SimpleXmlRpcServer and character encoding

[shymon]

I'm using SimpleXmlRpcServer class. Although I set encoding parameter in the
constructor, I have to return all strings in default platform encoding
(windows-1250/win32 or iso-8859-2/linux in my case). When I send values in,
for example, UTF-8, string received by client is messed up.

The client is written in java using Apache XmlRpc library 2.0.

Is there any solution other than sending all string values in Base64
encoding?

 

[Diez B. Roggisch]

I'm using SimpleXmlRpcServer class. Although I set encoding parameter in
the constructor, I have to return all strings in default platform encoding
(windows-1250/win32 or iso-8859-2/linux in my case). When I send values
in, for example, UTF-8, string received by client is messed up.

The client is written in java using Apache XmlRpc library 2.0.

Is there any solution other than sending all string values in Base64
encoding?

Use unicode-objects. And unicode IS NOT utf-8. The encoding parameter will
affect the xml generated & send over the wire - *not* what strings you
pass/return to your implementation.

So I think you should remove the encoding parameter alltogether, as this
will make the transport being in utf-8. Then use only unicode-objects in
your python code. And on the java-side, things *should* be in order.

Diez

 

[shymon]

Use unicode-objects. And unicode IS NOT utf-8. The encoding parameter will
affect the xml generated & send over the wire - *not* what strings you
pass/return to your implementation.

So I think you should remove the encoding parameter alltogether, as this
will make the transport being in utf-8. Then use only unicode-objects in
your python code. And on the java-side, things *should* be in order.

Diez

I have tried unicode strings also, let's say u"miłość".
Result received by the client was the same as if I sent UTF-8 encoded
string.

 

[Diez B. Roggisch]

I have tried unicode strings also, let's say u"miłość".
Result received by the client was the same as if I sent UTF-8 encoded
string.

Please show concrete code-examples, and the results, preferably as
repr-string, as these contain the hex-chars:

repr("ö") -> "'\\xc3\\xb6'"

The above shows that the ö is encoded in utf-8.

Diez