How do I include &, <, > etc in XML attribute values

I want to create an XML file which will be used to store the structure of a Java program. I am able to successfully parse the Java program and create the tags as required. The problem arises when I try to include the source code inside my tags, since Java source code may use a vast number of entity reference and reserved characters like &, < ,> , &. I am not able to create a valid XML.

My XML should go like this:

<?xml version="1.0"?> <prg name="prg_name"> <class name= "class_name> <parent>parent class</parent> <interface>Interface name</interface> . . . <method name= "method_name"> <statement>the ordinary java statement</statement> <if condition="Conditional Expression"> <statement> true statements </statement> </if> <else> <statement> false statements </statement> </else> <statement> usual control statements </statement> . . . </method> </class> . . . </prg> 

Like this, but the problem is conditional expressions of if or other statements have a lot of & or other reserved symbols in them which prevents XML from getting validated. Since all this data (source code) is given by the user I have little control over it. Escaping the characters will be very costly in terms of time.

I can use CDATA to escape the element text but it can not be used for attribute values containing conditional expressions. I am using Antlr Java grammar to parse the Java program and getting the attributes and content for the tags. So is there any other workaround for it?

2 Answers

You will have to escape

" to &quot; ' to &apos; < to &lt; > to &gt; & to &amp; 

for xml.

In XML attributes you must escape

" with &quot; < with &lt; & with &amp; 

if you wrap attribute values in double quotes ("), e.g.

<MyTag attr="If a&lt;b &amp; b&lt;c then a&lt;c, it's obvious"/> 

meaning tag MyTag with attribute attr with text If a<b & b<c then a<c, it's obvious - note: no need to use ' to escape ' character.

If you wrap attribute values in single quotes (') then you should escape these characters:

' with &apos; < with &lt; & with &amp; 

and you can write " as is. Escaping of > with > in attribute text is not required, e.g. <a b=">"/> is well-formed XML.