axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <stev...@iseran.com>
Subject Re: cvs commit: xml-axis/java/src/org/apache/axis AxisFault.java
Date Thu, 16 Jan 2003 00:24:07 GMT

----- Original Message -----
From: "Glen Daniels" <gdaniels@macromedia.com>
To: <axis-dev@xml.apache.org>
Sent: Wednesday, January 15, 2003 15:43
Subject: RE: cvs commit: xml-axis/java/src/org/apache/axis AxisFault.java


>
> +1 to the concept
>
> -0 to the implementation
>
> Wouldn't it be clearer to call it dumpToEncodedString() or somesuch and
avoid the boolean arg?

one thing is that there is some ugly hacker in the dumpToString() stuff to
deal with non namespaced stuff in the fault details (like, um, all the
strings we create), and I didnt want to replicate all of it...the only
difference between the normal and the encoded dump is a bit in the middle.

I thought about making the default action simply encode everything...we do
that with the fault details in the dumo already, after all; if that is the
action you'd prefer its trivial to do that.

The main place that we dump faults other than in the response to a GET is
debug traces. When there isnt stuff to escape in the fault string or
similar, encoding makes no visible difference, and so we have to ask 'do we
mind if log traces escape invalid XML chars?'. If the answer is no, then
I'll make escaping the only way to get a dump, and life is much simpler.

-steve


Mime
View raw message