Apache Axis 1.2 Final ChangeLog (Changes after 1.2 RC3)

Designed for use with Ant.

2005-05-02 18:04 dims 
much better....at least i can understand what's happening :)
2005-05-02 17:00 dims 
Clean up the accumulated junk :) fixes marshall3 failures as well.
2005-05-02 16:59 dims 
more tests from Guillaume. Thanks Guillaume.
2005-05-02 15:23 dims 
compile _Helper's as well
2005-05-02 12:06 dims 
Additional tests from Guillaume.
2005-05-02 11:39 dims 
Patch for AXIS-1970 - Workaround for wsdl4j OperationType not being correctly serializable
2005-05-01 19:39 dims 
Got the client-side working for (AXIS-1926) wrapped document/literal generates wrong SOAP message for arrays within arrays
2005-05-01 15:02 dims 
Sort the TypeEntries...test/wsdl/echo fails on some machines for test2ComplexEchoServiceEcho21 and not on others because of the ordering. make the behavior consistent.
2005-05-01 04:33 dims 
comment out temporarily
2005-05-01 04:05 dims 
prevent clash with type
2005-04-30 21:38 dims 
Test case (in progress) for AXIS-1926 - wrapped document/literal generates wrong SOAP message for arrays within arrays

URL: http://issues.apache.org/jira/browse/AXIS-1926
2005-04-30 18:31 dims 
Fix for AXIS-1935 - .NET does not like xmlns=""
2005-04-30 13:03 dims 
Fix for AXIS-1887 -  tcpmon unable to display messages: faulty lookup of encoding
2005-04-30 00:19 dims 
Fix for (AXIS-1966) .NET interop: rpc/encoded array of array: regression from RC2

URL: http://issues.apache.org/jira/browse/AXIS-1966
2005-04-29 18:48 dims 
As promised the test case for replicating jonas failures.
2005-04-29 17:57 dims 
one more fix for jonas...will check in the test case in a bit. (doc/lit wsdl, --helperGen --wrapArrays true -t --typeMappingVersion 1.3, minus getter/setter indexed methods)
2005-04-29 16:41 tomj 
Put back in setDefaultTypeMapping() API to the Emitter since
a) it isn't nice to remove a public API
b) my product is *using* the public API, so it breaks :-)

Make sure it is depricated, but still does something sane.
Also fix a potential problem that the TypeMappingRegistry can be null
in the Emitter and dereferenced.  Create a default registry.
2005-04-29 15:35 dims 
additional tests from guillaume
2005-04-29 13:54 dims 
populate metadata for MaxOccurs slightly better (uncovered by Jonas tck)
2005-04-28 04:08 dims 
Fix for AXIS-1919 - Deserialization of derived types gets messed up
2005-04-27 16:33 dims 
Slightly better server-side log message...no idea why this is not being thrown back to the client.

(courtesy of Guillaume)
2005-04-27 13:37 dims 
Fix for  AXIS-1942 - 1.2RC3: rpc/enc mode, encoding long[] as xsd:anyType broken from 1.2RC2

URL: http://issues.apache.org/jira/browse/AXIS-1942
2005-04-27 02:26 dims 
all-tests works fine with dotNetSoapEncFix=true...any objections, please chime in quickly :)
2005-04-26 22:01 dims 
Fix for AXIS-1957  - BeanSerializer.writeSchema(Class,Types) stop class bug
from David Green

http://issues.apache.org/jira/browse/AXIS-1957
2005-04-26 21:06 dims 
Fix for AXIS-1910 - CLONE -axis serializes unwanted attributes (e.g. lowestSetBit of java.math.BigInteger)
2005-04-26 13:15 dims 
Patch from gdamour (geronimo)

Notes:
- AttributeDesc were not properly registered by the fieldNameMap Map when setFields was used.
- This mirrors the TypeDesc.addFieldDesc() method behavior.
2005-04-26 01:14 dims 
adding convenience constructors
2005-04-25 20:54 dims 
try to make tomj happy :)

- reinstate the working of getCall() in service. (with explicit javadoc about thread safety)
- remove my HACK and store last _call in o.a.a.c.Stub in the new _createCall method.
2005-04-25 20:51 dims 
Fix for all-tests failure
2005-04-25 14:51 dims 
enhancing the tests a bit (sending nulls and throwing exception on nulll array on server-side)
2005-04-25 14:44 venkat 
Applying the patch from Ashutosh for the issue AXIS-1757. I think the code change is too trivial to add a test case. If anybody thinks that this needs a test case, then i will add one.
2005-04-25 10:02 venkat 
Test case for the fix AXIS-1725. Added a new test case for passing javax.xml.soap.SOAPEnvelope.
2005-04-24 15:49 venkat 
Fix for the issue AXIS-1725. Message style service should be able to recieve javax.xml.soap.SOAPEnvelope as request parameter.
2005-04-24 03:00 dims 
Fix for AXIS-1682 - setting CHARACTER_SET_ENCODING in org.apache.axis.client.Call
from Bernard Bernstein

URL: http://issues.apache.org/jira/browse/AXIS-1682
2005-04-24 02:17 dims 
Fix for zero size attachment and minor code cleanup.
2005-04-23 21:37 dims 
Fix for AXIS-1906 - CLONE -dime attachment content is empty if datahandler's source is not FileDataSource
Fix for AXIS-1126 - dime attachment content is empty if datahandler's source is not FileDataSource

