Author | Date | Description |
Robert Hastings | 2005-07-19 |
Added RegEnumKey
|
Adam Perry-Pelletier | 2005-07-20 |
Added documentation files to the jsegue website that were
artifacts created during the creation of the 1.4.1.111 release.
The changes-1.4.0.32.html file was somehow omitted from last
post-release check-is, so adding it now.
|
Robert Hastings | 2005-07-25 |
Limitted support for wrapping a java implementation in a native COM object. Currently only dispinterface wrapping is supported.
|
Adam Perry-Pelletier | 2005-07-29 |
Renamed new wrap method so that it conveys that the java object
gets wrapped by IDispatch on the native side.
|
build | 2005-07-30 |
Committed_by_Anthill
|
Robert Hastings | 2005-08-04 |
Added a small swing UI for managing java.util.logger.
|
Robert Hastings | 2005-08-04 |
Expose a getByteBuffer property that returns a byte buffer that
is a duplicate of the underlying byte buffer. It is positioned at 0.
|
Adam Perry-Pelletier | 2005-08-05 |
Reworked the example xml building code so when multiple filesets
are sent its way, only one example element gets injected into the examples.xml
file.
|
Adam Perry-Pelletier | 2005-08-05 |
Whitespace change in build.xml to trigger cruisecontrol build.
|
Robert Hastings | 2005-08-08 |
Added a fromString method to GUID to create a GUID from a string. Same as CLSIDFromString or IIDFromString
|
Robert Hastings | 2005-08-18 |
Fixed bad size computation
|
Adam Perry-Pelletier | 2005-09-14 |
The last release of jSegue was a 1.4.1.* so this new release will be a
1.4.2.* release.
|
Robert Hastings | 2005-09-20 |
Run with Java 5 as well as Java 4
|
Robert Hastings | 2005-09-20 |
Tesing w/VC++ .NET
|
Robert Hastings | 2005-09-30 |
Ignore a few more output files.
|
Robert Hastings | 2005-10-13 |
Sample vtbl code.
|
Robert Hastings | 2005-10-13 |
New native structure for BITMAP.
|
Robert Hastings | 2005-10-14 |
Added BITMAP
|
Robert Hastings | 2005-10-16 |
Tests for returning double, and float working
|
Robert Hastings | 2005-10-18 |
more testing, return float, double, structs all working
|
Adam Perry-Pelletier | 2005-10-18 |
Tlb2java now supports the notion of a parameter rewrite. So you can retype a parameter, and then provide the actual custom C++ conversion code that knows how to handle the new parameter type. Definitions for this code can occur in the default header file for the generated library. The command line args for the retype is given as follows: -retype_param IPictureDef.Bitmap.param.0 com.moesol.bindings.platform_sdk.windows_api.BITMAP 'BITMAPHolder a0(env, p0)'
|
Adam Perry-Pelletier | 2005-10-18 |
Added a BITMAPHolder that knows how to convert a bitmap as needed in a retype.
|
Adam Perry-Pelletier | 2005-10-18 |
Added constness to a few members of primitive_array_critical .
|
Adam Perry-Pelletier | 2005-10-18 |
Added a Java class BITMAP that corresponds to the MS BITMAP structure.
|
Adam Perry-Pelletier | 2005-10-18 |
Added the getObject method on HBITMAP to correspond to the ::GetObject windows_api function to retrieve that actual BITMAP from the HBITMAP handle.
|
Adam Perry-Pelletier | 2005-10-18 |
Generation differences from h-gen working against different JDK.
|
Adam Perry-Pelletier | 2005-10-19 |
Changed the BITMAP.set/get_bmBits to take a ByteBuffer vs the original byte []. This will facilitate retyping the return type.
|
Robert Hastings | 2005-10-20 |
Refactored the core of the interface thunking into jcom.h
|
Adam Perry-Pelletier | 2005-10-25 |
Added a way for users to retype a parameter that is also a IDL retval. Changed the BITMAPHolder to have cast operators and overloaded operators that allow it to serve as the return converter for a BITMAP. Fixed a memory leak created by storing ParamConversionCfg as a pointer type in a map - changed it to be a reference type and gave it a copy constructor.
