org.json.XML.unescape("
"); throws org.json.JSONException: Invalid numeric character reference: 


[S-wald]

org.json.XML.unescape("
"); throws org.json.JSONException: Invalid numeric character reference: 
 in version 20260522

Works in version 20251224.

[stleary]

@S-wald Thanks for submitting this issue.

Alternate form of the issue:
// throws an exception after v20260522, but used to work in v20251224
XML.toJSONObject("<a>&#10;</a>");

In XML, the &#...; sequence defines a codepoint, not a numeric value. The codepoint 10 is a linefeed, which is not a valid JSON character. I believe it was fixed correctly in #1045.
However, a mechanism can be provided to support legacy code that relies on the v20251224 behavior, for the special case of codepoints that evaluate to one of the common whitespace chars (TAB, LF, CR). Will this work for you?

[S-wald]

@stleary IMHO org.json.XML.unescape should just unescape XML and output the whitespace char (e.g. '\n'). After that it may be escaped for JSON again.

[stleary]

v20251224 behavior:
Given: XML.toJSONObject("<a>&#10;</a>");


 is parsed to int 10 and converted to a string containing a linefeed, this value is stripped when trim() is called, so the resulting value is an empty string.

@S-wald What is your use case? Is your code calling XML.unescape() directly?

[S-wald]

We use XML.toJSONObject to convert some XML we get from some other system component.