Author | Date | Description |
Robert Hastings | 2005-03-02 |
Committed C++Builder project files.
Added a C++Builder test program with a main entry point.
|
Robert Hastings | 2005-03-04 |
During mock objects transformations I discovered that IDispatch and IUnknown were off by one
level of indirection, this fixes that
|
Robert Hastings | 2005-03-18 |
Ignore word temp files
|
Robert Hastings | 2005-04-01 |
When using -disp_only we still generate java code to use JNI when we run into an interface that does NOT
derive from IDispatch. We mostly do this so that the java code will still compile.
|
Adam Perry-Pelletier | 2005-04-01 |
A common programming error when using OleThread is users don't call
startAndInit() before invoking a thread operation through the various
invoke calls. This commits adds code that will throw an IllegalStateException
if one of the invoke methods has been called before startAndInit has been
called.
|
Robert Hastings | 2005-04-08 |
Workaround buggy type libraries that are missing [source] interface markers on their coclass'es.
New public static COM.addListener/removeListener methods let you add/remove listeners if you know the DIID.
Also, fixed a bug where the listener cookie map might not be found (e.g. two java instances for the same coclass).
Refactored so that the cookie managment happens in IUnknown using helper classes.
|
Robert Hastings | 2005-04-08 |
Had to make MoeSolInternal public
|
Robert Hastings | 2005-04-08 |
Fixed bug in code that tries to find IFace.Jni from IFace. IUnknown.Jni.MoeSolInternal was being found
|
Robert Hastings | 2005-04-08 |
h-gen
|
Robert Hastings | 2005-04-14 |
For getting the version information from a cvs tag we were missing a way to set the cvs.tag.prefix
|
Robert Hastings | 2005-04-14 |
When we have trouble with the writing of version properties throw a runtime error with
the real exception as the cause so that we get better error diags.
|
Robert Hastings | 2005-04-14 |
Reworked the version logic so that we use the version defaults if we cannot find a version
entry that matches what we are looking for.
Better error diags.
|
Robert Hastings | 2005-04-14 |
Refactored to share toString template function
|
Robert Hastings | 2005-04-14 |
Added support for structures to idl to xml.
|
Robert Hastings | 2005-04-14 |
Fixed regression. Also, fixed the fact that sometimes the time would change causes this
unit test to fail
|
Robert Hastings | 2005-04-19 |
Added a manifest with version info, <issue id="CJB15">
|
Robert Hastings | 2005-04-19 |
<issue id="CJB15"> this test code prints out how java sees the info in the manifest file.
|
Robert Hastings | 2005-04-20 |
Expanded tlb2java to generate java wrapper code for structures defined in a typelibrary.
|
Robert Hastings | 2005-04-20 |
Added new files
|
Robert Hastings | 2005-04-21 |
Added support for accessing arrays of native structures that are nested in native structures.
When getting, no bytes are copied, structure wrapper classes are just bound to the correct byte
offsets in the byte buffers.
Refactored: pulled up equals and hashCode from GUID to NativeStructure.
|
Robert Hastings | 2005-04-21 |
Generate copyless code for arrays of structures.
Use length_is() attribute when supplied with the -missing option.
Error if a field comes after a variable sized array
Refactor so that NativeStructDTP is simpler and shared more code
|
Robert Hastings | 2005-04-22 |
Updated structure support to handle variable sized arrays.
|
Robert Hastings | 2005-04-22 |
Fixed bug in replacer it was dropping the char after the $
|
Robert Hastings | 2005-04-22 |
Fixed bug
|
Robert Hastings | 2005-04-22 |
Fixed, was generating invalid xml
|
Robert Hastings | 2005-05-06 |
Fixed a bug in the computation of the dynamic size
|
Robert Hastings | 2005-05-11 |
Correctly compute getDynamicSize when the last field of a structure is itself a dynamically resizable structure.
note that unions cannot be dynamic.
|
Adam Perry-Pelletier | 2005-05-13 |
Properly detached the the SAFEARRAYs that are elements in the returned
array to java so that they are not deleted out from underneath us. This
fixes the problem where getting of the attribute TRK_RTN caused a JVM
crash.
|
Robert Hastings | 2005-05-20 |
Pass length_is attribute through to the xml
|
Robert Hastings | 2005-05-20 |
Fixed bug. The length_is attribute was not being used for primitive arrays.
|
Robert Hastings | 2005-05-20 |
Emit the offset of each field in the structure as a constant for use outside of the get/set.
|
Robert Hastings | 2005-05-25 |
Added new method for structures that have variable sized arrays.
getDynamicSize(count) computes what the dynamic size would be for count elements in the variable sized array.
|
Robert Hastings | 2005-05-28 |
New factory method for creating structs with variable sized arrays, does not nest.
|
Robert Hastings | 2005-06-04 |
Found and fixed two bugs when dealing with unsigned values. For unsigned byte/int 0xFF or 0xFFFFFFFF would
be returned as -1.