|
Adam Perry-Pelletier | 2005-10-26 |
This commit adds a JUnit test for the new retyping of a parameter functionality of jSegue. The ComServer.idl now has a property RetypeLongToBitmap of type LONG. The LONG is retyped to a BITMAP which is carried through the binding on the java side. See com.moesol.tests.com_server.IDualTest.java to see its usage in the test. Make not of
-retype_param IDual.RetypeLongToBitmap.param.0 com.moesol.bindings.platform_sdk.windows_api.BITMAP BITMAPHolder
in the build.xml file. Also see BITMAPHolder to see how that actual conversion is implemented on the C++ side.
|
Adam Perry-Pelletier | 2005-10-26 |
Removed a dead include file.
|
Adam Perry-Pelletier | 2005-10-26 |
Replace jniRegistrationStr with getSigClassName as the latter already existed and does the same as the former.
|
Adam Perry-Pelletier | 2005-10-26 |
Added hashCode and equals methods on BITMAP that reflect that it has a separate field of type ByteBuffer to hold the actual bitmap bits. A unit test exercises these methods when either the stucture bytes change or the bits ByteBuffer change.
|
Adam Perry-Pelletier | 2005-10-26 |
instanceof takes into account object nullness so I removed the explicit null check in equals.
|
Robert Hastings | 2005-10-28 |
Added checkBitsForHandle
|
Robert Hastings | 2005-10-28 |
Fixed hashCode when m_bits == null
|
Robert Hastings | 2005-10-28 |
The check for ensuring the cached class for cached method ID's was too strict. It was reloading the cache if the value of cached_class and class did not match, but it is safe to NOT reload if they are the same object with different refs. Fixed!
|
Robert Hastings | 2005-10-28 |
Added support to tlb2java for passing simple structures as [in] [out] and [in,out] parameters.
|
Robert Hastings | 2005-10-29 |
Make -disp_only compile for records.
|
Robert Hastings | 2005-10-30 |
Made the version-compile target just compile com/moesol/util/*Version*.java files so that if there are source files that are generated but they have errors in them the bulid still will compile.
|
Robert Hastings | 2005-10-30 |
com.moesol.util.*Version*.java depend on com.moesol.generator.
|
Robert Hastings | 2005-10-31 |
Fixed assertion in unit test. Was a bug in the unit test.
|
Robert Hastings | 2005-11-02 |
Added NULL so that you can write
GUID.NULL.equals(x)
|
Robert Hastings | 2005-11-02 |
Added VT_RECORD support
Refactored.
Added try catch of _com_error to jni_Invoke_impl
|
Robert Hastings | 2005-11-02 |
Added VT_RECORD support
|
Robert Hastings | 2005-11-02 |
Added VT_RECORD support for -disp_only.
|
Robert Hastings | 2005-11-02 |
Added VT_RECORD support for -disp_only.
Additionally, added special case handling for STRUCT * so that in
java that is the same as STRUCT.
|
Adam Perry-Pelletier | 2005-11-02 |
Updated the comment to reflect the function actually called.
|
Adam Perry-Pelletier | 2005-11-02 |
Manually created the 1.5.0.176 version for bundled release with CJB.
|
Robert Hastings | 2005-11-02 |
A little more javadoc for TYPELIB fields.
|
Robert Hastings | 2005-11-03 |
Start of support for fields that are BSTR, VARIANT, SAFEARRAY, or interfaces.
|
Robert Hastings | 2005-11-03 |
Start of support for fields that are BSTR, VARIANT, SAFEARRAY, or interfaces.
Added some missing put/get's.
|
Robert Hastings | 2005-11-04 |
Added structure with BSTR, VARIANT, SAFEARRAY, and interface
|
Robert Hastings | 2005-11-04 |
Added VARIANT support.
|
Robert Hastings | 2005-11-04 |
Fixed test, new interface methods.
|
Robert Hastings | 2005-11-04 |
Fixed broken build.
|
Robert Hastings | 2005-11-04 |
Added javadoc.
|
Robert Hastings | 2005-11-04 |
Fixed broken unit test (partial commit allowed some vtbl stuff to get commited early).
|
Adam Perry-Pelletier | 2005-11-04 |
Moved BITMAPHolder down into the windows_api namespace and directory. Reworked the export declaration of BITMAPHolder to get rid of compiler warnings introduced as class member vars were exported.
|
Robert Hastings | 2005-11-11 |
Moved
|
Robert Hastings | 2005-11-11 |
Added accessors for type info and TYPEATTR * to the TypeInfoHolder
|
Robert Hastings | 2005-11-11 |
Made VT_DISPATCH and VT_UNKNOWN generate correct code when in records.
|
Robert Hastings | 2005-11-11 |
Moved toString into stdafx.h so that it is shared by all.
|
Robert Hastings | 2005-11-11 |
Added more information to the xml output, such as the funcdesc so that we can see in xml the vtble offset, etc.
|
Robert Hastings | 2005-11-11 |
Refactored the DispatchJavaMethod code so that DispatchJavaStructure can share code with it.
|
Robert Hastings | 2005-11-11 |
Created new jni_copy_bytes method that can copy bytes from a NativeStructure or a ByteBuffer. It does the "right" thing if the underlying ByteBuffer is a HeapByteBuffer or a DirectByteBuffer.
