Question about using res.end() vs. returning response object

[xia-stan]

In this comment on #497, @The-EDev states the following:

I can see a few issues with this implementation. For starters, if you're going to use a response it would be best to use ([](crow::response& res) { and res.end() rather than defining it later and then using return res.

I've seen some discussion around using res.end(), but nothing explicit around why this is preferred over returning a response object defined in the lambda. Could you provide some additional clarity around preferring res.end() over the returned object?

[The-EDev]

Hi, sorry for the late response. Ultimately the way Crow works is that res.end() triggers the data inside it to be sent to the client. when you return a response Crow is calling res.end() on it behind the scenes. The only reason to return a response is if you have some function that creates a response object. Otherwise it would save a few cpu cycles to just use the handler provided response object and call end() on that.