Notes:
- Need to use HTTP 1.1 to prevent the problem if getting the size of the stream.
- After this fix, CommonsHTTPSender should work fine out of the box.
- If you want to use HTTPSender, then switch on the HTTP11 chunked support using the following code:
        call.setProperty(MessageContext.HTTP_TRANSPORT_VERSION,
                HTTPConstants.HEADER_PROTOCOL_V11);
        Hashtable httpconf = new Hashtable();
        httpconf.put(HTTPConstants.HEADER_TRANSFER_ENCODING,
              HTTPConstants.HEADER_TRANSFER_ENCODING_CHUNKED);
        call.setProperty(HTTPConstants.REQUEST_HEADERS,httpconf);
2005-04-23 01:49 dims 
fix typo and don't add empty vectors as collision candidates.
2005-04-23 01:47 dims 
Fix for AXIS-1890 - Support additional constructors in generated fault stubs
2005-04-23 01:46 dims 
Fix for AXIS-1928 - Non-Exception Throwables are not logged
2005-04-22 15:25 dims 
Fix for AXIS-1895 - Use of hard coded value instead of constant

URL: http://issues.apache.org/jira/browse/AXIS-1895
2005-04-22 15:14 dims 
oops forgot the test cases.
2005-04-22 14:51 dims 
fix jar names
2005-04-22 14:42 dims 
Fix for AXIS-1806 - axis does not include license agreements for third party jar files it re-distributes
Fix for AXIS-1808 - Include version number for ALL jar files in the file name (like log4j-1.2.8jar)
2005-04-22 13:59 dims 
Fix for AXIS-935 - Using ThreadLocal Call objects in a servlet memory leak both on client and server side

After reading tom's comment...moved the _getCall to stub, which ensures that Service is threadsafe. Added a deprecated flag and an exception in Service's getCall to point people in the right direction.
2005-04-21 22:56 dims 
belated Fix for AXIS-1331 - WSDL2Java genetates incorrect code from WSDL
2005-04-21 19:47 dims 
Fix for AXIS-1941 - WSDDDeployment MBean contribution

heavily tweaked code contrib originally sent by Algirdas Veitas. Tested using a combination of Tomcat55, MX4J, MC4J, JDK15, JCONSOLE.

URL: http://issues.apache.org/jira/browse/AXIS-1941
2005-04-21 11:42 dims 
Fix for AXIS-1939 - Default Configuration of CommonHTTPSender
from Marco van Eck

URL: http://issues.apache.org/jira/browse/AXIS-1939
2005-04-21 02:26 dims 
Fix for AXIS-1943 - ID dups with FastUUIDGen
from M Kisimoto

URL: http://issues.apache.org/jira/browse/AXIS-1943
2005-04-21 02:09 dims 
Fix CR/LF problems.
2005-04-20 23:55 dims 
Fix for AXIS-1754 - Service ThreadLocal previousCall should not be static
2005-04-20 22:18 dims 
Fix for  AXIS-895 - Axis doesn't maintain more than one cookie with http
2005-04-19 02:17 dims 
Updated TerraService.WSDL
2005-04-18 17:17 gdaniels 
Flesh out echo2 test case to actually check the XML, and also make
it work.

Pass itemQName info through the generated Stub and also the WSDD
when appropriate, which allows the server to do the right thing when
serializing arrays.

This isn't 100% right yet, in that the xsi:type on the maxOccurs test for
echo2 shouldn't be "soapenc:Array", but it's close (and heck, .NET
ignores xsi:type attributes anyway... :))

NOTE : Checking this in without a successful all-tests run, because the
terra server (for test/wsdl/terra) seems broken - damned external
dependencies. :(
2005-04-18 17:08 dims 
Partial fix for AXIS-1932 - global configuration properties axis.disableServiceList and axis.servicesPath don't work

Notes:
- just prevent the NPE for now to take this off the blocker list.
2005-04-17 02:25 gdaniels 
Check in Guillaume's test, which works for now - because we're
managing to deserialize XML which doesn't match the schema (we
don't care what the "item" QName is when deserializing an array).

Need to update the test to check the XML manually, and fix the
array serialization problem.
2005-04-13 16:22 gdaniels 
unwrapArrays -> wrapArrays.  Makes much more sense. :)

Default is now OFF (don't wrap arrays with beans)
2005-04-13 16:22 gdaniels 
unwrapArrays -> wrapArrays.  Makes much more sense. :)

Default is now OFF (don't wrap arrays with beans)
2005-04-12 23:42 dims 
oops forgot the main file
2005-04-12 23:07 dims 
- Upgrade to Commons HTTPClient RC2
- Fix several deprecated API usage
- Get NTLM auth working for the endpoint (prev code took care of NTLM proxy auth)
2005-04-12 23:07 dims 
- Upgrade to Commons HTTPClient RC2
- Fix several deprecated API usage
- Get NTLM auth working for the endpoint (prev code took care of NTLM proxy auth)
2005-04-12 22:18 gdaniels 
Support "wrapped" literal arrays, of the following form:

<element name="foo">
 <complexType>
  <sequence>
   <element name="item" maxOccurs="unbounded" type="xsd:string"/>
  </sequence>
 </complexType>
</element>

We used to WSDL2Java this into a bean with a String[] member, but
we will now by default represent the "foo" field/parameter as a
String [] itself.  We will also by default write out WSDL using
this style, so we'll prefer this type of array to repeating
the "foo" element itself.  This last bit is critical for non-
wrapped document services returning arrays, since we were
formerly writing illegal schema (global elements are not
allowed to have maxOccurs)!

There is a WSDL2Java flag "unwrapArrays" which defaults to
"yes".  If you set it to "no", we won't do this stuff (and
remain backwards compatible).