Added new vtbl thunking. This is still a work in progress.
Fixed up some of the record handling to use malloc and free instead of IRecordInfo::RecordCreate and RecordDestroy. We use malloc and free because we pass owner ship of the nested fields to the Java side.
|
Robert Hastings | 2005-11-11 |
Start of implementing the native code to handle VT_RECORD. [in] and [out] for BSTR, VARIANT, and interfaces are working, but [in,out] is not.
|
Robert Hastings | 2005-11-11 |
Added missing trait methods for call_method_a
|
Robert Hastings | 2005-11-11 |
VT_RECORD tests are not ready for primetime!
|
Robert Hastings | 2005-11-11 |
Progress w/reverse COM.
We removed all of the C++ type information from addEntry, so now we look it up from the LIBID and IID. That gets us the ITypeInfo. We then try to drill down into the [dual] side of the interface. Once we do that we have to recurse through all of our base interfaces and have them fill in the entry info that we need. After we visit all bases and the type info for the interface we assert that we have updated all m_entries.
JUnit passes for one call. More type support forth coming.
|
Adam Perry-Pelletier | 2005-11-11 |
Added IRM_EXPORT to the methods rather than the class to fix compiler
warnings associated with exporting class member variables.
|
Robert Hastings | 2005-11-11 |
Incremental step toward reverse COM.
Replaced the hand coded conversion of int with table driven model.
Refactored some of the files, pulled some of the code out of jcom.h into a new jni_ComInterfaceThunk.h
|
Robert Hastings | 2005-11-12 |
Fixed problems with second runs of tests.
|
Robert Hastings | 2005-11-12 |
Move unit tests.
|
Robert Hastings | 2005-11-12 |
Moved auto_delete_vector to jcom.h to share w/jni_ComInterfaceThunk
|
Robert Hastings | 2005-11-12 |
Added std:: so that jcom.h will compile even if no using namespace std has been seen.
|
Robert Hastings | 2005-11-12 |
Refactored to share jni_convert_value. Table driven system for creating parameter handlers is working. Many more types should be working. JUnit tests for INT, BYTE, DWORD, and BSTR (sans as an out parameter) are working
|
Robert Hastings | 2005-11-12 |
Moved DumpBytes class from Cga and extended it to support dumping ByteBuffer data (heap or direct).
Added some JUnit tests.
|
Robert Hastings | 2005-11-12 |
Added toString to help w/debugging on the java side.
Started a small unit test
|
Robert Hastings | 2005-11-12 |
Added GEN_CONST_NAME to help with generating the mapping of VT_FOO to the string "VT_FOO" on the java side.
|
Robert Hastings | 2005-11-12 |
Added support for getting number of outstanding interfaces
|
Robert Hastings | 2005-11-12 |
Added code for computing the size of each argument on the stack.
Cleaned up some minor todo's.
|
Robert Hastings | 2005-11-12 |
Added missing BSTR/jobject conversion, refactoed all of the BSTR conversions to just call the BSTR/jstring conversion.
|
Robert Hastings | 2005-11-12 |
Refactored vtable tests into their own ComInterfaceThunkTest class.
|
Robert Hastings | 2005-11-12 |
Added support for passing in and out a VARIANT of VT_ERROR. Before we would throw unsupported variant type. Added unit test for it on disp only side.
|
Robert Hastings | 2005-11-14 |
Javadoc
|
Robert Hastings | 2005-11-14 |
added getIID
|
Robert Hastings | 2005-11-14 |
Beginning of query interface implementation
|
Robert Hastings | 2005-11-15 |
h-gen
|
Robert Hastings | 2005-11-15 |
Partially working "reverse COM" QueryInterface. The main feature missing at this time is support for always returning the same interface pointer for IUnknown.
|
Robert Hastings | 2005-11-15 |
Partially working "reverse COM" QueryInterface. The main feature missing at this time is support for always returning the same interface pointer for IUnknown.
