21 +02:00
14c551bc3a
rpcrt4/tests: Win64 printf format warning fixes.
2006-10-09 12:44:16 +02:00
82babbee0b
Add missing '\n's to Wine traces.
2006-10-05 17:46:48 +02:00
2324aeec33
rpcrt4: Use I_Rpc{Allocate, Free} instead of Heap{Alloc, Free} as that
...
the former are exported by rpcrt4 seemingly to allow callers of tower
functions to free the allocated memory.
2006-10-03 16:14:57 +02:00
7dab87e2e9
rpcrt4: RpcBindingCopy is implemented.
2006-10-03 16:14:49 +02:00
b9423076ba
rpcrt4: Don't use HEAP_ZERO_MEMORY when most of the fields will be initialised to non-zero values.
2006-10-03 16:14:47 +02:00
1011bc47bf
rpcrt4: Don't return void values. This fixes Visual C++ warnings.
2006-10-02 11:54:43 +02:00
68e84a7a53
Add missing '\n's to ok() calls.
2006-09-29 17:53:39 +02:00
85ce5117c5
Add a WINE_NO_LONG_INT define to many makefiles to ease the transition of DWORD/LONG/ULONG from long to int.
2006-09-29 16:08:30 +02:00
bc0f9c49b0
rpcrt4: Fix RpcMgmtSetServerStackSize prototype.
2006-09-28 14:03:50 +02:00
8221beb225
rpcrt4: Cast-qual warnings fix.
2006-09-26 11:40:52 +02:00
3f123e51bf
rpcrt4: Cast-qual warnings fix.
2006-09-26 11:40:47 +02:00
269f77d872
rpcrt4: Cast-qual warnings fix.
2006-09-26 11:40:41 +02:00
7147f79153
rpcrt4: Cast-qual warnings fix.
2006-09-26 11:40:33 +02:00
32ab895e46
makefiles: Generate the dependencies line to avoid some code duplication.
2006-09-14 09:38:16 +02:00
06f9532dda
Assorted spelling fixes.
2006-09-08 11:04:53 +02:00
a6e545a27c
rpcrt4: Raise an exception if IRpcChannelBuffer_GetBuffer fails.
...
Make StdProxy_GetIID and StdProxy_GetChannel return void.
Don't bother checking for This and pRpcChannelBuffer being NULL as
native doesn't.
2006-09-07 11:42:05 +02:00
bd9891ffed
rpcrt4: Add delegating stub support.
2006-09-07 10:57:39 +02:00
b058e6e729
rpcrt4: Add infrastructure to create and manage a variable sized vtbl
...
that will be used as the server object for delegated stubs.
2006-09-07 10:57:27 +02:00
9b91307677
rpcrt4: Tests on NdrCStdStubBuffer_Release to show that we shouldn't call Disconnect.
...
Also call IPSFactoryBuffer_Release on the 2nd arg, rather than on This->pPSFactory.
2006-09-07 10:24:54 +02:00
86aa95fd2b
rpcrt4: Tidy up CStdStubBuffer_QueryInterface.
2006-09-07 10:23:00 +02:00
dffd3e5034
rpcrt4: Tidy up CStdStubBuffer_AddRef,Release.
2006-09-07 10:22:35 +02:00
ada61620de
rpcrt4: Make the reference counting in the standard proxy thread-safe.
2006-09-07 10:19:38 +02:00
3e01da1838
rpcrt4: Fix NdrStubGetBuffer to use the IRpcChannelBuffer object since
...
IRpcChannelBuffer_GetBuffer may put extra data onto the wire for ORPC
calls.
2006-09-07 10:18:52 +02:00
3535b70a0b
rpcrt4: Remove some unused code.
2006-09-07 10:18:48 +02:00
5515633517
rpcrt4: Rename ndr_midl.c to ndr_clientserver.c to more accurately
...
idenitfy what the code contained does - it is a number of helpers for
generated client/server code.
2006-09-07 10:18:31 +02:00
583ae12233
rpcrt4: Move NdrStub* functions to cstub.c and move NdrProxy*
...
functions to cproxy.c since both sets of functions depend on the
implementations of the standard stubs and proxies.