Note that all of this is only in non-encoded mode.  SOAP RPC
encoding continues to work as usual (with soapenc:Arrays).

TODO:

There are still some edge cases to test out, but checking this
in for now so people concerned with these types of issues can
test with the current CVS.  I suspect we need an "unwrapArrays"
flag on Java2WSDL and in the engine as well.  A little cleanup
wouldn't hurt.

Details:

* ElementDesc and ParameterDesc both carry an "itemQName"
  which enables us to keep track of this kind of usage, and
  get the "inner" QName for these types of collections to
  correctly match the schema.  This needs to propagate to
  various places (esp. SerializationContext and then
  ArraySerializer), so you'll see it pop up a bunch.

* Fix up a little logic + formatting in ArrayDeserializer

* ArraySerializer plucks the itemQName from the current
  SerializationContext and uses that instead of always
  using "item".

* BeanDeserializer uses the presence of itemQName metadata
  to tell if it's a "normal" repeated element (and thus
  each one should be deserialized by the component type's
  deser) or a literal array (in which case the Array
  Deserializer should be used)

* BeanSerializer does the same

* RPCHandler makes sure the ArrayDeserializer gets used
  when appropriate (all array params in document style)

* Support "unwrapArrays" switch in various places (ant
  task, command line, thread value down through Emitter
  and SymbolTable...)

* Shut off unwrapArrays for various tests which would
  break otherwise

* Make sure to correctly write WSDL for wrapped arrays
  and also to correctly write metadata into generated
  Java classes

* Change test.wsdl.arrays3 test to use unwrapped arrays
2005-04-12 22:18 gdaniels 
Support "wrapped" literal arrays, of the following form:

<element name="foo">
 <complexType>
  <sequence>
   <element name="item" maxOccurs="unbounded" type="xsd:string"/>
  </sequence>
 </complexType>
</element>

We used to WSDL2Java this into a bean with a String[] member, but
we will now by default represent the "foo" field/parameter as a
String [] itself.  We will also by default write out WSDL using
this style, so we'll prefer this type of array to repeating
the "foo" element itself.  This last bit is critical for non-
wrapped document services returning arrays, since we were
formerly writing illegal schema (global elements are not
allowed to have maxOccurs)!

There is a WSDL2Java flag "unwrapArrays" which defaults to
"yes".  If you set it to "no", we won't do this stuff (and
remain backwards compatible).

Note that all of this is only in non-encoded mode.  SOAP RPC
encoding continues to work as usual (with soapenc:Arrays).

TODO:

There are still some edge cases to test out, but checking this
in for now so people concerned with these types of issues can
test with the current CVS.  I suspect we need an "unwrapArrays"
flag on Java2WSDL and in the engine as well.  A little cleanup
wouldn't hurt.

Details:

* ElementDesc and ParameterDesc both carry an "itemQName"
  which enables us to keep track of this kind of usage, and
  get the "inner" QName for these types of collections to
  correctly match the schema.  This needs to propagate to
  various places (esp. SerializationContext and then
  ArraySerializer), so you'll see it pop up a bunch.

* Fix up a little logic + formatting in ArrayDeserializer

* ArraySerializer plucks the itemQName from the current
  SerializationContext and uses that instead of always
  using "item".

* BeanDeserializer uses the presence of itemQName metadata
  to tell if it's a "normal" repeated element (and thus
  each one should be deserialized by the component type's
  deser) or a literal array (in which case the Array
  Deserializer should be used)

* BeanSerializer does the same

* RPCHandler makes sure the ArrayDeserializer gets used
  when appropriate (all array params in document style)

* Support "unwrapArrays" switch in various places (ant
  task, command line, thread value down through Emitter
  and SymbolTable...)

* Shut off unwrapArrays for various tests which would
  break otherwise

* Make sure to correctly write WSDL for wrapped arrays
  and also to correctly write metadata into generated
  Java classes

* Change test.wsdl.arrays3 test to use unwrapped arrays
2005-04-12 22:18 gdaniels 
Support "wrapped" literal arrays, of the following form:

<element name="foo">
 <complexType>
  <sequence>
   <element name="item" maxOccurs="unbounded" type="xsd:string"/>
  </sequence>
 </complexType>
</element>

We used to WSDL2Java this into a bean with a String[] member, but
we will now by default represent the "foo" field/parameter as a
String [] itself.  We will also by default write out WSDL using
this style, so we'll prefer this type of array to repeating
the "foo" element itself.  This last bit is critical for non-
wrapped document services returning arrays, since we were
formerly writing illegal schema (global elements are not
allowed to have maxOccurs)!

There is a WSDL2Java flag "unwrapArrays" which defaults to
"yes".  If you set it to "no", we won't do this stuff (and
remain backwards compatible).

Note that all of this is only in non-encoded mode.  SOAP RPC
encoding continues to work as usual (with soapenc:Arrays).

TODO:

There are still some edge cases to test out, but checking this
in for now so people concerned with these types of issues can
test with the current CVS.  I suspect we need an "unwrapArrays"
flag on Java2WSDL and in the engine as well.  A little cleanup
wouldn't hurt.

Details:

* ElementDesc and ParameterDesc both carry an "itemQName"
  which enables us to keep track of this kind of usage, and
  get the "inner" QName for these types of collections to
  correctly match the schema.  This needs to propagate to
  various places (esp. SerializationContext and then
  ArraySerializer), so you'll see it pop up a bunch.

* Fix up a little logic + formatting in ArrayDeserializer

* ArraySerializer plucks the itemQName from the current
  SerializationContext and uses that instead of always
  using "item".