Added missing singleton accessor instance() to IUnknown and IDispatch
|
Robert Hastings | 2005-11-15 |
Skeleton in proc server that will eventually host java COM components
|
Robert Hastings | 2005-11-15 |
Include dependent headers, make in_proc depend on bindings
|
Robert Hastings | 2005-11-15 |
Refactored, extracted IUnknown::detach
|
Robert Hastings | 2005-11-15 |
Start of dllGetClassObject
|
Robert Hastings | 2005-11-15 |
Start of thread local to store/access the raw HRESULT
|
Robert Hastings | 2005-11-15 |
fixed build error in in_proc
|
Adam Perry-Pelletier | 2005-11-15 |
Initial commit of the new com_moesol_java_launch project which builds the com_moesol_java_launch.dll. This project extracts the java launching capabilities used by JavaMapInjector for use by other applications, notably the C2PC Gateway Java CGI functionality.
|
Robert Hastings | 2005-11-15 |
Copy dependent dll, com_moesol_bindings
|
Robert Hastings | 2005-11-15 |
Do not post a dialog box when an unhandled exception occurs. Instead call back to COM.uncaughtException and let it map (if it can) and log the exception.
|
Robert Hastings | 2005-11-15 |
IUnknown_00024::detach was broken, fixed
|
Robert Hastings | 2005-11-16 |
Updated CLSCTX enums to match latest Microsoft docs.
|
Adam Perry-Pelletier | 2005-11-16 |
Initial commit of the JavaLauncher application that will launch a Java class's main method as specified in a configuration properties file.
|
Adam Perry-Pelletier | 2005-11-16 |
Fixed Release build configuration.
|
Robert Hastings | 2005-11-16 |
Better error reporting. Log any java exception even COMException. Use COM::uncaughtException to map java exception to HRESULT (including COMException).
|
Robert Hastings | 2005-11-16 |
Fixed memory leak if the thunk will not be used because an error is thrown, delete it
|
Robert Hastings | 2005-11-16 |
Implemented dllGetClassObject.
|
Robert Hastings | 2005-11-16 |
ComClassFactory is a Java implementation of IClassFactory. Since a java Class objects can act as a class factory ComClassFactory simply wraps IClassFactory around a Class object.
|
Robert Hastings | 2005-11-16 |
Brought over all COM error codes.
Added HRESULT_FROM_NT and HRESULT_FROM_WIN32 from MSDN documentation.
|
Robert Hastings | 2005-11-16 |
Start of thunking for IDispatch
|
Robert Hastings | 2005-11-16 |
Added support for custom thunk methods to handle IUnknown, IDispatch, and IClassFactory. In general, the custom thunk can be used for any interface that does not have a typelib or that you need to override.
|
Robert Hastings | 2005-11-16 |
Fixed warning by making some private methods that get called from JNI default
|
Robert Hastings | 2005-11-16 |
Bindings for IClassFactory, forward and reverse COM.
|
Robert Hastings | 2005-11-16 |
comment
|
Robert Hastings | 2005-11-16 |
Implement QueryInterface vi COM.providedQueryInterface
|
Robert Hastings | 2005-11-16 |
Moved LastHRESULT from tests to src</dn
|
Robert Hastings | 2005-11-16 |
Add static accessor
|
Robert Hastings | 2005-11-16 |
Oops, backout testDllGetClassObject since its unit test is not quite ready for prime-time
|
Robert Hastings | 2005-11-17 |
Added registerServer and unregisterServer to COM.
|
Robert Hastings | 2005-11-17 |
Removed commented out code
|
Robert Hastings | 2005-11-17 |
Added Close helper method.
|
Robert Hastings | 2005-11-17 |
Deprecated RegSetKeyEx, it should have been named RegSetValueEx
|
Robert Hastings | 2005-11-17 |
Fixed getBindingsPath
|
Robert Hastings | 2005-11-17 |
Fixed testDllGetClassObject to work on any maching by registering
MyVtableServer. Also, note that this unit test unregisters MyVtableServer
when the test finishes.
|
Adam Perry-Pelletier | 2005-11-17 |
Heavy refactoring in prep for final logging of errors. Added a way to set a debug flag in the properties file. Created a method that adds the executable module's name to pass down to Java.
|
Robert Hastings | 2005-11-18 |
Renamed deprecated method
|
Robert Hastings | 2005-11-18 |
Moved in proc server into com_moesol_bindings
|
Robert Hastings | 2005-11-18 |
Added comment
|
Robert Hastings | 2005-11-18 |
Added a sample C++ program to create the java component and use reverse COM to access it.
