fetch resolves even if 404?
Use this code...
fetch(`https://fercarvo.github.io/apps/integradora/DB/corpus.json`)
.then(async (data) => {
if (data.ok) {
data = await data.json()
//Here you have your data...
}
}).catch(e => console.log('Connection error', e))
A fetch()
call is only rejected if the network request itself fails for some reason (host not found, no connection, server not responding, etc...).
Any result back from the server (404, 500, etc...) is considered a successful request from the promise point of view. Conceptually, you made a request from the server and the server answered you so from the networking point of view, the request finished successfully.
You need to then test that successful response to see if has the type of answer you wanted. If you want a 404 to be a rejection, you could code that yourself:
fetch('notExists').then(function(response) {
if (!response.ok) {
// make the promise be rejected if we didn't get a 2xx response
throw new Error("Not 2xx response", {cause: response});
} else {
// got the desired response
}
}).catch(function(err) {
// some error here
});
You could even make your own myFetch()
that just does this automatically for you (converts any non-200 response status to a rejection).
What is the reason behind a resolved promise for a a completely bad request ( non existing resource / server down).
First off, server down will not generate a successful response - that will reject.
A successful response is generated if you successfully connect to the server, send it a request and it returns a response (any response). As for "why" the designers of the fetch()
interface decided to base the rejection on this, it's a bit hard to say without talking to someone who was actually involved in the design of that interface, but it seems logical to me. This way the rejection tells you whether the request got through and got a valid response. It's up to your code to decide what to do with the response. You can, of course, create your own wrapper function that modifies that default behavior.