* BeanDeserializer uses the presence of itemQName metadata
  to tell if it's a "normal" repeated element (and thus
  each one should be deserialized by the component type's
  deser) or a literal array (in which case the Array
  Deserializer should be used)

* BeanSerializer does the same

* RPCHandler makes sure the ArrayDeserializer gets used
  when appropriate (all array params in document style)

* Support "unwrapArrays" switch in various places (ant
  task, command line, thread value down through Emitter
  and SymbolTable...)

* Shut off unwrapArrays for various tests which would
  break otherwise

* Make sure to correctly write WSDL for wrapped arrays
  and also to correctly write metadata into generated
  Java classes

* Change test.wsdl.arrays3 test to use unwrapped arrays
2005-04-12 22:18 gdaniels 
Support "wrapped" literal arrays, of the following form:

<element name="foo">
 <complexType>
  <sequence>
   <element name="item" maxOccurs="unbounded" type="xsd:string"/>
  </sequence>
 </complexType>
</element>

We used to WSDL2Java this into a bean with a String[] member, but
we will now by default represent the "foo" field/parameter as a
String [] itself.  We will also by default write out WSDL using
this style, so we'll prefer this type of array to repeating
the "foo" element itself.  This last bit is critical for non-
wrapped document services returning arrays, since we were
formerly writing illegal schema (global elements are not
allowed to have maxOccurs)!

There is a WSDL2Java flag "unwrapArrays" which defaults to
"yes".  If you set it to "no", we won't do this stuff (and
remain backwards compatible).

Note that all of this is only in non-encoded mode.  SOAP RPC
encoding continues to work as usual (with soapenc:Arrays).

TODO:

There are still some edge cases to test out, but checking this
in for now so people concerned with these types of issues can
test with the current CVS.  I suspect we need an "unwrapArrays"
flag on Java2WSDL and in the engine as well.  A little cleanup
wouldn't hurt.

Details:

* ElementDesc and ParameterDesc both carry an "itemQName"
  which enables us to keep track of this kind of usage, and
  get the "inner" QName for these types of collections to
  correctly match the schema.  This needs to propagate to
  various places (esp. SerializationContext and then
  ArraySerializer), so you'll see it pop up a bunch.

* Fix up a little logic + formatting in ArrayDeserializer

* ArraySerializer plucks the itemQName from the current
  SerializationContext and uses that instead of always
  using "item".

* BeanDeserializer uses the presence of itemQName metadata
  to tell if it's a "normal" repeated element (and thus
  each one should be deserialized by the component type's
  deser) or a literal array (in which case the Array
  Deserializer should be used)

* BeanSerializer does the same

* RPCHandler makes sure the ArrayDeserializer gets used
  when appropriate (all array params in document style)

* Support "unwrapArrays" switch in various places (ant
  task, command line, thread value down through Emitter
  and SymbolTable...)

* Shut off unwrapArrays for various tests which would
  break otherwise

* Make sure to correctly write WSDL for wrapped arrays
  and also to correctly write metadata into generated
  Java classes

* Change test.wsdl.arrays3 test to use unwrapped arrays
2005-04-12 17:35 gdaniels 
Get all-tests working on LiteralArrayWork branch.
2005-04-12 17:25 dims 
Updated to latest WSDL4J (1.5.1)
2005-04-12 14:34 gdaniels 
Make sure SimpleListDeserializer is on the list of non-indexy ones.
2005-04-12 12:28 gdaniels 
Checking in current state of literal array changes on a branch
so Tom and I can easily share it this AM.
2005-04-12 12:28 gdaniels 
Checking in current state of literal array changes on a branch
so Tom and I can easily share it this AM.
2005-04-12 12:28 gdaniels 
Checking in current state of literal array changes on a branch
so Tom and I can easily share it this AM.
2005-04-12 12:28 gdaniels 
Checking in current state of literal array changes on a branch
so Tom and I can easily share it this AM.
2005-04-12 12:23 gdaniels 
branches:  1.1.2;
Test doc/lit service with arrays at the ParameterDesc level (just checking in so I can make a branch)
2005-04-12 12:23 gdaniels 
Test doc/lit service with arrays at the ParameterDesc level (just checking in so I can make a branch)
2005-04-08 14:22 dims 
Fix for AXIS-1889 - Problem with namespace handling in Axis 1.2 : break XML Signature
from Guillaume Sauthier
2005-04-06 04:59 toshi 
Changed the link for the Axis Wiki in Japanese.
2005-03-31 10:30 toshi 
branches:  1.114.2;
On the i18n task, we've found a lot of unused keys at the current
resource file. To make clear the status, I've added the following
changes to the file.

 BEFORE: resource.properties
   +-----------------+
   | key-1 = ....    | <-- (In-use)
   | key-2 = ....    | <-- (Unused)
   | key-3 = ....    | <-- (In-use)
   | key-4 = ....    | <-- (Unused)
   | key-5 = ....    | <-- (Unused)
   | key-6 = ....    | <-- (In-use)
   :                 :
   +-----------------+

 AFTER: resource.properties
   +-----------------+
   | ### In-use keys |
   | #               |
   | key-1 = ....    |
   | key-3 = ....    |
   | key-6 = ....    |
   |                 |
   | ### Unused keys |
   | #               |
   | key-2 = ....    |
   | key-4 = ....    |
   | key-5 = ....    |
   :                 :
   +-----------------+

Credit:
 On this task, Shouhei Kodama has been a great help in creating
the new resource file.

