Il have many proxies uri (http and socks5) and i'm using reqwest for sending some http request throught theses proxies and i wanna drop a proxy if he didn't work.
for proxy in proxies {
let proxy = match Proxy::all(proxy_url) {
Ok(proxy) => proxy,
Err(e) => {
eprintln!("creating proxy failed! {:?}", e);
continue;
}
};
let client = match Client::builder()
.proxy(proxy)
.build()
let client = match client {
Ok(client) => client,
Err(e) => {
eprintln!("building client failed! {:?}", e);
continue;
}
}
loop {
match client.get(&config.endpoint).send().await {
Ok(_) => {}
Err(e) => {
if e.is_proxy_error() { // this method doesn't exist
eprintln!("Dropping proxy...");
break;
} else {
eprintln!("client error! {:?}", e);
}
}
}
}
}
And i got many kind of Reqwest::Error
reqwest::Error { kind: Request, url: "http://example.com/", source: hyper::Error(Connect, "socks connect error: Invalid response version") }
reqwest::Error { kind: Request, url: "http://example.com/", source: hyper::Error(Connect, "socks connect error: Proxy server unreachable") }
reqwest::Error { kind: Request, url: "http://example.com/", source: hyper::Error(Connect, "socks connect error: Connection refused") }
reqwest::Error { kind: Request, url: "http://example.com/", source: hyper::Error(Connect, "socks connect error: Connection reset by peer (os error 104)") }
The error message is explicit in most case, but how can i handle each of them differently ?
The reqwest::Error have inner field which is private so i cant access him.
And if i get the source of the reqwest::Error, i just have an Box<syn Error> which i cant treat like a hyper::Error
You can downcast the result of Error::source() to a concrete error type, e.g.
use std::error::Error;
...
Err(reqwest_error) => {
let hyper_error: Option<&hyper::Error> = reqwest_error
.source()
.unwrap()
.downcast_ref();
...
}
This code uses unwrap() to extract the error from the Option returned by source(). If you can't guarantee that the error will always have a source, you should use some conditional way to unwrap the source, e.g.
let hyper_error: Option<&hyper::Error> = reqwest_error
.source()
.and_then(|e| e.downcast_ref());
Related
I have set up a proxy in my NestJs Application to bypass the CORS restrictions.
The EP for the service is -
https://api.niftyuat.com/cors?cd=https://google.com
And it will give me the body of the google home page hence bypassing the CORS restriction.
But the issue is, when I try to download any image or some file, it gives me some kind of encoded input.
https://api.niftyuat.com/cors?cd=https://niftypm-production-assets-user.s3-accelerate.amazonaws.com/a895af03-aa9e-4b60-a3b4-54776b240c3e-Screen%20Shot%202021-03-18%20at%2012.03.24%20PM.png
I am also sharing my code.
#Controller('cors')
export class CorsController {
constructor(private readonly corsService: CorsService) { }
#Get()
async get(#Query('cd') uri: string, #Response() res: ExpressResponse) {
try {
let url = uri.charAt(0) == '/' ? uri.substring(1) : uri;
const response: AxiosResponse = await axios.get(url, {
timeout: 5000,
})
res.writeHead(200, { ...response.headers })
res.write(JSON.stringify(response.data))
res.end()
} catch (error) {
return error.message
}
}
}
Any idea what I am doing wrong here?
You forgot to set Content-Type header in response. Set it to 'image/png' or 'image/x-png'.
I try to use RSocketRequester to send a message from the server to the specific client, but I don't know how to handle it on the frontend. The server is Spring Webflux with the controller like this:
data class Message(val message: String)
#Controller
class RSocketController {
private val log = LoggerFactory.getLogger(RSocketController::class.java)
#MessageMapping("say.hello")
fun sayHello(message: String): Flux<Message> {
log.info("say hello {}", message)
return Flux.just(Message("server says hello"))
}
#MessageMapping("say.hi")
fun sayHi(message: String, rSocketRequester: RSocketRequester): Flux<Message> {
log.info("say hi {}", message)
rSocketRequester
.route("say.hello")
.data(Message("server says hi hello ;)"))
.send()
.subscribe()
return Flux.just(Message("server says hi!!"))
