Bug Report

| Q | A | | ------------------- | ------ | | BC Break | no | | Version | 1.5.2 |

Summary

It seems that there is still something going on with HTTPMetrics in the 1.5.2 version. I can't reproduce it since I don't really know what's going on, but our crash report shows a considerable amount of these issues...

FYI, we use the Sentry and NewRelic SDKs and I am not sure if they could interfere with RealHTTP somehow 🤔

OS Version: iOS 15.5 (19F77)
Report Version: 104

Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: BUS_NOOP at 0x0000000000000419
Crashed Thread: 1

Application Specific Information:
addObject: > countByEnumeratingWithState:objects:count: > transactionMetrics >
Attempted to dereference garbage pointer 0x419.

Thread 1 Crashed:
0   libobjc.A.dylib                 0x33045f118         objc_retain
1   CFNetwork                       0x302374644         CFURLRequestSetShouldStartSynchronously
2   CFNetwork                       0x30237a400         CFURLCacheCurrentDiskUsage
3   CFNetwork                       0x3023783c8         CFURLRequestSetHTTPRequestBodyParts
4   ZVVOneApp                       0x203193e1c         HTTPMetrics.init (HTTPMetrics.swift:61)
5   ZVVOneApp                       0x203165858         HTTPResponse.init (HTTPResponse.swift:126)
6   ZVVOneApp                       0x203174368         [inlined] HTTPResponse.__allocating_init (HTTPResponse.swift:120)
7   ZVVOneApp                       0x203174368         HTTPDataLoader.completeTask (HTTPDataLoader.swift:434)
8   ZVVOneApp                       0x2031745cc         [inlined] HTTPDataLoader.urlSession (HTTPDataLoader.swift:268)
9   ZVVOneApp                       0x2031745cc         HTTPDataLoader.urlSession (<compiler-generated>:266)
10  NewRelic                        0x107941e28         -[NRURLSessionTaskDelegateBase URLSession:task:didCompleteWithError:] (NRMAURLSessionTaskDelegateBase.m:55)
11  CFNetwork                       0x30234fa74         CFURLRequestCopyHTTPRequestMethod
12  Foundation                      0x3041f43e8         __NSBLOCKOPERATION_IS_CALLING_OUT_TO_A_BLOCK__
13  Foundation                      0x304205a9c         -[NSBlockOperation main]
14  Foundation                      0x3041dfad8         __NSOPERATION_IS_INVOKING_MAIN__
15  Foundation                      0x3041f0018         -[NSOperation start]
16  Foundation                      0x3041f34e4         __NSOPERATIONQUEUE_IS_STARTING_AN_OPERATION__
17  Foundation                      0x304200cf8         __NSOQSchedule_f
18  libdispatch.dylib               0x300da2044         _dispatch_block_async_invoke2
19  libdispatch.dylib               0x300dc0090         _dispatch_client_callout
20  libdispatch.dylib               0x300d97ab4         _dispatch_continuation_pop$VARIANT$armv81
21  libdispatch.dylib               0x300d971f0         _dispatch_async_redirect_invoke
22  libdispatch.dylib               0x300da462c         _dispatch_root_queue_drain
23  libdispatch.dylib               0x300da4de8         _dispatch_worker_thread2
24  libsystem_pthread.dylib         0x3b7a29dd0         _pthread_wqthread

I have a custom validator like this:

public class ErrorMessageValidator: HTTPValidator {

    public func validate(response: HTTPResponse, forRequest request: HTTPRequest) -> HTTPResponseValidatorResult {
        if [HTTPStatusCode.ResponseType.success, .redirection].contains(response.statusCode.responseType) {
            return .nextValidator
        }

        guard let data = response.data else {
            return .failChain(HTTPError(.invalidResponse))
        }

        let jsonData = JSON(data)

        guard let message = jsonData.dictionary?["message"]?.string else {
            return .nextValidator
        }

        return .failChain(MyError(title: message))
    }
}

And I get to the last line as expected, so a MyError is returned, wrapped into the failChain.