Thanks,
Toshi
2005-03-29 04:17 dims 
Fix for AXIS-1902 - Serializable HandlerInfoChainFactory not serializable.
2005-03-25 10:02 jongjinchoi 
Ant 1.5 support reverted
2005-03-25 10:02 jongjinchoi 
branches:  1.142.2;
Ant 1.5 support reverted
2005-03-24 21:43 tomj 
Fix bug Axis-1900 - WSDL2Java fails on SQL reporting service.
  http://issues.apache.org/jira/browse/AXIS-1900

When we check to see if an operations parameter schema can be unwrapped
we did not allow choice elements in the squence.

This does not address the real problem, which is that "wrapped" in the symbol
table is a global for the *service* and not the operation.  If we turn it off,
it should be for the operation only.  This would involve revamping the whole
system, so Axis2 will have to solve this.
2005-03-24 21:43 tomj 
branches:  1.53.2;
Fix bug Axis-1900 - WSDL2Java fails on SQL reporting service.
  http://issues.apache.org/jira/browse/AXIS-1900

When we check to see if an operations parameter schema can be unwrapped
we did not allow choice elements in the squence.

This does not address the real problem, which is that "wrapped" in the symbol
table is a global for the *service* and not the operation.  If we turn it off,
it should be for the operation only.  This would involve revamping the whole
system, so Axis2 will have to solve this.
2005-03-24 21:43 tomj 
branches:  1.117.2;
Fix bug Axis-1900 - WSDL2Java fails on SQL reporting service.
  http://issues.apache.org/jira/browse/AXIS-1900

When we check to see if an operations parameter schema can be unwrapped
we did not allow choice elements in the squence.

This does not address the real problem, which is that "wrapped" in the symbol
table is a global for the *service* and not the operation.  If we turn it off,
it should be for the operation only.  This would involve revamping the whole
system, so Axis2 will have to solve this.
2005-03-23 02:24 jongjinchoi 
branches:  1.1.2;
add testcase for AXIS-1891
2005-03-23 02:24 jongjinchoi 
add testcase for AXIS-1891
2005-03-23 02:22 jongjinchoi 
separate NUll and non-covertable in ArrayOfT --> T[] conversion
2005-03-22 19:56 gdaniels 
Account for new (correct) ordering of params which matches schema.
2005-03-22 12:24 dims 
Fix for AXIS-1873 - JavaBeanWriter reorders variable names in generated constructors
Fix for AXIS-1865 - Axis 1.2 RC3 wsdl2Java problem
2005-03-21 15:00 jongjinchoi 
Fix for AXIS-1891's first patch.
(http://issues.apache.org/jira/browse/AXIS-1891)
2005-03-21 12:37 dims 
branches:  1.85.2;
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-21 12:37 dims 
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-21 12:37 dims 
branches:  1.84.2;
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-21 12:37 dims 
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-21 12:37 dims 
branches:  1.112.2;
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-21 12:37 dims 
Fix for AXIS-1891 - Expanding Java2wsdl to support the starting-from-java wrapped style web service
from Jongjin Choi

URL: http://issues.apache.org/jira/browse/AXIS-1891
2005-03-20 18:58 gdaniels 
branches:  1.50.2;
Fix:

http://issues.apache.org/jira/browse/AXIS-1869

Don't walk the ref chain unless the TypeEntry is a CollectionTE - this should
DTRT for true derived types.
2005-03-19 03:48 dims 
javax.wsdl.extensions.soap.SOAPBody.getParts is returning a List consisting of String's. (NOT Parts as assumed in the code)
2005-03-19 03:10 dims 
Set the QName to that got from getBindedParameterName only if it is not null.
2005-03-17 18:12 gawor 
on one way invocation always release the connection
2005-03-17 17:19 dims 
prevent NPE.
2005-03-17 06:39 gawor 
improved CommonsHTTPSender. 1) the entire message is not serialized in memory anymore, 2) made it easier for others to extend the sender code, 3) added support for same cfg options as HTTPSender
2005-03-16 06:55 dims 
Fix for AXIS-1878 - Serializable OperationDesc with method instance
from David Blevins
2005-03-15 22:33 dims 
Fix for AXIS-1855 - Add XPath in Castor(De)Serializer on ValidationException
from Jene Jasper  

URL: http://issues.apache.org/jira/browse/AXIS-1855
2005-03-15 22:02 dims 
Fix for AXIS-1867 - Exception thrown in Tomcat 5.5.7 in AxisServletBase
from Alexandre ROMAN 

URL: http://issues.apache.org/jira/browse/AXIS-1867
2005-03-15 21:42 dims 
Fix for AXIS-1872 - Allow for streaming of DIME attachments
from Marc Dumontier

URL: http://issues.apache.org/jira/browse/AXIS-1872

Example in bug report:
DynamicContentDataHandler datahandler = null;
URL url = new URL("http",host,port,QUERY_STRING);
URLDataSource dataSource = new URLDataSource(url);
datahandler = new DynamicContentDataHandler(dataSource);
datahandler.setChunkSize(10 * 1024 * 1024);
2005-03-15 17:27 gawor 
read last \r\n if this is the last token
2005-03-15 13:33 dims 
Fix for AXIS-1875 : Serializable JavaServiceDesc et. all
from David Blevins

URL: http://issues.apache.org/jira/browse/AXIS-1875
2005-03-10 03:23 dims 
Fix for AXIS-1862 - Significant performance problem with NodeImpl.getOwnerDocument
from Michael Thome (mthome@bbn.com)
2005-03-08 15:50 dims 
Fix for AXIS-1857 - PositiveInteger fields in a bean are incorrectly serialized (lowestSetBit attribute)
from Yves Langisch

URL: http://issues.apache.org/jira/browse/AXIS-1857
2005-03-05 01:25 dims 
Fix for AXIS-1841 - QSWSDLHandler does not correctly substitute soap:address location URLs when WSDL contains multiple ports
from Ian P. Springer

URL: http://issues.apache.org/jira/browse/AXIS-1841
2005-03-04 08:43 venkat 
Checking in the patch from Hans for AXIS-66 (xsd:list support). The changes are hand-merged since the patch is not in valid format.

