How to set http or https based on window.location.protocol - api

I am working on a flutter web application.I want to prevent the CORS error whether my API Endpoint runs on https:// or http:// How can I modify my EndPoint Class below to set http or https based on the current window.location.protocol?
Here is my service class:
import 'package:universal_html/html.dart';
class ApiEndPoint {
Location currentLocation = window.location;
static
const host = 'api.xyz.com';
static
const http = 'http://';
static
const https = 'https://';
static String baseUrl = "";
// get base {
// if (currentLocation.protocol == 'https') {
// return baseUrl = "$https$host/api";
// } else if (currentLocation.protocol == 'http') {
// return baseUrl = "$http$host/api";
// }
// }
}

I'm not sure I understand your problem:
If currentLocation.host != apihost (for instance currentLocation could be https://app.xyz.com) you'll have to allow CORS on your API host anyways. Why then still support unsecure http:// protocol for your API. It's no problem to access https:// resources from an unsecured website. Just always use https://api.xyz.com and allow CORS origins http://app.xyz.com and https://app.xyz.com
If currentLocation.host == apihost just ditch the protocol://host part from your apirequests and just access /api/.... The browser (resp. webview) will add the correct protocol and host anyways ...
Other than that, what's wrong with your current approach? It should work. But you could probably simplify it to
get base {
return baseUrl = "${currentLocation.protocol}$host/api"
}
and ditch the const static http and const static https string constants.
I don't have any experience with flutter app architecture, nor do I know the architecture of your app. So I don't know whether it's possible to have multiple instances of that ApiEndPoint with different window.locations. This will probably lead to a problem with your static String baseUrl, as a static variable is shared among all instances of a class. So when you have different window.locations at the same time, this will be inconsistent. So maybe you should ditch the baseUrl variable alltogether and just do
get base {
return "${currentLocation.protocol}$host/api"
}
as this will always have the correct protocol for the location of the current instance.
EDIT Regarding your comment:
Of course you cannot access the base property in a static way. Ie ApiEndPoint.base won't work, because base is not a static property. So in principle you have 2 options (depending on your architecture)
Change base (and of course also currentLocation) to be static. Then you can access it vie ApiEndPoint.base
static Location currentLocation = window.location;
static get base {
return "${currentLocation.protocol}$host/api"
}
Create an instance of your ApiEndPoint class, and access the base property only via that instance
final api = ApiEndPoint();
...
final baseUrl = api.base;
In any case, you should read about static keyword ...