Added put/getUnsignedShortArray.
|
Robert Hastings | 2005-06-04 |
Fixed bug in test.
|
Robert Hastings | 2005-06-04 |
Fixed regression. Called clear, but that is not right because that sets limit to the capacity.
Ended up calling position(0) to ensure that we have all our bytes remaining. Otherwise, we were
sometimes hashing over nothing. Also, just use the byte buffer hashCode method.
|
Robert Hastings | 2005-06-04 |
Added support for -missing <type>.<field> [string]. When tlb2java
sees this for a field it assumes that the field is a C style string
and creates a getter and setter that use the java String type. The old
getter and setter are available as methods prefixed with underscore.
|
Robert Hastings | 2005-06-04 |
Fixed a bug in getString(offset, length).
Also, made getSlice return a "tighter" byte buffer.
The old form did not set the capacity of the returned buffer.
The new form sets the limit and capacity.
|
Robert Hastings | 2005-06-06 |
By using dup before creating the slice we make sure that the original
position and limit are not modified.
|
Robert Hastings | 2005-06-06 |
By using dup before creating the slice we make sure that the original
position and limit are not modified.
Also, went back to not reducing the capacity in the slice. Need to verify
no code takes advange of the fact that a slice's capacity may be larger
than the limit.
|
Robert Hastings | 2005-06-06 |
fixed typo
|
Robert Hastings | 2005-06-06 |
Fixed missing getDynamicSize on array fields.
|
Robert Hastings | 2005-06-07 |
Added a comment about long being a C/C++ 4 byte long */
|
Robert Hastings | 2005-06-09 |
Update equals and hashCode. Fast return if this == other. Dup the byte buffer and set position to 0 before
calling bb.equals or bb.hashCode.
|
Robert Hastings | 2005-06-09 |
Added a check for other == null in equals to avoid a NullPointerException.
|
Robert Hastings | 2005-06-10 |
Added -xjavadoc support to tlb2java. -xjavadoc causes tlb2java to look for
package.xml in the package output directory. If not found it emits an error
and exists. If found it merges the text content found into the output javadoc
comments.
|
Robert Hastings | 2005-06-10 |
Refactored to use new getContainingTypeLib.
|
Robert Hastings | 2005-06-10 |
Refactored to use computeOutputFilename
|
Robert Hastings | 2005-06-10 |
Added a dispinterface with properties (exposes a place where tlb2java outputs just comments).
|
Robert Hastings | 2005-06-10 |
ws
|
Robert Hastings | 2005-06-10 |
Fixed bug in arg parsing
|
Robert Hastings | 2005-06-10 |
Better error message when we cannot open a package.xml file.
|
Robert Hastings | 2005-06-10 |
Since javadoc and tlb2java -xjavadoc look for files in the package directory we
can no longer delete the entire folder during clean. Instead we delete all the files
except *.xml and *.html (ant also skips the .cvsignore file).
|
Robert Hastings | 2005-06-10 |
better test docs
|
Robert Hastings | 2005-06-10 |
Added getHelpContext and getHelpFile.
Protect getSuper for IUnknown case.
|
Robert Hastings | 2005-06-10 |
Process dispatch, alias, and module.
|
Robert Hastings | 2005-06-10 |
Process dispatch, alias, and module.
Added some missing attributes.
|
Robert Hastings | 2005-06-11 |
ignore generated files
|
Robert Hastings | 2005-06-13 |
Updated javadoc
|
Robert Hastings | 2005-06-13 |
Remove usused local
|
Robert Hastings | 2005-06-13 |
Use -xjavadoc for all tlb generations.
|
Robert Hastings | 2005-06-13 |
Started fleshing out the xjavadoc documentation.
|
Robert Hastings | 2005-06-13 |
Implement -xjavadoc for all java code generation.
|
Robert Hastings | 2005-06-13 |
Use a nested <javadoc> element and copy all of its children to the javadoc comment.
This allows use to use html elements like <p> in the javadoc xml and have them show
up in the javadoc comment.
|
Robert Hastings | 2005-06-14 |
Start of module support, made some big refactorings to share Jni code gen
|
Robert Hastings | 2005-06-14 |
Fix, when we moved to using <javadoc> we now need to check to see if zero nodes
are returned to decide to output the missing xpath expression
|
Robert Hastings | 2005-06-14 |
Generate xjavadoc for stdole
|
Robert Hastings | 2005-06-14 |
Fixed up javadoc formating for coclass.
|
Robert Hastings | 2005-06-14 |
Removed IDL: from javadoc is we find a user supplied javadoc using -xjavadoc.
|
Robert Hastings | 2005-06-15 |
Switched to using #import in an attempt to get the msxml stuff to compile under vc6.
|
Robert Hastings | 2005-06-15 |
Make calling Release from finalizer thread safer. Before a thrown Error could hang the ole thead.
|
Robert Hastings | 2005-06-15 |
Refactored to share getDllName, added getDllEntry
|
Robert Hastings | 2005-06-15 |
Added support for type library modules.