2006-09-07 10:17:28 +02:00
d28698dd73
rpcrt4: Add a test for IRpcStubBuffer_Connect on a delegated stub.
...
This shows that we have a vtbl ptr at (void**)This-2 that is passed to
the base interface's IRpcStubBuffer_Connect. This vtbl should
presumably forward calls to the real object's vtbl. We can't simply
pass the real object to the base interface's stub, as the real
object's QI may not provide us with the base interface.
2006-08-29 12:26:31 +02:00
161f60071e
rpcrt4: Add a helper function to create a stub.
2006-08-29 12:26:22 +02:00
5cf75e615a
makefiles: Only run test and documentation targets in directories that need them.
2006-08-28 13:37:44 +02:00
487c691dc5
rpcrt4: Tests for CStdStubBuffer_Disconnect and a tidy up of the implementation.
2006-08-23 20:10:36 +02:00
f42f727f7c
rpcrt4: Tests for CStdStubBuffer_Connect and fixes to make them pass.
2006-08-23 20:10:15 +02:00
cadc39842b
rpcrt4: Add tests for CreateStub and fix up CStdStubBuffer_Construct to match.
2006-08-23 20:10:06 +02:00
6a0fccd3a9
rpcrt4: Add tests for NdrStubForwardingFunction and write a commented out implementation.
...
This shows that the IRpcStubBuffer for a delegated stub's base
interface is stored at (void**)This-1.
2006-08-23 20:09:52 +02:00
a5dcb3e39d
rpcrt4: Tests for NdrDllGetClassObject.
...
These show that delegated stubs need different Connect, Disconnect and
CountRefs methods than non-delegated stubs (ie they don't use all of
the CStdStubBuffer_* functions).
2006-08-23 20:09:04 +02:00
998e256c82
Fix some -Wstrict-prototype warnings () -> (void).
2006-08-21 12:06:26 +02:00
59be6e0600
rpcrt4: Remember to skip over the conformance description.
2006-08-18 11:34:26 +02:00
d9b6672530
rpcrt4: Add tests for NdrAllocate.
2006-08-17 11:19:17 +02:00
4bbc3ca8dc
rpcrt4: Move processing of the arguments from NdrClientCall2 to two new functions.
2006-08-16 16:41:42 +02:00
68e867da81
rpcrt4: Forward NdrClientCall to NdrClientCall2.
...
NdrClientCall should be implemented as a separate function when
NdrClientCall2 has been factorised enough.
2006-08-09 20:41:34 +02:00
f953cc406a
rpcrt4: Add some more traces and make the stack offsets easier to read in the stubless interpreter.
2006-08-09 20:41:21 +02:00
88bbbed176
rpcrt4: Change call_memory_sizer to preserve the buffer contents and
...
initialise the memory size field to zero before calling the memory
sizer function.
2006-08-09 20:41:13 +02:00
d29108c396
rpcrt4: Set retval_ptr and allocate memory for out pointers for the old format stubless intepreter.
2006-08-09 20:41:01 +02:00
3db5f872a0
rpcrt4: Support fixed arrays in calc_arg_size.
2006-08-09 20:40:40 +02:00
e26bcb7fb2
rpcrt4: Complex types should be dereferenced an extra time in the old format stubless code.
2006-08-09 20:40:30 +02:00
64828b7520
rpcrt4: Fix the stack offset for the old format stubless code when an object's method is being called.
2006-08-09 20:40:25 +02:00
488c0a0423
rpcrt4: Make the error output on a failed binding more informative
...
by dumping information about the interface it is trying to bind to.
2006-08-04 20:27:13 +02:00
dd000074fe
rpcrt4: Support for SMVARRAY in EmbeddedComplexSize.
2006-07-25 11:31:59 +02:00
5700fe8a61
rpcrt4: Set BufferMark before calling EmbeddedPointerMarshall, EmbeddedPointerUnmarshall or EmbeddedPointerMemorySize.
2006-07-24 13:47:40 +02:00
ba9158e2e1
rpcrt4: Improve the stubs for RpcImpersonateClient and RpcRevertToSelf.
2006-07-24 13:47:36 +02:00
a8531067c4
rpcrt4: Support small fixed arrays in EmbeddedComplexSize.
