Prevent encoding of #text values when XML is built

[gavinharriss]

Is there any way to turn off the encoding of #text elements when XML is being built from JSON?

I have a scenario where the destination system will take the raw value without unencoding text, so the encoded text shows up in the UI still encoded. This is beyond my control.

I'm wanting to capture a JSON representation of the raw source data for audit purposes. Though any #text field I use elsewhere is at risk of the same issue with user inputted values.

From this:

{
    "Note": {
        "Description": {
            "#text": "Raw source data"
        },
        "NoteText": {
            "#text": JSON.stringify(obj)
        }
    }
  }

The XMLBuilder outputs:

<Note>
	<Description>Raw source data</Description>
	<NoteText>{
  &quot;SomeProperty&quot;: {
    &quot;ChildProp&quot;: {
      &quot;x&quot;: 123,
      &quot;y&quot;: &quot;qwerty&quot;,
      &quot;z&quot;: &quot;xxx&quot;
    }
  }
}</NoteText>
</Note>

But what I need is:

<Note>
	<Description>Raw source data</Description>
	<NoteText>{
  "SomeProperty": {
    "ChildProp": {
      "x": 123,
      "y": "qwerty",
      "z": "xxx"
    }
  }
}</NoteText>
</Note>

Is there a way to achieve this?

I've tried experimenting with the <![CDATA[]]> tag, but unfortunately the destinations system doesn't care much for the CDATA tag.

[amitguptagwl]

What I understand here is that builder is not encoding the data. Can you please hardcode the value instead of JSON.stringify and debug?

[gavinharriss]

Thanks for taking a look, much appreciated :)

I've just tried passing const chars = '"&<>' in instead and the resulting output is <NoteText>&quot;&amp;&lt;&gt;</NoteText>, so it's definitely encoding the #text value.

[gavinharriss]

I did stumble across the XmlBuilderOptions tagValueProcessor option, but experimenting with this it only deals with the pre-unencoded value.

[gavinharriss]

I believe I've found what I need = setting the processEntities: false builder option. And if I need fine grained control around what chars are encoded it looks like I can use the entities setting.

[amitguptagwl]

This will help you to solve your problem.

https://github.com/NaturalIntelligence/fast-xml-parser/blob/master/docs/v4/5.Entities.md

[gavinharriss]

Unfortunately entities isn't exposed via the typescript definitions, but this is what worked for me in case anyone else has the same issue:

this.builder = new XMLBuilder({
    ignoreAttributes: false,
    // CargoWise only likes < and > chars encoded in text, so override encoding and provide our own encoder
    // NOTE: There appears to be an `entities` setting in js where we can set explicit chars to be encoded,
    //       but this doesn't seem to be exposed via the ts type definitions. 
    processEntities: false,
    tagValueProcessor: (tagName, a) => { 
        if (a) {
            return a.toString()
                .replace(/</g, '&lt;')
                .replace(/>/g, '&gt;');
        }
        return a; 
    },
});

[amitguptagwl]

Yes, it is somehow missed from typings.