|
Robert Hastings | 2005-11-18 |
Lowered log level from INFO/WARNING to FINE because these exceptions are getting converted into HRESULTs which then get converted back to reasonable messages. Also, QI is going to fail for any interface the object does not implemnt and this is OK. To see more detail about an exception thrown from a Java COM server you just turn up the log level for com.moesol.bindings.platform_sdk.component_services.COM.
|
Robert Hastings | 2005-11-18 |
Added missing module .def file
|
Robert Hastings | 2005-11-18 |
Refactored, and added a CoCreateInstance test. This test is currently not being run (XXXtest previx) because the start_vm code in DllGetClassObject needs to dynamically build the CLASSPATH
|
Adam Perry-Pelletier | 2005-11-19 |
The JavaLauncher now send error information to the Windows Event Log. Error reporting was improved across the application. Custom messages resources were added so that the event log has nice error output.
|
Adam Perry-Pelletier | 2005-11-19 |
Change from Binary to ASCII.
|
Adam Perry-Pelletier | 2005-11-23 |
Added better logging information and caught specific exceptions to drive specific logging.
|
Robert Hastings | 2005-11-23 |
Added fullpath to devenv.com launch.
|
Robert Hastings | 2005-11-23 |
Made calling vtable methods with IUnknown and IDispatch work
|
Karl Schricker | 2005-11-24 |
Added more instances of ${devenv.home}
|
Adam Perry-Pelletier | 2005-11-24 |
Moved Java.cpp and Java.h into com_moesol_bindings so that we don't have to ship an extra dll. However, due to heap allocations across build configurations, the Java.cpp and Java.h are not actually built into the com_moesol_bindings.dll but are included in projects that build other executables.
|
Adam Perry-Pelletier | 2005-11-24 |
Got rid of the last vestiges of com_moesol_java_launch.
|
Robert Hastings | 2005-11-24 |
Ignore Debug directory
|
Robert Hastings | 2005-11-24 |
Refactored to extract a jni_Param class
|
Robert Hastings | 2005-11-24 |
Partial handling of COCLASS and INTERFACE user defined types
|
Robert Hastings | 2005-11-24 |
fixed build error in projects that do not use the std namespace.
|
jlcos | 2005-11-26 |
Configure Eclipse to use an external Ant builder
|
Robert Hastings | 2005-11-26 |
"Reverse COM" now works for all interface pointer types, the two special cases VT_UNKNOWN, VT_DISPATCH, and the three VT_PTR,VT_USERDEFINED cases TKIND_DISPATCH, TKIND_INTERFACE, and TKIND_COCLASS.
Created two classes derived from param_handler_info_impl<VT_UNKNOWN>, coclass_handler_info_impl, and interface_handler_info_impl.
|
jlcos | 2005-11-26 |
Disable ant builder for incremental Eclipse builds
|
Robert Hastings | 2005-11-28 |
Partial start of managing IUnknown identity
|
Robert Hastings | 2005-11-28 |
Deleted tmp file, breaks build
|
Robert Hastings | 2005-11-28 |
Fixed broken build.
|
Adam Perry-Pelletier | 2005-11-29 |
Prepared JavaLauncher for inclusion into CJB's MSI. Added a trivial simple app that pops up a dialog in java to demonstrate simple usage of JavaLauncher.
|
Adam Perry-Pelletier | 2005-11-29 |
Pulled a JavaLauncher.properties file up from jSegue to here. Although seemingly generic, this file is specific for the installed location of CJB and not jSeuge, so moved it out of jSegue. jSegue now has a similar file that works for local directory.
|
Adam Perry-Pelletier | 2005-11-29 |
Fixed a bug where java exceptions weren't getting posted to the event log properly. The ostringstream didn't know how to convert ::java::lang::String so it only put the address of it out to the event log. Now the full message appears as the ::java::lang::String was converted to a const char *.
|
Robert Hastings | 2005-11-29 |
Fixed eclipse warning
|
Robert Hastings | 2005-11-29 |
For reverse COM this implements the COM requirement that when you query to IUnknown from any interface to the COM server you must get back the same pointer so that clients can compare for equality.
|
Robert Hastings | 2005-11-29 |
Renamed ComCoClass to ComObjectRoot since it functions more like the ATL CComObjectRoot.
