Client "Can you change your API? If a POST return 201 ours system crashes reading it"
Me "Your system WHAT?????"

Bahaha

My reply would be "Change your system then, not gonna break it for everyone else"

Ya filter is ip adresse and return {"message":"fuck you"}

I’ve had this before, the whole “everything must return 200, or our system automatically elevates it to a full on panic-mode error; just send us a 200 response with an error field instead” *bangs head*

@Brolls
The contractor dev at work (the only other dev...) insists on this. It drives me crazy!

Though he had any okay reason: the previous dev would say APIs were finished, but in reality hadn't even started, so 404s could indicate either a missing record or a missing API. I can see that getting very infuriating very quickly.

@Root yeah, the missing endpoint vs missing resource is a tricky one to solve; though I suppose the REST standard takes the position that as urls should map to resources, it’s essentially the same error (though not really).

@Root I guess you can always add an error message to differentiate them

@Brolls Yeah. But professionalism and documentation solve that 😋. Also my preferred method of adding a /help to the end of the API URLs that document them. If the help isn't present, the endpoint isn't either.

E.g. /api/v4/gift/help shows data about that endpoint and all child endpoints.

/api/v4/help therefore would list documentation about all v4 APIs.

(I may start using/docs instead.)

@guio I do.

@Root nice! Using /help or /docs sounds like a good way to go

@Root What about 501 Not Implemented?

Holy shit. They have no competent dev OR QA.

@k0pernikus That works as well 😊
But the route would only return that if I've already begun working on the api. If I haven't, it'd still 404.