}
}
On the frontend I use rsocket-js. The sayHello method works just fine (request-stream), but when I call the sayHi method I want to send two messages from the server. The first one to say.hello endpoint, and the second to say.hi endpoint. I've got rsocket-js implementation like this:
sayHello() {
console.log("say hello");
this.requestStream("say.hello");
},
sayHi() {
console.log("say hi");
this.requestStream("say.hi");
},
connect() {
const transport = new RSocketWebSocketClient({
url: "ws://localhost:8080/rsocket"
});
const client = new RSocketClient({
serializers: {
data: JsonSerializer,
metadata: IdentitySerializer
},
setup: {
keepAlive: 60000,
lifetime: 180000,
dataMimeType: "application/json",
metadataMimeType: "message/x.rsocket.routing.v0"
},
transport
});
client.connect().subscribe({
onComplete: socket => {
this.socket = socket;
console.log("complete connection");
},
onError: error => {
console.log("got connection error");
console.error(error);
},
onSubscribe: cancel => {
console.log("subscribe connection");
console.log(cancel);
}
});
},
requestStream(url) {
if (this.socket) {
this.socket
.requestStream({
data: url + " from client",
metadata: String.fromCharCode(url.length) + url
})
.subscribe({
onComplete: () => console.log("requestStream done"),
onError: error => {
console.log("got error with requestStream");
console.error(error);
},
onNext: value => {
// console.log("got next value in requestStream..");
console.log("got data from sever");
console.log(value.data);
},
// Nothing happens until `request(n)` is called
onSubscribe: sub => {
console.log("subscribe request Stream!");
sub.request(2147483647);
// sub.request(3);
}
});
} else {
console.log("not connected...");
}
}
I can see both messages in Google Chrome DevTools -> Network -> rsocket. So the client receives them but I can't catch in the code the one sent by RSocketRequester.
It seems that the server uses fireAndForget method. How to handle it on the client side?
As #VladMamaev said, we can provide a responder to the client like in this example https://github.com/rsocket/rsocket-js/blob/master/packages/rsocket-examples/src/LeaseClientExample.js#L104
For me, fireAndForget method is enough.
export class EchoResponder {
constructor(callback) {
this.callback = callback;
}
fireAndForget(payload) {
this.callback(payload);
}
}
import { EchoResponder } from "~/assets/EchoResponder";
...
const messageReceiver = payload => {
//do what you want to do with received message
console.log(payload)
};
const responder = new EchoResponder(messageReceiver);
connect() {
const transport = new RSocketWebSocketClient({
url: "ws://localhost:8080/rsocket"
});
const client = new RSocketClient({
serializers: {
data: JsonSerializer,
metadata: IdentitySerializer
},
setup: {
keepAlive: 60000,
lifetime: 180000,
dataMimeType: "application/json",
metadataMimeType: "message/x.rsocket.routing.v0"
},
responder: responder,
transport
});
I have a controller method
// ... inside a controller class
#get('/error', {})
async error() {
throw new Error("This is the error text");
}
The response I'm getting from this error front-end is:
{
"error": {
"statusCode": 500,
"message": "Internal Server Error"
}
}
What I would like the error to be is:
{
"error": {
"statusCode": 500,
"message": "This is the error text"
}
}
How do I return an error from a controller in Loopback 4?
Hello from the LoopBack team 👋
In your controller or repository, you should throw the Error exactly as shown in your question.
Now when LoopBack catches an error, it invokes reject action to handle it. The built-in implementation of reject logs a message via console.error and returns an HTTP response with 4xx/5xx error code and response body describing the error.
By default, LoopBack hides the actual error messages in HTTP responses. This is a security measure preventing the server from leaking potentially sensitive data (paths to files that could not be opened, IP addresses of backend service that could not be reached).
Under the hood, we use strong-error-handler to convert Error objects to HTTP responses. This module offers two modes:
Production mode (the default): 5xx errors don't include any additional information, 4xx errors include partial information.
Debug mode (debug: true): all error details are included on the response, including a full stack trace.