|
Robert Hastings | 2005-11-29 |
Freeze the COM map on the first QueryInterface. This is another COM requirement of IUnknown::QueryInterface, quote:
The set of interfaces accessible on an object through IUnknown::QueryInterface must be static, not dynamic. This means that if a call to QueryInterface for a pointer to a specified interface succeeds the first time, it must succeed again, and if it fails the first time, it must fail on all subsequent queries.
|
Robert Hastings | 2005-11-29 |
Test that FinalRelease is called during server teardown.
|
Robert Hastings | 2005-11-29 |
Added one more assert and comments to indicate what might
be happening
|
Robert Hastings | 2005-11-29 |
Make sure we work in the face of "bad clients" that free com servers via the wrong interfaces
|
Robert Hastings | 2005-11-29 |
Added a main method so we can quickly register this server for develmental testing
|
Robert Hastings | 2005-11-29 |
Updated how reverse COM handles COM interface reference counts. The goal of this update is to make it so that when the server implements a COM interface the resulting implementation can be used from native COM clients or directly from Java and still have the same reference count behavior. Take for example, calling
IUnknown result = idual.getPropIUnknown();
From the forward COM perspective the caller must later call result.Release(), (actually the GC will eventually call this method, but this discussion is sans what help the GC gives). So, when java code directly calls idual.getPropIUnknown without and jSegue layers we need to have that method create a new COM reference (via QueryInterface) that the client can call Release on without releasing the interface the server holds. So we had to update reverse COM to match this expectation on the server.
Now both methods work as expected.
|
Robert Hastings | 2005-11-30 |
Moved DumpBytes from the ui package to the util package.
|
Robert Hastings | 2005-11-30 |
Added support for using inner classes to implement interfaces.
|
Robert Hastings | 2005-11-30 |
Added support for registering COM component categories.
|
Robert Hastings | 2005-11-30 |
Re-enabled all these unit tests
|
Robert Hastings | 2005-11-30 |
More inner implementation testing. This tests that you can create a top-level class that implements an interface and then add it to the ComObjectRoot and things still work.
|
Robert Hastings | 2005-11-30 |
Relaxed assert for methods with zero args. In these cases is makes sense to pass NULL for the args parameter.
|
Robert Hastings | 2005-11-30 |
Added another unit test to test that we can call inner/inner IDual2 methods.
|
Robert Hastings | 2005-11-30 |
Start of COM aggregation support
|
Robert Hastings | 2005-12-01 |
Javadoc warnings
|
Robert Hastings | 2005-12-01 |
When reviewing code for aggregation support I realized that calling QueryInterface,AddRef, or Release through an interface that was implemented with InnerInterface would cause a ClassCastException. This commit fixes that bug.
|
Robert Hastings | 2005-12-02 |
Partial index property support for mx build which now needs this for the tms mock
|
Robert Hastings | 2005-12-02 |
Correctly build output java arrary for array types that contain object elements.
|
Robert Hastings | 2005-12-02 |
Fixed bug in out handling of COM interface pointers.
|
Robert Hastings | 2005-12-02 |
Renamed ComObjectRoot to ComObject, merged the functionaltiy of InnerInterface with ComObject.
Added support for Java aggregation which is very similar to COM aggregation.
and will be the basis of the COM aggregation support
|
Robert Hastings | 2005-12-02 |
Missing h-gen files
|
Robert Hastings | 2005-12-03 |
Updated ReverseCOM diagrams
|
Robert Hastings | 2005-12-03 |
Fixed warning
|
Robert Hastings | 2005-12-03 |
Refactored to catch array, reduces one JNI call
|
Robert Hastings | 2005-12-03 |
Fixed bug with [in, out] on COM interfaces, was causing VM crash, includes unit test
|
Robert Hastings | 2005-12-03 |
ByteBuffer memcpy support, still need some refactoring with this
|
Robert Hastings | 2005-12-03 |
Start of COM aggregation support.
Renamed some support classes. Have common architecture for Java
and COM aggregation. Have basic graph setup.
|
Robert Hastings | 2005-12-05 |
Added more to the unit test for IUnknownAsInOut
|
Robert Hastings | 2005-12-05 |
added missing line in aggregation graph
|
Robert Hastings | 2005-12-06 |
Cleaned up the option handling, added abort and exit handlers.
|
Robert Hastings | 2005-12-06 |
Expose the COM map for unit testing and modification.