2006-07-24 13:47:31 +02:00
4c4fc14fe9
rpcrt4: Allocate memory based on MaxCount, not ActualCount in NdrConformantStringUnmarshall.
2006-07-24 13:47:25 +02:00
98c26a81fe
rpcrt4: Add support for marshaling the FC_HYPER base type in complex types.
...
Change the FIXME messages to print out unrecognised format characters in
hex rather than decimal.
2006-07-24 13:47:13 +02:00
d2cb9ec567
rpcrt4: Display a big message box if the user tries to use typelib
...
marshaling in native oleaut32 with builtin rpcrt4 to tell them what is
wrong.
2006-07-18 11:13:36 +02:00
9318cb0fbc
rpcrt4: Handle more structure padding values.
2006-07-18 11:12:47 +02:00
2cb3b22c26
rpcrt4: Fix Variance Callback Functions.
...
A callback function for computing the variance puts the computed value
into the same field as a conformance, so make sure to copy the value
into pCount before returning from ComputeConformanceOrVariance.
2006-07-14 12:38:53 +02:00
a02034a609
rpcrt4: Add the threading model and a description when registering a CLSID for the PSFactoryBuffer.
2006-07-14 12:38:48 +02:00
2538490d0b
rpcrt4: Add a stub for I_RpcMapWin32Status.
2006-07-13 16:25:33 +02:00
eb8a40c368
libs: Merged libwine_unicode into libwine.
...
Most dlls use libwine_unicode at this point, so there's nothing to
gain by having two separate libraries.
2006-07-11 19:33:22 +02:00
0fa8baafcb
make_dlls: Recursively ignore .ok files in all tests directories.
2006-07-10 21:26:50 +02:00
edd100b899
make_dlls: Recursively ignore testlist.c in all tests directories.
2006-07-10 21:19:42 +02:00
5258e061dc
make_dlls: Ignore generated import libraries from the top-level .gitignore.
2006-07-10 20:55:58 +02:00
cc2b3b48e6
Take advantage of the recursive nature of .gitignore for Makefile entries.
...
Prefix other entries with '/' to make them non-recursive.
2006-07-10 20:47:54 +02:00
00e779370f
janitorial: Put "inline static" at the beginning of a declaration.
2006-07-07 14:30:35 +02:00
85ed9a5905
rpcrt4: Don't advance the buffer for FC_STRUCTPAD2.
2006-07-05 11:44:19 +02:00
e7df38c27f
rpcrt4: Add a stub for RpcMgmtSetServerStackSize.
2006-06-30 20:43:53 +02:00
93fb9d10f5
rpcrt4: Calculate size of FC_STRUCT arguments correctly.
2006-06-26 12:36:10 +02:00
7d92b89abf
rpcrt4: Stubs for RpcRevertToSelf and RpcMgmtSetComTimeout.
2006-06-21 13:39:52 +02:00
45d42528a0
rpcrt4: Add stub for RpcMgmtEnableIdleCleanup.
2006-06-21 13:39:38 +02:00
3a56eca1c7
rpcrt4: Don't use a short to hold the packet length, since it can
...
result in premature termination of the send.
2006-06-20 20:35:38 +02:00
0da8836ff7
rpcrt4: Move the packet manipulation from the exception filter to the exception handler.
...
Remove the unused server_sem variable.
2006-06-19 23:04:19 +02:00
1437f8bd8c
rpcrt4: Use the thread pool in the RPC server for processing packets.
2006-06-19 10:28:50 +02:00
26c2953e58
rpcrt4: Fix the argument size of conformant arrays.
2006-06-16 14:42:36 +02:00
b3c04e032d
rpcrt4: getaddrinfo returns any non-zero value on failure, not just negative values.
2006-06-14 12:47:05 +02:00
2a91838f0c
rpcrt4: Don't use sizeof in traces to avoid printf format warnings.
2006-06-13 14:11:28 +02:00
d90f68166c
rpcrt4/tests: Don't use sizeof in ok() to avoid printf format warnings.
2006-06-13 14:03:44 +02:00
1c04d9ab20
rpcrt4: Add an exception handler for CStdStubBuffer_Invoke.
