You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@axis.apache.org by Sv...@gebit.de on 2008/10/09 22:00:55 UTC
Re: BUG [?] in BeanUtil
Hi,
i can confrm that this is indeed a bug in BeanUtil. There are already two
issues open for this (AXIS2-3538,AXIS2-3765)
and i would be really happy if someone would commit my proposed bugfix (
AXIS2-3538) since we are using a self-patched version in our project and
are going into qa shortly.
Best wishes,
Sven.
Andrea Spinelli <an...@imteam.it>
10.09.2008 16:06
Bitte antworten an
axis-user@ws.apache.org
An
axis-user@ws.apache.org
Kopie
Thema
BUG [?] in BeanUtil
Hi everybody,
our company is using axis2 to make available to multiple platforms a
document management service.
Documents are represented by Java beans, with a property of type byte[],
the content of the document. We have a method which sends a document from
the client to the server.
I noticed that the content of the document is received empty or garbled at
the server side. I googled a lot about this issue, but found nothing. So I
tried to debug the axis2 code (using Eclipse) and I think I isolated the
problem. The XML SOAP request contains a fragment like (simplified and
translated):
...<document><field1>foo</field1>... <documentText>...(base64 encoded
content>...</documentText> ... <lastField>...</lastField></document>...
The axis2 code is contained in the ADB module, in the class
org.apache.axis2.databinding.utils.BeanUtil. The program iterates over the
properties of the javabean "document" until it finds the property
"documentText"; it recognizes it is an array and calls recursively the
deserialize method:
} else if (parameters.isArray()) {
partObj = deserialize(parameters,
(OMElement)parts.getParent(),
objectSupplier, prty.getName());
Here "parts" is the element <documentText>, and its getParent() is
"<document>".
Then the following code is executed:
if (beanClass.isArray()) {
ArrayList valueList = new ArrayList();
Class arrayClassType = beanClass.getComponentType();
if ("byte".equals(arrayClassType.getName())) {
return
Base64.decode(beanElement.getFirstElement().getText());
Here "beanElement" is the second parameter of "deserialize", a structure
representing the whole <document> element.
So the Base64 decoder is called on the text of the first element of
"<document>", which is "foo" in the above example, and is actually the
value of field1 - not good.
I modified the code deleting the call to getParent (passing directly
"parts"), and calling the Base64 decoder on the text of the current
element:
return Base64.decode(beanElement.getText());
and it works fine for me, both from a .NET client and a axis1 client.
Moreover, all the tests of the ADB module are OK.
After all this explanation, my question is: is this a bug, or I
misunderstood something? If my modifications are actually a fix, how can I
submit it?
TIA
Andrea
--
dott. Andrea Spinelli - IMTeam
e-mail: andrea dot spinelli at imteam dot it
phone: +39-035-636029
fax: +39-035-638129
www: http://www.imteam.it/
IANA enterprise number: 1.3.6.1.4.1.30916
--------------------------------------------------------------------- To
unsubscribe, e-mail: axis-user-unsubscribe@ws.apache.org For additional
commands, e-mail: axis-user-help@ws.apache.org