Now you can declare arbitrary functions in IDL and have tlb2java generate JNI to call them.
|
Robert Hastings | 2005-06-15 |
Removed unused private fields
|
Robert Hastings | 2005-06-15 |
Added a test target tlb-tests-xml to build the xml for com server
|
Robert Hastings | 2005-06-15 |
Removed [in] attribute
|
Robert Hastings | 2005-06-15 |
Added javadoc for module functions.
|
Robert Hastings | 2005-06-15 |
Refactored, moved outJavadocMethod to base
|
Robert Hastings | 2005-06-15 |
Added <return/> element to describe method and function return types.
Refactored getMethodAttrStr.
Added dll entry attributes for type infos that are modules.
|
Robert Hastings | 2005-06-15 |
Added javadoc merging for modules.
|
Robert Hastings | 2005-06-15 |
Changed type kind names to match those output in xml, should probably refactor xml
to use getTypeKindName as well...
Changed method signature to use FUNCDESC * instead of funcdesc_holder & so that xml can can call.
|
Robert Hastings | 2005-06-15 |
Fixed @return value javadoc generation. It uses the [retval] param if it exists, otherwise it uses
the functions return value unless that got mapped to void.
|
Robert Hastings | 2005-06-15 |
Compile new classes under vc6.
|
Adam Perry-Pelletier | 2005-06-16 |
Added support to generate XML for both the vtable and dispatch
sides of an dual interface.
|
Robert Hastings | 2005-06-16 |
Reformat source code as XML and pull // <frag comments out as actual xml elements.
|
Adam Perry-Pelletier | 2005-06-16 |
Updated the javadoc generation of capabilities of jSegue to
search for javadoc items by keying off a properties propget and propput
rather than the renamed Java properties accessors names (ie Latitude
rather than getLatitude).
|
Robert Hastings | 2005-06-17 |
Added ant.jar to build ExamplesXmlTask.
|
Robert Hastings | 2005-06-17 |
Ant task to generate examples.xml
|
Robert Hastings | 2005-06-17 |
Expanded classpath to handle ant task
|
Robert Hastings | 2005-06-17 |
Commented taskdef example
|
Robert Hastings | 2005-06-17 |
Suppress IDL for enums that have javadoc
|
build | 2005-06-17 |
Committed_by_Anthill
|
Robert Hastings | 2005-06-18 |
Fixed missing [ when dumping coclass interfaces.
|
Robert Hastings | 2005-06-21 |
Don't show the enum value if javadoc is supplied, the value is still available as a constant field value
|
Robert Hastings | 2005-06-21 |
Use end of line for end of frag markers to allow multiple frags on one one.
|
Robert Hastings | 2005-06-21 |
Changed the xml schema for package.xml files. Added a toplevel xjavadoc element. This allows a Java package
to be documented that has more than one type library mapped to it.
|
Robert Hastings | 2005-06-22 |
Include links to implemented interfaces in the coclass section of javadoc.
|
Robert Hastings | 2005-06-22 |
Include the source forge logo on our web page.
|
Robert Hastings | 2005-06-24 |
Fixed xpath expressions used in some xjavadoc lookups.
Allow parameters to be renamed so that javadoc names are more meaningful.
|
Robert Hastings | 2005-06-24 |
Added a comment and a tracker item, since I've looked at this a couple of times now
|
Robert Hastings | 2005-06-24 |
If xjavadoc is turned off return idl.
|
Robert Hastings | 2005-06-28 |
Switched from using an assert to throwing a java exception when we are asked
to convert a object to a variant. This means that the object is of a type we don't support
such as java.io.File.
|
Robert Hastings | 2005-06-28 |
Made the native leak detection thread a daemon thread. Otherwise, behavior is different between -ea
run none -ea runs. -ea could cause a program that would normally exit to not exit because of the leak
detection thread.
|
Robert Hastings | 2005-06-28 |
Allow parameter renames to use either the actual param name or the param number during the rename.
Since typical use of this feature is to rename __MIDL0087 parameters to something meaningful. __MIDL0087
names are automatically generated by midl when a parameter is unnamed. The number used may change from
one typelib version to the next. Using the parameter number allows easy handling because IFoo.Method.param.0
will be the same. Otherwise we might have IFoo.Method.__MIDL0078 for one version of the typelib,
but IFoo.Method.__MIDL0079 for the next.
|
Robert Hastings | 2005-06-29 |
Refactored to expose the garbage collection loop. New method COM.gc runs System.gc in a loop until there are no unreleased objects
or until 1 second has passed.
|
Robert Hastings | 2005-06-29 |
javadoc
|
Robert Hastings | 2005-07-04 |
Ignore t.exe
|
Adam Perry-Pelletier | 2005-07-14 |
Commit of first simple minded maven script. This script in now way
replaces our automatic build yet. It only does two thing: compiles java, and
tests the java. The properties file associated with it is equivalent to setting
-source 1.4 on the command line and gets us by the assert() not being found.
To test this out. Download and install maven, update path to find maven bin,
the simply run >maven java:compile or >maven test.
|