The debug mode can be enabled by adding the following line to your Application constructor:
this.bind(RestBindings.ERROR_WRITER_OPTIONS).to({debug: true});
Learn more in our docs: Sequence >> Handling errors
Alternatively, you can implement your own error handler and bind it as the sequence action reject. See Customizing sequence actions in our docs.
export class MyRejectProvider implements Provider<Reject> {
constructor(
#inject(RestBindings.SequenceActions.LOG_ERROR)
protected logError: LogError,
#inject(RestBindings.ERROR_WRITER_OPTIONS, {optional: true})
protected errorWriterOptions?: ErrorWriterOptions,
) {}
value(): Reject {
return (context, error) => this.action(context, error);
}
action({request, response}: HandlerContext, error: Error) {
const err = <HttpError>error;
const statusCode = err.statusCode || err.status || 500;
const body = // convert err to plain data object
res.statusCode = statusCode;
res.setHeader('Content-Type', 'application/json; charset=utf-8');
res.end(JSON.stringify(body), 'utf-8');
this.logError(error, statusCode, request);
}
}
If you just want to show error message, you just extend Error object and throw it like below. (Loopback documentation didn't mention this anyway)
Avoid using 5xx error and use 4xx error to show some important thing to user is best practice and so that Loopback4 was implemented like this.
class NotFound extends Error {
statusCode: number
constructor(message: string) {
super(message)
this.statusCode = 404
}
}
...
if (!await this.userRepository.exists(id)) {
throw new NotFound('user not found')
}
For my situation, I found a catch in my sequence.ts file. Inside the catch, it checked if the error had a status code of 4xx, and if not, it just returned a anonymous 500.
Here's the code I was looking for to do the logic:
// sequence.ts
...
} catch (err) {
console.log(err);
let code: string = (err.code || 500).toString();
if (code.length && code[0] === '4') {
response.status(Number(code) || 500);
return this.send(response, {
error: {
message: err.message,
name: err.name || 'UnknownError',
statusCode: code
}
});
}
return this.reject(context, err);
}
...
Here's how you tell it what to do:
// ... inside a controller class
#get('/error', {})
async error() {
throw {
code: 400,
message: "This is the error text",
name: "IntentionalError"
}
}
To throw custom validation error I use this method:
private static createError(msg: string, name?: string): HttpErrors.HttpError {
const error = new HttpErrors['422'](msg);
error.name = name ?? this.name;
return error;
}
Catch error examples here are for defaultSequence, overriding the handle method.
But nowdays app template uses MiddlewareSequence.
So here is the example, how tomodify the response in middleware sequence, you can use this example:
import { Middleware, MiddlewareContext } from '#loopback/rest';
export const ErrorMiddleware: Middleware = async (middlewareCtx: MiddlewareContext, next) => {
// const {response} = middlewareCtx;
try {
// Proceed with next middleware
return await next();
} catch (err) {
// Catch errors from downstream middleware
// How to catch specific error and how to send custom error response:
if (HttpErrors.isHttpError(err) || (err as HttpErrors.HttpError).statusCode) {
const code: string = (err.statusCode || 500).toString();
if (code.length && code[0] === '4') {
response.status(Number(code) || 500);
return response.send({
error: {
message: err.message,
name: err.name || 'UnknownError',
statusCode: code
}
});
}
}
throw err;
}
};
And register the middleware in application.ts
this.middleware(ErrorMiddleware);
I use the following function to Post a object of a given class.
public Post<T>(object: T, url: string, httpOptions: {}): Observable<T> {
return this.httpClient.post<T>(`${environment.apiEndpoint}` + url, object, httpOptions)
.pipe(
catchError(this.handleError)
);
}
This function is called in all the service that wants to post something. Like this.
public addEquipment(equipment: Equipment): Observable<Equipment> {
return this.shared.Post<Equipment>(equipment, this.url, this.header);
}
addEquipment is then executed within the component that uses that service. Like this.
this.equipmentService.addEquipment(result)
.subscribe((data: any) => { this.alertService.success(data) }, (error: any) => this.alertService.error(error));
The problem is when the API returns a error (that I can see includes a error message, in the network tab) it tells me that there is no body in the response. The API returns a HttpResult where the error message is added to the response field.
return new HttpResult { StatusCode = HttpStatusCode.Conflict, Response = "Error message"}
I use the following function to handle the errors.
private handleError(error: HttpErrorResponse) {
if (error.error instanceof ErrorEvent) {
// A client-side or network error occurred. Handle it accordingly.
console.error('An error occurred:', error.error.message);
}
else {
console.log(error);
console.error(
`Backend returned code ${error.status}, ` +
`body was: ${error.error}`);
}
console.log(error);
return throwError(
error.error)
};
It is Angular 6 and a ServiceStack API.
