Convert XML with quotes (and other characters) in text to JSON encoding

[schallm]

I'm trying to convert XML (from a system I don't control) to JSON. The XML sometimes contains double quotes and other "special" characters in the text content. Dasel version 2.8.1 currently encodes:

• " as " where I would like it to be \"
• > as > where I would like it to be >

I assume this is html encoding? Is there an argument that controls this that I am missing?

Here is a sample string showing what I'm sending:

echo '<foo>"bar>baz"</foo>' | dasel.exe --read xml --write json

It returns:

{
  "foo": ""bar>""
}

I need it to return:

{
  "foo": "\"bar>baz\""
}

I'm running this on Windows if that matters.

[agammann]

I reproduced the HTML entities with v2.8.1. In v3.11.2, the same conversion preserves the actual string value. The v3 flags are -i and -o:

'<foo>"bar>baz"</foo>' | dasel -i xml -o json

The output is:

{
    "foo": "\"bar\u003ebaz\""
}

Here \u003e is a JSON escape for >. Parsing this JSON gives the string "bar>baz", including the two quote characters. The "/> spelling in v2 represents different string content after JSON parsing, which is why that output was troublesome.

I checked both releases and decoded the v3 JSON to verify the value. This gives the value you need, although the JSON serializer writes > as \u003e rather than the literal character.

The v3 JSON writer uses JSON string encoding.