But in my calling code:

let request = myServerCall(myParameters).httpRequest()
let response = try await request.fetch(MyModel.self)

an error is thrown (expected) but it is an HTTPError with an empty error property, so my MyError isn't available here

Platform Version

iOS 16

SDK Version

1.7.2

Xcode Version

Xcode 14

Steps To Reproduce

I have this request

try await HTTPRequest {
            $0.path = "businesscards"
            $0.headers = HTTPHeaders([
                .authorization:"Bearer \(token)",
                .contentType:"multipart/form-data"
            ])
            
            $0.method = .post
            $0.body = try .multipart(boundary: nil, { form in
                if let name = businessCardCreation.name {
                    try form.add(string: name, name: "name")
                }
                if let surname = businessCardCreation.surname {
                    try form.add(string: surname, name: "surname")
                }
                if let email = businessCardCreation.email {
                    try form.add(string: email, name: "email")
                }
            })
        }.fetch().decode(BusinessCard.self, decoder: Decoders.decoder)

that is not sending Authorization header to the backend.

Expected Behavior

I expect to see the Authorization header in the backend.

Actual Incorrect Behavior

I can't see the authorization header in the backend. Here is a screenshot of the headers sent in a local call, but the same happens on the cloud.

Trying the same request using Alamofire everything works.

Platform Version

Any

SDK Version

1.7.0

Xcode Version

13.4.1 (13F100)

Steps To Reproduce

Embed the latest 1.7.0 in any bigger project and run project.

Xcode has over 100% usage as soon as the play button is being hit. Also restart of Xcode with project or restart of Mac and opening the project leads to 100% CPU usage. Xcode is extremely slow and doesn't show errors in the code anymore.

Expected Behavior

Version not affecting Xcode CPU usage.

Actual Incorrect Behavior

100% CPU usage when SDK is embedded in project.

Bug Report

| Q | A | | ------------------- | ------ | | BC Break | no | | Version | 1.5.2 |

Summary

Hello, I am experiencing some crash using RealHTTP on my app. Not sure if it comes from the lib or the use I do.

Current behavior

The app crash when deallocating a pointer on line 438 of the file HTTPDataLoader.swift.

How to reproduce

2% of my users are experiencing the crash but not sure of how to reproduce for the moment.

Here are the stack trace : crash.txt

Thanks for your help

When transferMode = .largeData the data is downloaded to a temporary file in Downloads folder. However, when the Downloads folder does not yet exists the following error is shown:

2022-06-01 09:01:04.079666+0200 Project[3490:61771] open on .../Library/Developer/CoreSimulator/Devices/0FAC81F6-4CC8-492C-AB90-BD7A148107C9/data/Containers/Data/Application/9BE2AA69-AFBB-4796-90A0-9BAC0CE7CCE3/Downloads/ACEC951C-84BC-44E4-8B27-0B6BC2623240: No such file or directory

The pull requests resolves the issue and creates the temporary folder if it does'n yet exists.

I have a custom validator that parses the JSON response and tries to build an Error struct. If it builds the error successfully, my validator will return with a .failChain(responseError), but in my catch at the call site, the error is still embedded into a HTTPError, which makes the catch a bit convoluted:

import RealHTTP
import MyService

do { 
  let result = try await service.callSomeEndpoint()
} catch let httpError as HTTPError {
  if let responseError = httpError.error as? ResponseError {
    // handle my custom response error here
  } else {
    // handle the http error?
  }
} catch {
  // 
}

Question: what's your rationale behind automatically nesting the validator custom errors in a HTTPError ?