Related

How to authenticate Shopware 6 <base-app-url> correctly

With the Admin SDK it's possible to further enrich the administration in Shopware 6. As in the installation guide for apps stated, an entry point (base-app-url) needs to be provided in the manifest file of an app.
Since every request needs to be authenticated properly, this GET request also needs authentication. However, I am not able to authenticate this one in the same way as I am successfully doing it with the GET request from modules.
The base-app-url request looks the following (in my case with some [custom] entity privileges):
http://localhost:3000/sdk?location-id=sw-main-hidden&privileges=%7B%22read%22%3A%5B%22language%22%2C%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22create%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22update%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22delete%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%7D&shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045964&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE&shopware-shop-signature=e7b20a46487046a515638f76c6fadab6b1c749ea4a8ac6e7653527e73ba18380
The shop has the following data
Shop {
_id: 'sbzqJiPRrbHAlC2K',
_url: 'http://localhost:8888',
_secret: '3c5a2f031006791f2aca40ffa22e8febbc8a53d8',
_apiKey: 'SWIAB2PVODCWSLZNDMC5ZM1XWA',
_secretKey: 'VnNwM0ZOMnN1Y05YdUlKazlPdlduWTdzOHhIdFpacjVCYkgzNEg'
}
I am currently authenticating my modules like the following (Node.js):
const SHOPWARE_SHOP_SIGNATURE = 'shopware-shop-signature';
export function authenticateGetRequest(req: Request, shop: Shop): void {
// e7b20a46487046a515638f76c6fadab6b1c749ea4a8ac6e7653527e73ba18380
const signature = getSignatureFromQuery(req);
verifySignature(shop.secret, removeParamsFromQuery(req), signature);
}
function getSignatureFromQuery(req: Request): string {
if (!req.query[SHOPWARE_SHOP_SIGNATURE]) {
throw new Error('Signature is not present in request!');
}
return req.query[SHOPWARE_SHOP_SIGNATURE] as string;
}
function removeParamsFromQuery(req: Request): string {
// Some code
// Returns following string - Does neither work for base-app-url nor for module GET requests:
// 'shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045964'
// If the string follows this pattern, it works only for modules:
// shop-id={id}&shop-url={url}&timestamp={ts}&sw-version={v}&sw-context-language={cl}&sw-user-language={ul}
}
function verifySignature(secret: string, message: string, signature: string): void {
const hmac = crypto.createHmac('sha256', secret).update(message).digest('hex');
if (hmac !== signature) {
throw new Error('Signature could not be verified!');
}
}
However the base-app-url cannot be verified correctly and the "Signature could not be verified!" error is thrown.
What am I doing wrong here?
More info:
Additionally I added a GET request for a module where everything is working:
http://localhost:3000/faq?shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045963&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE&shopware-shop-signature=0f0889c9e8086c6c3553dc946a01f2ef27b34cd1c55b0c03901b6d8a6a9b6f53
The resulting string can be verified:
shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045963&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE
Try out following code in some php sandbox environment:
<?php
$message = 'shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045963&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE';
$secret = '3c5a2f031006791f2aca40ffa22e8febbc8a53d8';
$signature = '0f0889c9e8086c6c3553dc946a01f2ef27b34cd1c55b0c03901b6d8a6a9b6f53';
$hmac = hash_hmac('sha256', $message, $secret);
if (!hash_equals($hmac, $signature)) {
echo 'Signature not valid';
} else {
echo 'Signature valid';
}
SOLUTION:
Express decodes the query strings automatically with req.query depending on your express configuration. Keep in mind to validate the hmac with encoded query params as they are passed from shopware.
In my case the only difference where the decoded privileges and they looked like this:
&privileges={"read":["language","ce_atl_faq_group_faqs","ce_atl_faq_group","ce_atl_faq"],"create":["ce_atl_faq_group_faqs","ce_atl_faq_group","ce_atl_faq"],"update":["ce_atl_faq_group_faqs","ce_atl_faq_group","ce_atl_faq"],"delete":["ce_atl_faq_group_faqs","ce_atl_faq_group","ce_atl_faq"]}
But they need to look like this:
&privileges=%7B%22read%22%3A%5B%22language%22%2C%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22create%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22update%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22delete%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%7D
Looking at the QuerySigner, this is how the signature is generated on the side of Shopware with the actual arguments:
hash_hmac(
'sha256',
'location-id=sw-main-hidden&privileges=%7B%22read%22%3A%5B%22language%22%2C%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22create%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22update%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22delete%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%7D&shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045964&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE',
'VnNwM0ZOMnN1Y05YdUlKazlPdlduWTdzOHhIdFpacjVCYkgzNEg'
);
// 8034a13561b75623420b06fb7be01f20d97556441268939e9a5222ffec12215a
Given on your side you remove the shopware-shop-signature query param AND that the secrets are equal on both sides, you should be able to regenerate the matching signature.
const crypto = require('crypto');
const message = 'location-id=sw-main-hidden&privileges=%7B%22read%22%3A%5B%22language%22%2C%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22create%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22update%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%2C%22delete%22%3A%5B%22ce_atl_faq_group_faqs%22%2C%22ce_atl_faq_group%22%2C%22ce_atl_faq%22%5D%7D&shop-id=sbzqJiPRrbHAlC2K&shop-url=http://localhost:8888&timestamp=1674045964&sw-version=6.4.18.0&sw-context-language=2fbb5fe2e29a4d70aa5854ce7ce3e20b&sw-user-language=de-DE';
const hmac = crypto.createHmac('sha256', 'VnNwM0ZOMnN1Y05YdUlKazlPdlduWTdzOHhIdFpacjVCYkgzNEg').update(message).digest('hex');
// 8034a13561b75623420b06fb7be01f20d97556441268939e9a5222ffec12215a
So in theory your code looks fine. Verify that the query string matches exactly. Things to check:
Maybe your node server decodes the url entities unwantedly?
Does your node serve escape special characters in the query string?
Do the secrets match on both sides?
To consider additionally:
Consider to just point the base-app-url to a static page outside of the scope of your app server instead. As that page will be loaded inside an iframe, you can use client side javascript to read the query parameters and, only if necessary, make requests to your app server using the credentials from inside the iframe. Keep in mind you really only need the authentication if you need to handle personalized data, otherwise you might as well serve static assets without the need for authentication.