Also i have ran both JAX-RPC and SAAJ TCK and ensured that all tests ran green.

-- venkat
2005-03-03 18:28 dims 
Fix for AXIS-530 - elements without minOccurs="0" produce isnill="true" on the wire
2005-03-01 00:13 dims 
Add extra null checks (some folks don't use AxisEngine :)
2005-02-28 15:57 dims 
add missing xml declaration
2005-02-26 21:37 gdaniels 
Some cleanup for TypeMappings - share defaults again, produce a
few less instances, delegate a little better.
2005-02-26 21:37 gdaniels 
branches:  1.106.2;
Some cleanup for TypeMappings - share defaults again, produce a
few less instances, delegate a little better.
2005-02-26 21:37 gdaniels 
Some cleanup for TypeMappings - share defaults again, produce a
few less instances, delegate a little better.
2005-02-25 04:06 dims 
Key off the WS-I BP flag.
2005-02-25 03:46 dims 
fix TCK failure
2005-02-25 03:28 dims 
remove unnecessary typecasting
2005-02-24 21:59 gdaniels 
* Re-do delegation logic.  There were various problems as a result
  of TypeMappingImpls being both the repository for mappings and
  also the links on delegation chains themselves.  Now TypeMappingImpl
  is purely an internal functional class, and all accesses to
  TypeMappings from the outside use TypeMappingDelegates (which
  might now be better named TypeMappingProxy).  A TypeMappingDelegate
  is responsible for a) delegating all functionality to the
  TypeMappingImpl it points to, and b) maintaining the link to the
  next TypeMappingDelegate in a scoping chain.  This lets us be
  more flexible about connecting them together.  If I have
  SOAPENC and default mappings in TMR1 and delegate to TMR2, the
  order is TMR1[soapenc]->TMR2[soapenc]->TMR1[default]->tmr2[default]

* Create TypeMappingDelegates not TypeMappingImpls wherever needed

* Centralize all the type mapping version code in the TMR.

* Rejigger the Emitters (from/to Java) - instead of having a separate
  default type mapping, just use a TypeMappingRegistry which
  will have its own defaults built in.  This means that the TM we
  use will be able to do defaulting itself.

* Use consistent type mapping versions in addrImplNoSEI/build.xml

* Use consistent type mapping versions in roundtrip2/build.xml
  (wsdl2java needs to be consistent with java2wsdl)

* Various tweaks to tests to accommodate above changes

* Don't block exceptions in JAXRPCSamples test

* Still have some cleanup to do, and right now we're creating way
  too many instances of TypeMappingImpls.  Next step is to change
  it so that only the engine-level TMRs contain the SOAP encoding
  mappings, and subsidiary TMRs (for services) only contain default
  mappings if they specify a "typeMappingVersion" different from
  that of the engine.
2005-02-24 21:59 gdaniels 
* Re-do delegation logic.  There were various problems as a result
  of TypeMappingImpls being both the repository for mappings and
  also the links on delegation chains themselves.  Now TypeMappingImpl
  is purely an internal functional class, and all accesses to
  TypeMappings from the outside use TypeMappingDelegates (which
  might now be better named TypeMappingProxy).  A TypeMappingDelegate
  is responsible for a) delegating all functionality to the
  TypeMappingImpl it points to, and b) maintaining the link to the
  next TypeMappingDelegate in a scoping chain.  This lets us be
  more flexible about connecting them together.  If I have
  SOAPENC and default mappings in TMR1 and delegate to TMR2, the
  order is TMR1[soapenc]->TMR2[soapenc]->TMR1[default]->tmr2[default]

* Create TypeMappingDelegates not TypeMappingImpls wherever needed

* Centralize all the type mapping version code in the TMR.

* Rejigger the Emitters (from/to Java) - instead of having a separate
  default type mapping, just use a TypeMappingRegistry which
  will have its own defaults built in.  This means that the TM we
  use will be able to do defaulting itself.

* Use consistent type mapping versions in addrImplNoSEI/build.xml

* Use consistent type mapping versions in roundtrip2/build.xml
  (wsdl2java needs to be consistent with java2wsdl)

* Various tweaks to tests to accommodate above changes

* Don't block exceptions in JAXRPCSamples test

* Still have some cleanup to do, and right now we're creating way
  too many instances of TypeMappingImpls.  Next step is to change
  it so that only the engine-level TMRs contain the SOAP encoding
  mappings, and subsidiary TMRs (for services) only contain default
  mappings if they specify a "typeMappingVersion" different from
  that of the engine.
2005-02-24 21:59 gdaniels 
* Re-do delegation logic.  There were various problems as a result
  of TypeMappingImpls being both the repository for mappings and
  also the links on delegation chains themselves.  Now TypeMappingImpl
  is purely an internal functional class, and all accesses to
  TypeMappings from the outside use TypeMappingDelegates (which
  might now be better named TypeMappingProxy).  A TypeMappingDelegate
  is responsible for a) delegating all functionality to the
  TypeMappingImpl it points to, and b) maintaining the link to the
  next TypeMappingDelegate in a scoping chain.  This lets us be
  more flexible about connecting them together.  If I have
  SOAPENC and default mappings in TMR1 and delegate to TMR2, the
  order is TMR1[soapenc]->TMR2[soapenc]->TMR1[default]->tmr2[default]

* Create TypeMappingDelegates not TypeMappingImpls wherever needed

* Centralize all the type mapping version code in the TMR.