All suggestions would be appreciated.
FYI it's preferable to return structured error responses in ServiceStack which you can do with:
HttpError.Conflict("Error message");
Which will let you catch it when using ServiceStack's TypeScript ServiceClient with:
try {
var response = await client.post(request);
} catch (e) {
console.log(e.responseStatus.message);
}
But from this answer for handling errors with Angular HTTP Client it suggests the error body should be accessible with:
this.httpClient
.get("data-url")
.catch((err: HttpErrorResponse) => {
// simple logging, but you can do a lot more, see below
console.error('An error occurred:', err.error);
});
I created 2 servers here
Router router = Router.router(vertx);
router.route().handler(BodyHandler.create());
router.post("/api/upload").handler(routingContext -> {
System.out.println(routingContext.fileUploads().size());
routingContext.response().end();
});
vertx.createHttpServer().requestHandler(req -> {
router.accept(req);
}).listen(8080, listenResult -> {
if (listenResult.failed()) {
System.out.println("Could not start HTTP server");
listenResult.cause().printStackTrace();
} else {
System.out.println("Server started");
}
});
// ==========================================
vertx.createHttpServer().requestHandler(req -> {
req.bodyHandler(buff -> {
System.out.println(buff.toString() + " from client");
req.response().end();
});
}).listen(8081, listenResult -> {
if (listenResult.failed()) {
System.out.println("Could not start HTTP server");
listenResult.cause().printStackTrace();
} else {
System.out.println("Server started");
}
});
The 1st one is from vertx documentation.
The 2nd one is from https://github.com/vert-x3/vertx-examples/blob/master/web-client-examples/src/main/java/io/vertx/example/webclient/send/stream/Server.java
When tested with Postman, both works.
When tested with other front-end codes, (example: https://github.com/BBB/dropzone-redux-form-example), only 2nd server works.
This is what I updated on the above github example.
fetch(`http://localhost:8081/api/upload`, {
method: 'POST',
headers: {
},
body: body,
})
.then(res => {
console.log('response status: ', res.statusText);
return res.json();
})
.then(res => console.log(res))
.catch(err => {
console.log("An error occurred");
console.error(err);
});
In practice, I prefer to use the approach to 1st server.
Since both are tested by Postman, I believe server is not an issue, and need to tweak on the client side.
Can anyone point out what I should be adding to the client?
Thanks.
Edit
axios.post('http://localhost:50123/api/upload', fileData)
.then(response => {
console.log('got response');
console.dir(response);
})
.catch(err => {
console.log("Error occurred");
console.dir(err);
});
axios works when passing a file from frontend.
Now the problem is unit-test using Vertx Web Client.
fs.open("content.txt", new OpenOptions(), fileRes -> {
if (fileRes.succeeded()) {
ReadStream<Buffer> fileStream = fileRes.result();
String fileLen = "1024";
// Send the file to the server using POST
client
.post(8080, "myserver.mycompany.com", "/some-uri")
.putHeader("content-length", fileLen)
.sendStream(fileStream, ar -> {
if (ar.succeeded()) {
// Ok
}
});
}
});
The above code from http://vertx.io/docs/vertx-web-client/java/#_writing_request_bodies doesn't work for 1st server. FileUploads is empty.
It works for 2nd.
Edit2
I decided to use a simple HttpClient code, and it works as well.
How can I make a multipart/form-data POST request using Java?
CloseableHttpClient httpClient = HttpClients.createDefault();
HttpPost uploadFile = new HttpPost("http://localhost:8080/upload");
MultipartEntityBuilder builder = MultipartEntityBuilder.create();
builder.addTextBody("field1", "yes", ContentType.TEXT_PLAIN);
// This attaches the file to the POST:
File f = new File("./test.txt");
builder.addBinaryBody(
"file",
new FileInputStream(f),
ContentType.APPLICATION_OCTET_STREAM,
f.getName()
);
HttpEntity multipart = builder.build();
uploadFile.setEntity(multipart);
CloseableHttpResponse response = httpClient.execute(uploadFile);
HttpEntity responseEntity = response.getEntity();
System.out.println(responseEntity.toString());
I don't see how your last example could work. You post to http://localhost:8080/upload, but your route is /api/upload. In your second example, with port 8081 you simply ignore the route, and assume that anything you receive is a file upload. That's the only reason second example "works".