Switched to using an unmodifiable hash map when the COM map
is frozen.
|
Robert Hastings | 2005-12-06 |
Completed "inner" aggregation testing. Java COM object servers can now be used as aggregates of other (even native) COM object servers.
|
Robert Hastings | 2005-12-06 |
Added testing of [in, out] IUnknown **
|
Robert Hastings | 2005-12-06 |
Added a todo unit test
|
Robert Hastings | 2005-12-06 |
Fixed broken unit test, forgot to commit this change
|
Adam Perry-Pelletier | 2005-12-06 |
Commit of resource only .dll and the event logging for com_moesol_bindings.
|
Adam Perry-Pelletier | 2005-12-06 |
Fixed broken release build.
|
Adam Perry-Pelletier | 2005-12-06 |
Refactored the event logging mechanism into a class named EventLog so other areas of code relying on MessageBox could be transitioned to using EventLog.
|
Robert Hastings | 2005-12-06 |
Refectored ByteBuffer handling
|
Adam Perry-Pelletier | 2005-12-06 |
The functor now throws an exception rather than ASSERTs when the underlying funtor member is NULL. Made com_moesol_bindings rethrow lower-leverl errors/exceptions. Removed generated files from the build.
|
Adam Perry-Pelletier | 2005-12-06 |
Added back deleted files. These are actually needed as com_moesol_bindings needs them.
|
Robert Hastings | 2005-12-06 |
Fixed bug with byte buffer refactoring, forgot to init m_direct to
NULL when the byte buffer is not a direct byte buffer.
|
Robert Hastings | 2005-12-07 |
Fixed a hard-coded path in the event registration.
Bound the event application name to the dll name to avoid conflicts
with hosting application logging.
|
Robert Hastings | 2005-12-07 |
Made JavaLauncher work with '.'s in the class name as well as with '/'.
So
foo.bar.MainClass now works
when before only
foo/bar/MainClass would work.
|
Adam Perry-Pelletier | 2005-12-08 |
Header files are ASCII not binary.
|
Adam Perry-Pelletier | 2005-12-08 |
Got rid of onerous event logging at the INFO level that was polluting the Windows Event Log.
|
Adam Perry-Pelletier | 2005-12-08 |
Removed a dead file from the repository and the build. This is a resource-only dll.
|
Adam Perry-Pelletier | 2005-12-08 |
Exported EventLog functionalrity for use in other dlls and exes.
|
Adam Perry-Pelletier | 2005-12-08 |
Now that EventLog is refactored and exported from com_moesol_bindings.dll, the custom logging that was used in JavaLauncher has been eliminated in favor of that in EventLog in com_moesol_bindings.dll.
|
Adam Perry-Pelletier | 2005-12-08 |
Fixed broken build due to including a deleted header file.
|
Adam Perry-Pelletier | 2005-12-08 |
Refactoring: decompose conditional statement for better clarity.
|
jlcos | 2005-12-09 |
Disable Eclipse ant builder until resource problems can be resolved.
|
Adam Perry-Pelletier | 2005-12-09 |
Decomposed conditional logic to a helper method. Created maybeDeReferncePtr to consolidate duplicative code.
|
Robert Hastings | 2005-12-10 |
Added unit test to see what happens when an event handler throws an exception
|
Robert Hastings | 2005-12-10 |
Fixed JUnit, forgot remove listener
|
derek | 2005-12-13 |
Added CoCreateGUID to COM.java. COM class now provides a unique GUID creation capability.
JUnit Test added t TestCOM.java
|
Robert Hastings | 2005-12-13 |
Removed empty file
|
Robert Hastings | 2005-12-13 |
Added code to create a class path builder from preferences setup
in the windows registry
|
Robert Hastings | 2005-12-14 |
Replaced emacs last edit time with cvs Id
|
derek | 2005-12-14 |
Updated name of CoCreateGUID method to CoCreateGuid
|
Robert Hastings | 2005-12-14 |
Finished reverse COM documentation.
|
Robert Hastings | 2005-12-14 |
Added support for com_inproc_groups
|
Robert Hastings | 2005-12-14 |
Reverse COM support for setting the system classpath from registry entries.
|
Adam Perry-Pelletier | 2005-12-14 |
Implemented code generation path for the VT_CARRAY c-style array type. Currently this code path only supports char arrays. So one can use idl notation
[id(78)] HRESULT InCArrayType([in] char pVal[20]);
[id(79)] HRESULT OutCArrayType([out] char pVal[20]);
|
Adam Perry-Pelletier | 2005-12-15 |
Made the memory allocation/manipulation calls a little bit more robust and better able to handle future mods where the array element size may change.