Dependency Injection Access While Configuring Service Registrations in asp.net Core (3+)

I have cases, where I want to configure services based on objects which are registered in the dependency injection container.
For example I have the following registration for WS Federation:
authenticationBuilder.AddWsFederation((options) =>{
options.MetadataAddress = "...";
options.Wtrealm = "...";
options.[...]=...
});
My goal in the above case is to use a configuration object, which is available via the DI container to configure the WsFederation-middleware.
It looks to me that IPostConfigureOptions<> is the way to go, but until now, I have not found a way to accomplish this.
How can this be done, or is it not possible?
See https://andrewlock.net/simplifying-dependency-injection-for-iconfigureoptions-with-the-configureoptions-helper/ for the I(Post)ConfigureOptions<T> way, but I find that way too cumbersome.
I generally use this pattern:
// Get my custom config section
var fooSettingsSection = configuration.GetSection("Foo");
// Parse it to my custom section's settings class
var fooSettings = fooSettingsSection.Get<FooSettings>()
?? throw new ArgumentException("Foo not configured");
// Register it for services who ask for an IOptions<FooSettings>
services.Configure<FooSettings>(fooSettings);
// Use the settings instance
services.AddSomeOtherService(options => {
ServiceFoo = fooSettings.ServiceFoo;
})
A little more explicit, but you have all your configuration and DI code in one place.
Of course this bypasses the I(Post)ConfigureOptions<T> entirely, so if there's other code that uses those interfaces to modify the FooSettings afterwards, my code won't notice it as it's reading directly from the configuration file. Given I control FooSettings and its users, that's no problem for me.
This should be the approach if you do want to use that interface:
First, register your custom config section that you want to pull the settings from:
var fooSettingsSection = configuration.GetSection("Foo");
services.Configure<FooSettings>(fooSettingsSection);
Then, create an options configurer:
public class ConfigureWSFedFromFooSettingsOptions
: IPostConfigureOptions<Microsoft.AspNetCore.Authentication.WsFederation.WsFederationOptions>
{
private readonly FooSettings _fooSettings;
public ConfigureWSFedFromFooSettingsOptions(IOptions<FooSettings> fooSettings)
{
_fooSettings = fooSettings.Value;
}
public void Configure(WsFederationOptions options)
{
options.MetadataAddress = _fooSettings.WsFedMetadataAddress;
options.Wtrealm = _fooSettings.WsFedWtRealm;
}
}
And finally link the stuff together:
services.AddTransient<IPostConfigureOptions<WsFederationOptions>, ConfigureWSFedFromFooSettingsOptions>();
The configurer will get your IOptions<FooSettings> injected, instantiated from the appsettings, and then be used to further configure the WsFederationOptions.

Ktor custom feature swap url

Im trying to add a custom feature in Ktor. It's basically a url swapper (we have a scenario where domains might be changed during anytime & can't update the client everytime).
We get the swapper list available and need a CustomFeature in Ktor to swap the url based on list. However, the context.request or request.url - everything is val and Im not able to assign new url to the request.
In Retrofit, it used to work like this
if (currentUrl.contains(urlSwapper.oldUrl)) {
val newUrl = currentUrl.replace(urlSwapper.oldUrl, urlSwapper.newUrl)
val newHttpUrl = request.url.newBuilder(newUrl)!!.build()
// build a new request with the new url. replace it
request = request.newBuilder().url(newHttpUrl).build()
break
}
}
In Ktor feature, Im trying something like this
scope.requestPipeline.intercept(HttpRequestPipeline.Transform) {
val currentUrl =
context.url.protocol.name + "://" + context.url.host + context.url.encodedPath
for (urlSwapper in feature.urlSwappers) {
if (currentUrl.contains(urlSwapper.oldUrl)) {
val newUrl = currentUrl.replace(urlSwapper.oldUrl, urlSwapper.newUrl)
val newHttpUrl = Url(newUrl)
context.url(url = newHttpUrl)
break
}
}
proceedWith(subject)
}
}
Is this the right way to do this ?
Generally yes, this is the right way. I have a few recommendations:
Intercept the sendPipeline instead of the requestPipeline.
An example:
client.sendPipeline.intercept(HttpSendPipeline.State) {
context.url(url = newUrl)
}
Get rid of proceedWith(subject) call because it's redundant.
Try to use Url objects instead of strings. You can get the current URL without affecting context by cloning UrlBuilder and building Url from it: context.url.clone().build()

How to dynamically resolve controller with endpoint routing?

