Should we keep the object prototype in req.body and similar fields

[expressjs]

Should we keep the object prototype in req.body and similar fields

[blakeembrey]

I think using prototype-less objects is definitely better and it avoid having odd gotchas in people's code as handling it would need to be done consistently (E.g. it either exists or it doesn't, it's not maybe a function, maybe something the user passed). That said, it's tricky because a lot of people do rely on it. With the introduction of node 6 setting the precedent (removing the object prototype from headers), I hope more people will be writing prototype-less objects for user input. Native methods like JSON.parse may be an issue though, so we need to figure out how to handle that - maybe we can move to Object.setPrototypeOf (with polyfill?).

[blakeembrey]

Also, FWIW, urlencoded will also have a null prototype in node 6 since they patched that with the headers. So technically that comment is incorrect, which just means we should be making it the most common denominator already (https://github.com/nodejs/node/blob/master/lib/querystring.js#L8-L12).

[dougwilson]

I am in general undecided on this. I personally lean more towards a null prototype, since it more accurately describes the type of object, but at the same time, it's actually pretty hard to interact with those types of objects in some cases. I feel like the main argument for the null prototype is to just to get developers to not even be able to make a dumb mistake when initially writing the code, but IDK

[LinusU]

This aws discussed at the Express TC Meeting 2016-08-10 and the consensus was to keep the prototype as null.