...
This is needed because IRpcStubBuffer::Invoke should not allow RPC
exceptions to be passed to the caller.
2006-06-12 13:21:46 +02:00
35982d72f5
rpcrt4: Initialise pStubMsg->MemorySize to zero before calling
...
ComplexStructMemorySize from NdrComplexArrayUnmarshall.
2006-06-12 13:21:46 +02:00
539b48bff4
rpcrt4: Use safe_mutiply in conformant array functions.
2006-06-12 13:21:46 +02:00
8dea3c2aa8
rpcrt4: Pass in a maximum variance value to ReadVariance to allow us
...
to validate the conformance values being read from the wire.
2006-06-12 13:21:46 +02:00
613ee6d01c
rpcrt4: Check that strings are null-terminated on unmarshaling of conformant-varying structs.
2006-06-12 13:21:46 +02:00
a946eb420b
rpcrt4: Use safe_multiply when multiplying anything by pStubMsg->ActualCount or pStubMsg->MaxCount.
2006-06-12 13:21:46 +02:00
cec6092aa2
rpcrt4: Raise an exception during unmarshaling if a conformant string isn't null-terminated.
2006-06-12 13:21:46 +02:00
6e95bfe85a
rpcrt4: Introduce a new function, safe_multiply, which will raise an
...
exception if a multiply overflows a 4-byte integer.
This will protect the unmarshaling code against attacks specifying a
large variance.
Use this new function in the conformant string functions to harden them
against attack.
2006-06-12 13:20:30 +02:00
d98b73a17a
rpcrt4: Clear an object pointer during unmarshaling, otherwise we
...
could end up using the previously allocated memory even if it isn't
the right size, causing heap corruption.
2006-06-09 21:07:57 +02:00
578df90d6d
rpcrt4: Initialise pStubMsg->MemorySize to zero before calling ComplexStructMemorySize from NdrComplexArrayUnmarshall.
2006-06-09 21:07:44 +02:00
e3df112a39
rpcrt4: Don't raise an exception if there is no freer function - not all types have a freer.
2006-06-09 21:07:37 +02:00
21673d03f9
rpcrt4: Implement more base types for complex types.
2006-06-09 21:07:22 +02:00
0b1e51ca25
rpcrt4: Now that there's a process-wide pool of connections we don't
...
need to copy the FromConn field when copying bindings.
2006-06-09 21:07:11 +02:00
56386bd88c
rpcrt4: Return the status received from RPCRT4_Send in I_RpcSend instead of always returning RPC_S_OK.
2006-06-09 21:07:02 +02:00
c32dc6744c
rpcrt4: Make the protseq_ops structure constant.
2006-06-09 13:12:58 +02:00
c0fcba798b
rpcrt4: Avoid potential redefinition of uuid_t.
2006-06-08 11:29:26 +02:00
58ec282bfb
rpcrt4: Implement a process-wide connection pool for client connections,
...
rather than relying on the restriction of one connection per
binding. This also avoids the problem of two threads using the same
connection at the same time.
2006-06-08 10:16:00 +02:00
1412c47a50
rpcrt4: Fix the retrieval of This for interpreted proxies.
2006-06-08 10:15:36 +02:00
7ef5c3c772
rpcrt4: Implement full-pointer support for interpreted stubs.
...
Move the calls to the full pointer init functions before
NdrClient/ServerInitializeNew, like in MIDL-generated code.
2006-06-08 10:15:22 +02:00
c9779ca16b
rpcrt4: Implement varying array NDR functions.
2006-06-05 13:25:09 +02:00
cd19c99284
rpcrt4: Don't call NdrBaseTypeMemorySize from within NdrBaseTypeUnmarshall.
2006-06-05 13:24:54 +02:00
0fad0895d7
rpcrt4: Don't try to resolve the name when constructing a protocol tower.
...
Instead just fill it in with all zeros.
2006-06-03 11:42:54 +02:00
ebcd6b6a24
rpcrt4: Change ComputeConformance and ComputeVariance from macros to
...
inline functions. Clear pStubMsg->Offset in ComputeVariance since it
isn't set by ComputeConformanceOrVariance and so could be left filled
with garbage.