* Rejigger the Emitters (from/to Java) - instead of having a separate
  default type mapping, just use a TypeMappingRegistry which
  will have its own defaults built in.  This means that the TM we
  use will be able to do defaulting itself.

* Use consistent type mapping versions in addrImplNoSEI/build.xml

* Use consistent type mapping versions in roundtrip2/build.xml
  (wsdl2java needs to be consistent with java2wsdl)

* Various tweaks to tests to accommodate above changes

* Don't block exceptions in JAXRPCSamples test

* Still have some cleanup to do, and right now we're creating way
  too many instances of TypeMappingImpls.  Next step is to change
  it so that only the engine-level TMRs contain the SOAP encoding
  mappings, and subsidiary TMRs (for services) only contain default
  mappings if they specify a "typeMappingVersion" different from
  that of the engine.
2005-02-23 13:22 venkat 
reverting the previous changes as that doesn't seem to be the right way to fix the issues with typemapping version.
2005-02-23 13:02 venkat 
Updated default typemapping from 1.2 to 1.3, hoping to fix the JAXRPC TCK failures.
2005-02-22 03:32 dims 
Mirror the changes made in org.apache.axis.wsdl.fromJava.Emitter
2005-02-21 21:37 gdaniels 
First cut at fix for roundtrip2 - this doesn't really solve the problem since it
re-creates the dependence on SOAP encoding whenver the typemapping
is "1.3".  Working on a better fix but having some issues, so checking this
in, since it doesn't do anything worse than it did with the original
SOAPENC-derived JAXRPCTypeMappingImpl
2005-02-21 21:37 gdaniels 
First cut at fix for roundtrip2 - this doesn't really solve the problem since it
re-creates the dependence on SOAP encoding whenver the typemapping
is "1.3".  Working on a better fix but having some issues, so checking this
in, since it doesn't do anything worse than it did with the original
SOAPENC-derived JAXRPCTypeMappingImpl
2005-02-21 03:04 dims 
Roundtrip test for typemappingversion=1.3 [edit the build.xml to recreate the problem]
2005-02-18 13:03 jongjinchoi 
branches:  1.195.2;
Fix for AXIS-1817, 1823.

- SOAPEnvelope#outputImpl() is modified so that envelope's each children's output or outputImpl methods is called in order. (What is the trailers in SOAPEnvelope?)

- MessageElement#getValue() :
   TextSerializationContext is not sufficient for getting the exact text node.
   So, comment it.

- MessageElement#getValue() looks for the first child text node among its children.  (not to check if just first node is text node)

- SOAPEnvelope's getValue() is removed. MessageElement#getValue() is used instead.

- SOAPHandler :
     - Add text child to current deserialing element (myElemetn) in onStartChild() for mixed contents.
     - To avoid appending text child for ignorable white spaces, the stripped string's length is checked, but it is necessary to append the original unstripped text child to keep its non-igrnorable white spaces.

- test/message/TestMessageElement.java
   fixed invalid test case for mixed contents according to SAAJ-spec and SUN's RI.

- test/saaj/TestEnvelope.java
   added test case for checking appended text node is position as expected (after soapbody)
2005-02-18 13:03 jongjinchoi 
branches:  1.19.2;
Fix for AXIS-1817, 1823.

- SOAPEnvelope#outputImpl() is modified so that envelope's each children's output or outputImpl methods is called in order. (What is the trailers in SOAPEnvelope?)

- MessageElement#getValue() :
   TextSerializationContext is not sufficient for getting the exact text node.
   So, comment it.

- MessageElement#getValue() looks for the first child text node among its children.  (not to check if just first node is text node)

- SOAPEnvelope's getValue() is removed. MessageElement#getValue() is used instead.

- SOAPHandler :
     - Add text child to current deserialing element (myElemetn) in onStartChild() for mixed contents.
     - To avoid appending text child for ignorable white spaces, the stripped string's length is checked, but it is necessary to append the original unstripped text child to keep its non-igrnorable white spaces.

- test/message/TestMessageElement.java
   fixed invalid test case for mixed contents according to SAAJ-spec and SUN's RI.

- test/saaj/TestEnvelope.java
   added test case for checking appended text node is position as expected (after soapbody)
2005-02-18 13:03 jongjinchoi 
Fix for AXIS-1817, 1823.

- SOAPEnvelope#outputImpl() is modified so that envelope's each children's output or outputImpl methods is called in order. (What is the trailers in SOAPEnvelope?)

- MessageElement#getValue() :
   TextSerializationContext is not sufficient for getting the exact text node.
   So, comment it.

- MessageElement#getValue() looks for the first child text node among its children.  (not to check if just first node is text node)

- SOAPEnvelope's getValue() is removed. MessageElement#getValue() is used instead.

- SOAPHandler :
     - Add text child to current deserialing element (myElemetn) in onStartChild() for mixed contents.
     - To avoid appending text child for ignorable white spaces, the stripped string's length is checked, but it is necessary to append the original unstripped text child to keep its non-igrnorable white spaces.

- test/message/TestMessageElement.java
   fixed invalid test case for mixed contents according to SAAJ-spec and SUN's RI.

