Johannes Waldmann wrote:

with haxr-20050621, I have trouble interfacing to Java.

When I build an XML RPC call with haxr, containing a struct, I get

<?xml version='1.0' ?>

...

<methodName >vorlesungen

but when I build the same call with org.apache.xmlrpc, I get

<?xml version="1.0"?><methodCall><methodName>vorlesungen</methodName> <params><param><value><struct><member><name>matrikel</name><value>531</value></member> <member><name>schule</name><value>HTWK</value></member> ...

the difference is that haxr has extra <string> tags for values of arguments. I don't think this is right?

Best regards,

Ah, I seem to have missed one important sentence in the XML-RPC spec [1]: "If no type is indicated, the type is string.". The way I understand the spec, having the <string>...</string> is optional. So what HAXR produces is valid and should be accepted by any implementation. However, I think HaXR won't accept a call without an explicit type tag. In other words, I would expect HaXR's calls to be understood by anyone, but that HaXR won't understand all valid calls. Is that what happens? /Björn [1] http://www.xmlrpc.com/spec

Johannes Waldmann wrote:

with haxr-20050621, I have trouble interfacing to Java.

When I build an XML RPC call with haxr, containing a struct, I get

<?xml version='1.0' ?>

...

<methodName >vorlesungen

but when I build the same call with org.apache.xmlrpc, I get

<?xml version="1.0"?><methodCall><methodName>vorlesungen</methodName> <params><param><value><struct><member><name>matrikel</name><value>531</value></member> <member><name>schule</name><value>HTWK</value></member> ...

the difference is that haxr has extra <string> tags for values of arguments. I don't think this is right?

There is an fix in darcs now. It's untested; I'll test it as soon as I get a test environment set up. Let me know how it works out for you. Thanks to Malcom Wallace for suggesting a low-resistance way to fix this. /Björn