2006-06-03 11:42:36 +02:00
a48433453b
rpcrt4: The lifetime of the binding handle stored in the context
...
handle is independent of the binding handle used in the call that
unmarshaled it, so we should use RpcBindingCopy to make a copy of it
and then call RpcBindingFree later to free it.
2006-06-03 11:42:25 +02:00
a4fff73ba3
rpcrt4: Fix an incorrect format string pointer being passed in to
...
ComputeVariance in NdrConformantVaryingStructBufferSize.
2006-06-03 11:42:13 +02:00
6a11aba526
rpcrt4: Skip over the right-hand-side data for the protocol floor so
...
that we can parse the towers received from native rpcrt4.
2006-06-03 11:42:03 +02:00
da44126395
rpcrt4: Align the buffer correctly when marshaling and unmarshaling context handles.
2006-06-03 11:41:51 +02:00
ec5b3c3ea6
rpcrt4: Freeing a NULL context handle is allowed.
2006-06-03 11:41:42 +02:00
1a346a4a13
rpcrt4: NdrClientContextMarshall/Unmarshall should both increment the buffer pointer.
2006-06-03 11:41:33 +02:00
dc6c640cd2
rpcrt4: Fix the unmarshaling of pointers in conformant varying arrays,
...
by passing the correct type format string to ReadConformance.
2006-06-03 11:41:22 +02:00
972ce2f635
rpcrt4: CContext can be NULL in NDRCContextMarshall, in which case we should marshall all-zeros.
2006-06-03 11:41:07 +02:00
678ad6286c
rpcrt4: Fix the pointer count used when handling embedded pointers in varying arrays.
2006-06-03 11:40:45 +02:00
423e7e3089
rpcrt4: Print some more informative error messages when getaddrinfo fails.
2006-06-03 11:40:42 +02:00
efc974d1f5
rpcrt4: All of the parameters to TowerExplode are optional.
2006-06-02 11:24:29 +02:00
e15e60b9f9
rpcrt4: Implement TowerConstruct and TowerExplode.
...
Add some tests for these undocumented functions.
2006-06-02 11:24:20 +02:00
37364b836f
rpcrt4: Add functions for generating and parsing the upper floors in endpoint-mapper towers for different transports.
2006-06-02 11:24:03 +02:00
84ab2e78a4
rpcrt4: Implement I_RpcAllocate and I_RpcFree.
2006-06-02 11:22:56 +02:00
cc9baf364e
rpcrt4: Don't send the unused part of the allocated buffer.
2006-06-02 11:22:44 +02:00
11da87d098
rpcrt4: NdrClientInitializeNew shouldn't clear all of the stub message, only selected fields.
2006-06-02 11:22:38 +02:00
7f98594f75
rpcrt4: Implement NDRCContext(Un)marshall and NDRCContextBinding.
2006-06-01 11:48:33 +02:00
bb835ae438
rpcrt4: Add ENUM32 support to ComplexStructMemorySize.
2006-05-31 15:48:11 +02:00
115525f3c4
rpcrt4: Make UserMarshalFlags static.
2006-05-30 12:25:22 +02:00
df57b025c6
rpcrt4: We should never pass an fMustAlloc value of TRUE into the
...
unmarshaler routines of embedded complexes in complex struct and array
unmarshaling.
Fixes a regression caused by commit
19fad8e710
2006-05-30 12:25:17 +02:00
cafd2ab37e
rpcrt4: Make sure to set a unique pointer with an id of 0 to NULL when unmarshaling.
2006-05-30 12:25:03 +02:00
11eb8ee042
rpcrt4: If the number of pointers is 0 for NdrFullPointerXlatInit then
...
set the number of pointers to a large default, so that NumberOfBuckets
doesn't end up less than zero.
2006-05-30 12:24:53 +02:00
ff9fd9c430
rpcrt4: Handle marshaling/unmarshaling full pointers.
2006-05-30 12:24:41 +02:00
bf985900e2
rpcrt4: For full pointers the state consists of flags, so or new flags into the field instead of assigning.
...
The return value of all full pointer functions that return an int
depends on the passed in query type being a flag in the state.