- test/saaj/TestEnvelope.java
   added test case for checking appended text node is position as expected (after soapbody)
2005-02-18 13:03 jongjinchoi 
Fix for AXIS-1817, 1823.

- SOAPEnvelope#outputImpl() is modified so that envelope's each children's output or outputImpl methods is called in order. (What is the trailers in SOAPEnvelope?)

- MessageElement#getValue() :
   TextSerializationContext is not sufficient for getting the exact text node.
   So, comment it.

- MessageElement#getValue() looks for the first child text node among its children.  (not to check if just first node is text node)

- SOAPEnvelope's getValue() is removed. MessageElement#getValue() is used instead.

- SOAPHandler :
     - Add text child to current deserialing element (myElemetn) in onStartChild() for mixed contents.
     - To avoid appending text child for ignorable white spaces, the stripped string's length is checked, but it is necessary to append the original unstripped text child to keep its non-igrnorable white spaces.

- test/message/TestMessageElement.java
   fixed invalid test case for mixed contents according to SAAJ-spec and SUN's RI.

- test/saaj/TestEnvelope.java
   added test case for checking appended text node is position as expected (after soapbody)
2005-02-18 12:36 dims 
Fix typo in glen's check-in
2005-02-17 20:05 gdaniels 
Fix up type mappings.

The DefaultJAXRPC11TypeMappingImpl shouldn't have extended
DefaultSOAPEncodedTMI, because that means that SOAP encoded
types would be polluting the default typemapping.  According to
JAX-RPC (and common sense) the default TM is not supposed to have
any encodingStyle-specific types in it.

JAXRPC type mapping now extends the normal type mapping, is
correctly registered as the defaultDelTM when using "1.3" as the
TypeMappingVersion (yuk), and there's a new test to check to make
sure SOAPENC_STRING is *not* mapped for the default TM no
matter which TMV we select.
2005-02-17 20:05 gdaniels 
Fix up type mappings.

The DefaultJAXRPC11TypeMappingImpl shouldn't have extended
DefaultSOAPEncodedTMI, because that means that SOAP encoded
types would be polluting the default typemapping.  According to
JAX-RPC (and common sense) the default TM is not supposed to have
any encodingStyle-specific types in it.

JAXRPC type mapping now extends the normal type mapping, is
correctly registered as the defaultDelTM when using "1.3" as the
TypeMappingVersion (yuk), and there's a new test to check to make
sure SOAPENC_STRING is *not* mapped for the default TM no
matter which TMV we select.
2005-02-17 15:34 dims 
Fix for the breaking test in marshall2
2005-02-17 13:52 dims 
Breaking all-tests intentionally - Test case from Guillaume.
2005-02-17 08:57 toshi 
Localizing the key 'numericEnc00' for Japanese.
2005-02-17 08:54 toshi 
Fixing AXIS-1815 - http://issues.apache.org/jira/browse/AXIS-1815.
Please test with the nightly build at:
<http://cvs.apache.org/dist/axis/nightly/axis-bin-1_2RC3.zip>

  Now, you should be able to see any mulit-byte characters if you
have enough font sets on your system. Otherwise you can check the
codes as a set of hexadecimal numeric character refernces such as
'&#x6C34;' by just turning on the 'Numeric' check box on the GUI.

Thanks,
Toshi
2005-02-15 10:15 toshi 
Changed the logic to check the delimiter - ';'.
2005-02-15 09:36 toshi 
Clean up to remain independent from the commons-lang code.
Feel free to add a patch, if it's not much to look at.
2005-02-15 09:36 toshi 
Clean up to remain independent from the commons-lang code.
Feel free to add a patch, if it's not much to look at.
2005-02-14 18:20 dims 
sample test case for rpc/lit with typemapping="1.3"
2005-02-14 18:20 dims 
branches:  1.1.2;
sample test case for rpc/lit with typemapping="1.3"
2005-02-14 13:19 dims 
Partially revert toshi's patch that adds commons-lang dependency. Copied relevant code from commons-lang (need to clean up when toshi's ready)
2005-02-14 13:19 dims 
Partially revert toshi's patch that adds commons-lang dependency. Copied relevant code from commons-lang (need to clean up when toshi's ready)
2005-02-14 13:17 dims 
Partially revert toshi's patch that adds commons-lang dependency. Copied relevant code from commons-lang (need to clean up when toshi's ready)
2005-02-14 09:10 toshi 
To support multi-language.
2005-02-14 09:09 toshi 
To support multi-language.
2005-02-14 09:08 toshi 
To support multi-language.
2005-02-14 09:07 toshi 
To support multi-language.
2005-02-14 04:05 dims 
branches:  1.144.2;
Restore JDK1.3 compat
2005-02-14 04:05 dims 
Restore JDK1.3 compat
2005-02-14 02:53 dims 
Revert changes for AXIS-1812
2005-02-14 00:42 dims 
Updated cvslog.html (differences between RC2 and RC3)
2005-02-12 21:58 dims 
Update to RC3
2005-02-12 18:04 dims 
Removing xsd:language, xsd:name from here as it breaks TCK tests.
2005-02-12 13:39 dims 
WSDL4J 1.5 + 1 late fix mentioned below

2005-02-12 07:09 wsdl4jadmin

Added base URI to InputSource so relative URIs get resolved when parsing the Document.
Made this change to readWSDL, parseImport, parseSchema. Also corrected related error
messages to use full URI where possible, not just the relative URI.
    * src/com/ibm/wsdl/xml/WSDLReaderImpl.java (1.38)