I would probably prefer to have 2 dedicated catches:

} catch let responseError as ResponseError {
  ...
} catch let httpError as HTTPError {
  ...
}

So I can ignore the HTTPError and fallback to a generic catch if I'd like to.. and I could also avoid leaking the RealHTTP implementation/imports in all my services ;)

I tried your example , but I get error

Instance method 'fetch' requires that 'HTTPRequest' conform to 'APIResourceConvertible'"

That error appeared hear

public extension HTTPClient {
func fetch<T: APIResourceConvertible>(_ convertible: T) async throws -> T.Result {
let result = try await fetch(convertible.request())
return try result.decode(T.Result.self)
}
}

I use several stubs, just implemented the third one and now running into problems:

extension HTTPStubRequest {
    static var stubXY: HTTPStubRequest {
        HTTPStubRequest().match(urlRegex: "*/resources/user/*/info").stub(for: .get) { urlRequest, matchedStubRequest in

        // getting here for the url https://example.com/resources/public/countries
        ...
        }
    }
}

This should never be the case, right? I'm having another stub added before this that should handle this case.

Bug Report

| Q | A | | ------------------- | ------ | | BC Break | no | | Version | 1.6.2 |

Summary

We have many (but intermittent) waitForStatusRecordUnlock crash and we doesn't know why,

Do you have any idea ? Misuse of the HTTPClient ?

Thanks for ur helping,

Bug Report

| Q | A | | ------------------- | ------ | | BC Break | no | | Version | 1.52 |

Summary

isAbsoluteURL regex doesn't work like expected

Current behavior

Impossible to add urlComponents.queryItems coz path is absolute :/

How to reproduce

/connect/login is an absolute URL connect/login is an absolute URL

Expected behavior

/connect/login is not an absolute URL connect/login is not an absolute URL

I think error is the ? after the first group in the regex

Thanks so much for this library! And thank you for great documentation! I sat down this evening and enjoyed what I was reading so much that I figured I might as well take a moment to help improve the guide, especially since a pair of fresh eyes are the best way to find little typos and such. Feel free to revert any of my suggestions, as I know some of them might be considered just an English style choice.

Hello, I did a take-home exercise for Walmart Labs and decided to use RealHTTP after doing a quick compare to other HTTP clients. My take-home was rejected and one of the three reasons was it "doesn't handle errors in a nice way". I didn't really handle errors as it was outside of the scope and 60 minute timeframe, but I did put in a catch:

struct APIService {
    
    struct InfoService {
        static func fetchCountries() async ->CountryInfo?  {
            do {
                guard let url = URL(string: "\(APIEnvironment.development.baseURL)/peymano-wmt/32dcb892b06648910ddd40406e37fdab/raw/db25946fd77c5873b0303b858e861ce724e0dcd0/countries.json") else { throw "Malformed URL" }
                if let countryInfo = try await HTTPRequest(url)
                    .fetch(CountryInfo.self) {
                    return countryInfo
                }
            } catch {
                //TODO: log error
            }
            return nil
        }
    }
}

class CapitalSearchViewModel {
    weak var dataSource : GenericDataSource<CountryInfoElement>?

    init(dataSource : GenericDataSource<CountryInfoElement>?) {
        self.dataSource = dataSource
    }
    
    func fetchCountries() {
        Task {
            await self.dataSource?.data.value = APIService.InfoService.fetchCountries() ?? []
        }
    }
}

Extending on what is above, what would be the preferred way to do error handling with RealHTTP? I can provide access to the project if anyone wants to take a look. I didn't really agree with the feedback and would appreciate more opinions from other iOS devs if anyone has time. My email is bhartsb at gmail.com. The other feedback was "iOS code is overly complex" and "boilerplate code". Boiler plate code was Xcode generated for the most part.

Hi, I am using HTTPStubber and love its simplicity and power.

I was wondering what would be the best approach to use it in a project with iOS 11+ support ? Since I can't use async/await, is there any way I can still use it? Or should I start looking at a custom implementation (or revert back to the old OHHTTPStubs 😒)?

Any thoughts?