Upgrading to asp.net core 2.2 in my hobby project there is a new routing system I want to migrate to. Previously I implemented a custom IRouter to be able to set the controller for the request dynamically. The incoming request path can be anything. I match the request against a database table containing slugs and it looks up the a matching data container class type for the resolved slug. After that I resolve a controller type that can handle the request and set the RouteData values to the current HttpContext and passing it along to the default implementation for IRouter and everything works ok.
Custom implementaion of IRouter:
public async Task RouteAsync(RouteContext context)
{
var requestPath = context.HttpContext.Request.Path.Value;
var page = _pIndex.GetPage(requestPath);
if (page != null)
{
var controllerType = _controllerResolver.GetController(page.PageType);
if (controllerType != null)
{
var oldRouteData = context.RouteData;
var newRouteData = new RouteData(oldRouteData);
newRouteData.Values["pageType"] = page.PageType;
newRouteData.Values["controller"] = controllerType.Name.Replace("Controller", "");
newRouteData.Values["action"] = "Index";
context.RouteData = newRouteData;
await _defaultRouter.RouteAsync(context);
}
}
}
A controller to handle a specific page type.
public class SomePageController : PageController<PageData>
{
public ActionResult Index(PageData currentPage)
{
return View("Index", currentPage);
}
}
However I got stuck when I'm trying to figure out how I can solve it using the new system. I'm not sure where I'm suppose to extend it for this behavior. I don't want to turn off the endpoint routing feature because I see an opportunity to learn something. I would aso appreciate a code sample if possible.
In ASP.NET 3.0 there is an new dynamic controller routing system. You can implement DynamicRouteValueTransformer.
Documentation is on the way, look at the github issue

Move Fastly config to CloudFlare Workers

How can the following code be re-written to work with the CF Workers feature?
# Start
if(req.url ~ "^/app" ) {
set req.url = regsub(req.url, "^/app/", "/");
set req.http.X-DR-SUBDIR = "app";
}
#end condition
Cloudflare Workers implements the Service Worker standard, so you will need to reimplement the VCL code snippet you posted in terms of a Service Worker.
Before I show you how to do that, consider what happens when a request for https://example.com/apple arrives at the proxy. I would expect the first regex for ^/app to match, but the second one for ^/app/ not to match -- i.e., the request would be passed through with no change to the URL, but with the addition of an X-DR-SUBDIR: app header.
I suspect that behavior is a bug, so I'll first implement a worker as if the first regex were ^/app/.
addEventListener("fetch", event => {
let request = event.request
// Unlike VCL's req.url, request.url is an absolute URL string,
// so we need to parse it to find the start of the path. We'll
// need it as a separate object in order to mutate it, as well.
let url = new URL(request.url)
if (url.pathname.startsWith("/app/")) {
// Rewrite the URL and set the X-DR-SUBDIR header.
url.pathname = url.pathname.slice("/app".length)
// Copying the request with `new Request()` serves two purposes:
// 1. It is the only way to actually change the request's URL.
// 2. It makes `request.headers` mutable. (The headers property
// on the original `event.request` is always immutable, meaning
// our call to `request.headers.set()` below would throw.)
request = new Request(url, request)
request.headers.set("X-DR-SUBDIR", "app")
}
event.respondWith(fetch(request))
})
To revisit the https://example.com/apple case, if we really wanted a Cloudflare Worker which pedantically reproduces the VCL snippet's behavior, we could change these lines (comments elided):
if (url.pathname.startsWith("/app/")) {
url.pathname = url.pathname.slice("/app".length)
// ...
}
to these:
if (url.pathname.startsWith("/app")) {
if (url.pathname.startsWith("/app/")) {
url.pathname = url.pathname.slice("/app".length)
}
// ...
}