2006-05-30 12:24:22 +02:00
63307241dd
rpcrt4: Store the QueryType in the state table in NdrFullPointerQueryRefId.
...
Implement NdrFullPointerFree.
2006-05-27 10:54:39 +02:00
7af506b2b4
rpcrt4: Fix NdrFullPointerQueryRefId to make more tests pass.
2006-05-27 10:54:21 +02:00
767631d853
rpcrt4: Implement remaining part of NdrFullPointerQueryPointer and implement NdrFullPointerInsertRefId.
2006-05-27 10:53:48 +02:00
eb0fe9e6e3
rpcrt4: Implement some full pointer functions and add stubs for others.
...
Add some tests.
2006-05-27 10:53:29 +02:00
a534fdec32
rpcrt4: Implement NdrClientContext(Un)Marshall.
2006-05-26 20:21:00 +02:00
013d00f094
rpcrt4: Even though FC_DEREFERENCE conformance is handled elsewhere in
...
ComputeConformanceOrVariance it is still necessary to set the count
variable to the value retrieved.
2006-05-26 20:15:49 +02:00
8b56e88276
rpcrt4: Return the actual status code sent back by the server, rather than RPC_S_CALL_FAILED.
2006-05-26 20:15:44 +02:00
42bf8254c0
rpcrt4: Stub implementations for NDRS marshalling functions.
2006-05-25 10:45:01 +02:00
ec2a9cef3a
rpcrt4: Set the Memory field of the stub message structure to the
...
start of the memory that an embedded pointer is embedded in so that
FC_POINTER conformance works correctly.
2006-05-25 10:40:35 +02:00
4e85273086
rpcrt4: Implement explicit generic binding handles.
2006-05-25 10:40:23 +02:00
0030b6dacd
rpcrt4: Move the getting of the client binding handle to a separate function.
2006-05-25 10:40:09 +02:00
fdda17deea
rpcrt4: Move the initialising of the stub message structure nearer to the top of NdrClientCall2.
...
Set stubMsg.StackTop after calling the initialise function.
2006-05-25 10:39:50 +02:00
9d87447455
rpcrt4: Store the arguments exclusively in stubMsg.StackTop.
2006-05-25 10:39:22 +02:00
b9f36c3fa2
rpcrt4: Implement RpcBindingCopy.
2006-05-25 10:39:02 +02:00
9ca8a86914
Link against libwine_unicode everywhere we use Unicode functions.
2006-05-24 18:05:40 +02:00
23f9e358ad
Tests should not use wine/unicode.h.
2006-05-24 18:05:09 +02:00
4def9cc022
rpcrt4: Add stub implementations for NDRCContext(Un)Marshall.
2006-05-24 14:38:14 +02:00
00598e49cf
rpcrt4: Uncomment test now that it works correctly.
2006-05-23 17:51:50 +02:00
69d18c48b1
rpcrt4: Add another static.
2006-05-23 17:51:44 +02:00
9ba57efc18
rpcrt4: Fix some comments and move the NdrSimpleType* functions out from amongst the NdrSimpleStruct* ones.
2006-05-23 17:51:19 +02:00
360a3f9142
Update the address of the Free Software Foundation.
2006-05-23 14:11:13 +02:00
dd362a62a1
rpcrt4: Fix copying too much data from the buffer by not taking into account the fact that auth_pad_len has been included in the fragment length.
2006-05-23 11:36:37 +02:00
e91da1458d
rpcrt4: Authentication padding should only be generated if there is authentication data.
2006-05-23 11:36:26 +02:00
6e7d06ad34
rpcrt4: Handle pointers to context handles in the stubless code.
...
Implement check for NULL context handles if requested.
2006-05-20 17:37:36 +02:00
9c885c9185
rpcrt4: We don't need to free object pointers if we are told we must allocate.
2006-05-20 17:37:25 +02:00
a2767e8d71
rpcrt4: Use macros to generate the code for unmarshaling base types.
2006-05-20 17:37:13 +02:00
19fad8e710
rpcrt4: Don't set the memory pointer to NULL before calling PointerUnmarshall.
...
Call it with fMustAlloc set to TRUE instead to achieve the same result
without side-effects.
2006-05-20 17:36:42 +02:00
Ge van Geldorp