[psa] JSONObject.optString doesn't do what you think it does

Myk Melez myk at mozilla.org
Tue May 27 17:09:38 PDT 2014


> Jim Chen <mailto:nchen at mozilla.com>
> 2014 May 23 at 11:23
> I think a good compromise between correctness and practicality would be
> to treat null and undefined as equivalent in NativeJSObject. If people
> are expecting and using them as equivalents, we can just treat them as
> such.
JS code often treats them as equivalent, since they both evaluate to 
boolean false in a logical expression. But some of the JS code in our 
tree distinguishes between them, using the "in" operator (or comparing 
the property's value to the undefined value) to determine if a JS object 
has a property.

> This is really simple to do (one-line patch), and will make opt*
> return the fallback value for any null properties.
That seems reasonable for the opt* methods, which optimize for convenience.

> It will also let you
> use NativeJSObject.has() in place of isNull().
I would maintain the current behavior of these methods, so code that 
cares can distinguish between null and undefined.

Unlike opt*, the usage of these methods is reasonably low, at 37 and 20 
lines, respectively, via a primitive grep through mobile/android/base. 
So they could be audited in a reasonable amount of time to ensure their 
behavior is what their callers expect.

-myk

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/mobile-firefox-dev/attachments/20140527/0b9a7ab6/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: compose-unknown-contact.jpg
Type: image/jpeg
Size: 770 bytes
Desc: not available
URL: <http://mail.mozilla.org/pipermail/mobile-firefox-dev/attachments/20140527/0b9a7ab6/attachment.jpg>


More information about the mobile-firefox-dev mailing list