Kay Kay - 6 months ago 26
JSON Question

json.dumps(): escaping forward slashes

Since forward slashes can only occur in strings inside a JSON serialized object and are not escaped (in the default settings), using

json.dump(some_dict).replace('/', r'\/')


reliably works, but it looks hacky.

I know that forward slashes don't have to be escaped, but you may escape them, and for my usecase I'd like to have them escaped.

Is there a way to to let the JSONEncoder escape forward slashes without manually escaping them?

Answer

Only escape forward slashes when encode_html_chars=True

Check out this- https://github.com/esnme/ultrajson/pull/114

The JSON spec says forward slahes shall be escaped implicitly.

Here is a solution to do it in JSONEncoder itself. Its just that you create an ESCAPE DICTIONARY and do computation before hand and do the encoding later.

https://chromium.googlesource.com/external/googleappengine/python/+/dc33addea2da464ca07e869cb11832e1ae82da9d/lib/django/django/utils/simplejson/encoder.py

Hope it helps.

-

Adding to the above solution, there is another reason to escape the characters. As kay said, it gives us some extra sleep. It prevents the attack. So the solution above takes care of all issues.

ESCAPE_DCT = {
    # escape all forward slashes to prevent </script> attack
    '/': '\\/',
    '\\': '\\\\',
    '"': '\\"',
    '\b': '\\b',
    '\f': '\\f',
    '\n': '\\n',
    '\r': '\\r',
    '\t': '\\t',
}
Comments