|
Adam Perry-Pelletier | 2005-12-15 |
The com_moesol_event_log project regenerated the messages file every time it built. This caused CVS to show these files as changed on every commit event when the file did not actually change. The generation of the message definitions has been removed from the visual studio project and moved to the ant script for generation only when it actually changes.
|
Robert Hastings | 2005-12-16 |
Moved UnsignedByteBuffer, UnsignedLongBuffer, and UnsignedShortBuffer into jSegue so that jSegue can use these classes.
|
Robert Hastings | 2005-12-16 |
Added a junit test for UnsignedByteBuffer
|
Robert Hastings | 2005-12-16 |
Added better support to SAFEARRAY conversions for VARIANTs. You can now use an UnsignedByteBuffer, UnsignedLongBuffer, or UnsignedShortBuffer as the object in setArray or makeArray. This is more convenient for these unsigned types than creating an array of the promoted java type (short, int, long).
|
Robert Hastings | 2005-12-16 |
Fixed field access when the field is a jbyteArray. Needed to static_cast the jobject to the return type
|
Robert Hastings | 2005-12-16 |
Fixed jni_bb_mempcy family of functions to work with read-only non-direct byte buffers. Also, read-only direct byte buffers throw a read-only exception if they are used as the destination of a jni_bb_memcpy.
|
Robert Hastings | 2005-12-17 |
Handle more types in Reverse COM: Typedefs and enums.
|
Robert Hastings | 2005-12-17 |
Made multiple args work with Reverse COM
|
Robert Hastings | 2005-12-17 |
Added is bad write pointer checking to reverse com in,out handler.
|
Robert Hastings | 2005-12-20 |
Fix warning
|
Robert Hastings | 2005-12-22 |
Added root aliases for HKEY from PlatformSDK to HKEY.
|
Robert Hastings | 2005-12-22 |
Deprecated simpleInvokeAndWait, since safeInvokeAndWait is better.
|
Robert Hastings | 2005-12-22 |
Fixed generic error message.
|
Robert Hastings | 2005-12-22 |
Fixed possible uncaught exception that could happen if the java code to convert the uncaught java exception threw.
|
Robert Hastings | 2005-12-22 |
Added support for registering the ThreadingModel when registering Java COM objects. The default ThreadingModel is Apartment.
Refactored the registry support out of COM and into a new class, ComRegistryHelper.
|
Robert Hastings | 2005-12-23 |
Log stack trace for Java exception in windows Event Log.
|
Robert Hastings | 2005-12-23 |
Added some convenience methods to HKEY to mirror what we do for HMENU, etc.
|
Adam Perry-Pelletier | 2006-01-03 |
When registering a Java COM object, it may or may not utilize the implemented categories features of COM. This commit allows registration to continue if a COM object doesn't use implemented categories.
|
Adam Perry-Pelletier | 2006-01-05 |
Fixed the build so it can locate the message compiler mc.exe without user modification of the 'path' environmental variable.
|
Adam Perry-Pelletier | 2006-01-19 |
Added documentation to the jSegue website that details how to configure JavaLauncher.
|
Adam Perry-Pelletier | 2006-01-19 |
Got rid of reference to irm in the jSegue example JavaLauncher.properties file.
|
Adam Perry-Pelletier | 2006-01-19 |
Added some user comments to the example section.
|
Robert Hastings | 2006-01-19 |
Fixed bug CJB60 by simplifying the factory_functor and new_instance_factory. Now we pass the factory_functor as a pointer from the top down. There is a new not_impl_factory_functor and a function_pointer_factory_functor that handle all of the non-reverse-com needs from before
|
Adam Perry-Pelletier | 2006-01-24 |
Fixed grammar error in docs.
|
Robert Hastings | 2006-02-01 |
Added a fine level log message so that you can know where
we are looking for the dev properties file
|
Robert Hastings | 2006-02-01 |
A few more error constants, we really need to add them all
|
Robert Hastings | 2006-02-01 |
Updated the event log message to include one possible cause,
the dreaded launch with remote debugging set to a port that is already
in use.
|
Robert Hastings | 2006-02-01 |
Possible VM bug. Pointing an object to another reference that is statically created in another class failed.
|
Robert Hastings | 2006-02-01 |
Allow deletes using a Value structure.
Don't fail if we unregister but a key is already gone
|
Robert Hastings | 2006-02-01 |
Find an already running Java VM when possible
|
Adam Perry-Pelletier | 2006-02-01 |
Added extra logging to print out what classpaths are appended. Also added windows event logging when "register_natives" fails.
|