getting json_encode to not escape html entities

Question

I send json_encoded data from my PHP server to iPhone app. Strings containing html entities, like '&' are escaped by json_encode and sent as &.

I am looking to do one of two things:

• make json_encode not escape html entities. Doc says 'normal' mode shouldn't escape it but it doesn't work for me. Any ideas?

• make the iPhone app un-escape html entities cheaply. The only way I can think of doing it now involves spinning up a XML/HTML parser which is very expensive. Any cheaper suggestions?

Thanks!

Solution 1

Neither PHP 5.3 nor PHP 5.2 touch the HTML entities.

You can test this with the following code:

<?php
header("Content-type: text/plain"); //makes sure entities are not interpreted
$s = 'A string with &amp; &#x6F8 entities';
echo json_encode($s);

You'll see the only thing PHP does is to add double quotes around the string.

Solution 2

json_encode does not do that. You have another component that is doing the HTML encoding.

If you use the JSON_HEX_ options you can avoid that any < or & characters appear in the output (they'd get converted to \u003C or similar JS string literal escapes), thus possibly avoiding the problem:

json_encode($s, JSON_HEX_TAG|JSON_HEX_AMP|JSON_HEX_QUOT)

though this would depend on knowing exactly which characters are being HTML-encoded further downstream. Maybe non-ASCII characters too?

Solution 3

Based on the manual it appears that json_encode shouldn't be escaping your entities, unless you explicitly tell it to, in PHP 5.3. Are you perhaps running an older version of PHP?

Comments

[psychotik]

Yea, I'm using 5.2.6 So pre-5.3 json_encode didn't escape html entities by default?