Chinmoy, I would say that in the case of MTOM (SwA is different), it is conceptually incorrect to assume that the DataSource name and content type are preserved when the message is sent over the wire. The reason is that MTOM defines a transformation between two different representations of the same information model: one that encodes binary data as base64 stored in a text node and one that encodes binary data as MIME parts. Since in the first representation the name and content type are unspecified, it would be wrong to assume that they are meaningful in the second representation (otherwise the representations would not be equivalent). In MTOM, content types (for the binary data) only appear because the message format is MIME, which uses content types. I guess that if the people who developed the MTOM specs had decided not to use MIME, they would not have introduced the concept of content types.
Andreas On Thu, Mar 26, 2009 at 14:13, Chinmoy Chakraborty <[email protected]> wrote: > Andreas, > > Thanks for your reply. Actually I am trying to invoke a service which > returns a DataHandler through MTOM. I have following code in the service > class: > > FileItem file = (FileItem) value; > ByteArrayDataSource bads = new ByteArrayDataSource(file.getInputStream(), > "application/msword"); > bads.setName("abc.doc"); > return new DataHandler(bads); > > > So far so good. The newly created DataHandler preserve the name. I have > following code in the client end from where I am invoking the service: > > ServiceClient sender = new ServiceClient(); > sender.setOptions(options); > OMFactory fac = OMAbstractFactory.getOMFactory(); > OMNamespace ns = > fac.createOMNamespace("http://www.abc.com/webservice";, "ns"); > OMElement payload = createPayload(ns, argNameList, args); > OMElement retVal = sender.sendReceive(payload); > if (retVal != null) { > OMElement ele = retVal.getFirstElement(); > OMText binaryNode = (OMText) ele.getFirstOMChild(); > binaryNode.setOptimize(true); > return (DataHandler) binaryNode.getDataHandler(); > } > > I am trying to return a .doc file. When I receive the datahandler in the > client end DataHandler#getName() returns "MyByteArrayDataSource" and the > content type becomes "application/octet-stream" meaning binary. > > Hope I made you understand the problem. May be the code > "binaryNode.setOptimize(true);" made the content-type binary. Then how to > get a .doc or other file type? > > Chinmoy > > > > > > On Wed, Mar 25, 2009 at 1:29 PM, Chinmoy Chakraborty <[email protected]> > wrote: >> >> Hi All, >> >> I noticed DataHandler.getName() always returns 'MyByteArrayDataSource' if >> the data source is ByteArrayDataSource from which the DataHandler has been >> created. Please look at the code below: >> >> FileItem file = (FileItem) value; >> ByteArrayDataSource bads = new ByteArrayDataSource(file.getInputStream(), >> "application/msword"); >> bads.setName("abc.doc"); >> DataHandler dh = new DataHandler(bads); >> >> Now if I do, dh.getName(), it always returns 'MyByteArrayDataSource". I >> guess this is a bug. Whats your opinion? >> >> >> Chinmoy >
