Many redux examples show making HTTP requests directly in an async action; however this will result in an exterme amount of duplicate code common to all requests, for example:
Retrying failed requests based on status code
Appending common HTTP Headers to each request.
Invoking additional http requests based on response (ie: oauth token refresh)
Aborting in-flight requests on route transitions.
My gut feeling is that a middleware could be used to make http requests (in the same vein as redux-api-middleware) which can keep the in-flight requests in the store - however I am also wondering if I'm leaning on bad habbits - is a little duplication a small price to pay for immutability?
Your action creators are just JavaScript functions.
If you have duplication between several functions, you extract the common code into another function. This is no different. Instead of duplicating the code, extract the common code into a function and call it from your action creator.
Finally, this pattern can be abstracted away with a custom middleware. Check out the “real world” example in Redux repo to see how it can be done.
All but the aborting could be accomplished while staying immutable by using a request factory which attaches .then and .catch (or equivalent, depending on promise flavor) to the request before returning it.
you can have a action which executes its operation in addition it calls another action, to achieve this you need to have a redux-async-transitions, the example code is given below
function action1() {
return {
type: types.SOMEFUNCTION,
payload: {
data: somedata
},
meta: {
transition: () => ({
func: () => {
return action2;
},
path : '/somePath',
query: {
someKey: 'someQuery'
},
state: {
stateObject: stateData
}
})
}
}
}
and here is for asynchronous call
function asynccall() {
return {
types: [types.PENDINGFUNCTION, types.SUCCESSFUNCTION, types.FAILUREFUNCTION],
payload: {
response: someapi.asyncall() // only return promise
}
meta: {
transition: (state, action) => ({
onPending: () => {
},
onSuccess: (successdata) => {
//gets response data can trigger result based on data
},
onFail: (promiseError) => {
//gets error information used to display messages
}
})
}
}
}
Calling http request in middleware is the same bad ideas as calling it in action creators. You should focus on making our reducers powerful enough to handle asynchronous effects as well as synchronous state transitions. The best way i found while working with redux is describe effects in the reducer(http request is the effect) and then handle it with library like redux-loop or redux-saga
For avoiding code duplication you can extract common code to request function and use it for handling http effects
Related
I have GET requests and normally when those succeeded I save data in store, but for POST requests I need to know if it succeeded or not, in order to execute some code (show a message and redirect), the docu says you can use an isLoading variable, but it just says if the service is working but not if it succeeded, if I try to create a new success variable in the store, it will be turned on forever after the request and I don't need that either. I tried returning a promise from the action creator and handle response directly inside the component but it looks like the same to call axios there instead of using redux.
My action creator looks like this:
export function createProject(userId, projectName) {
return function (dispatch) {
dispatch({ type: projectsActions.START_CREATE_PROJECT });
return ProjectsService.createProject(userId, projectName).then(() => {
dispatch({ type: projectsActions.SUCCESS_CREATE_PROJECT });
}).catch((error) => {
dispatch({ type: projectsActions.ERROR_CREATE_PROJECT });
throw error;
});
}
}
I understand where your doubts are coming from, it doesn't seem appropriate to have a field on your Redux store only to know the success of a one-time request.
If you only need to make a post request and only care about it's result once, the simplest way to do it is to use state in the component making the request. Component-level state is easily manageable and gets removed from memory when the component is unmounted, but on the other hand you may want to have a single source of truth for your app. You have to make a choice, but your Redux implementation is correct.
tl;dr: Within a Redux middleware function, is it okay to dispatch a new action after calling next to finish updating the store?
I'm building a HackerNews reader using Flutter and built-flutter-redux, based off of Brian Egan's TodoMVC example. It uses HN's Firebase-backed API to pull data:
https://github.com/HackerNews/API
My actions look like this right now:
ActionDispatcher<Null> fetchHackerNewsTopStories;
ActionDispatcher<List<int>> fetchHackerNewsTopStoriesSuccess;
ActionDispatcher<Null> fetchHackerNewsTopStoriesFailure;
ActionDispatcher<Null> fetchNextHackerNewsItem;
ActionDispatcher<HackerNewsItem> fetchHackerNewsItemSuccess;
ActionDispatcher<Null> fetchHackerNewsItemFailure;
There's a piece of middleware that listens for the fetchHackerNewsTopStories action and kicks off a call to the API:
MiddlewareHandler<AppState, AppStateBuilder, AppActions, Null>
createFetchHackerNewsTopStories(HackerNewsRepository service) {
return (MiddlewareApi<AppState, AppStateBuilder, AppActions> api,
ActionHandler next, Action<Null> action) {
service.fetchHackerNewsTopStories().then((ids) {
return api.actions.fetchHackerNewsTopStoriesSuccess(ids);
}).catchError(api.actions.fetchHackerNewsTopStoriesFailure);
next(action);
};
}
When it returns, I update my app's state with the list of IDs.
At some point I need to dispatch another action, fetchNextHackerNewsItem. There's another middleware function that will listen for that action and request the details for the the first story. When those details arrive, it'll request the next story, and so on until everything's updated.
What I'd like to know is whether I can do this:
// Invoked when REST call for the list of top story IDs completes.
MiddlewareHandler<AppState, AppStateBuilder, AppActions, List<int>>
createFetchHackerNewsTopStoriesSuccess() {
return (MiddlewareApi<AppState, AppStateBuilder, AppActions> api,
ActionHandler next, Action<List<int>> action) {
next(action);
api.actions.fetchNextHackerNewsItem(); // Is this cool?
};
}
// Initiates a request for a single story's details.
MiddlewareHandler<AppState, AppStateBuilder, AppActions, Null>
createFetchNextHackerNewsItem(HackerNewsRepository service) {
return (MiddlewareApi<AppState, AppStateBuilder, AppActions> api,
ActionHandler next, Action<Null> action) {
int nextId = api.state.topStoryIds[api.state.loadedUpToIndex];
service.fetchHackerNewsItem(nextId).then((item) {
return api.actions.fetchHackerNewsItemSuccess(item);
}).catchError(api.actions.fetchHackerNewsTopStoriesFailure);
next(action);
};
}
Because createFetchNextHackerNewsItem relies on the app's state (api.state.topStoryIds[api.state.loadedUpToIndex]), I'd like for it to run after the store is updated by the next(action) call.
Is it cool to dispatch new actions in Redux middleware after calling next, or is that some kind of anti-pattern? If it is an anti-pattern, what's the best way to implement this flow?
Yes, it's fine - a middleware can do literally anything it wants when an action is dispatched. That includes modifying / logging / delaying/ swapping / ignoring the original action, as well as dispatching additional actions.
I'm trying to mock the back-end for an AngularJS(1.3.8)-app with ngMockE2E as replacement until the back-end code has been written.
I'm using already existing services that also query other data, however they return a promise. I am aware that ngMockE2E is supposed to be synchronous, however I wanted to see if there's a way to do it asynchronously first.
Looking around the web I found this and put the mocking-related code into its own seperate module to see if this approach works.
$httpBackend.whenAsync('projects/').respond(function (promise, headers, status) {
var deferred = $q.defer();
_getProjectIndex().then(function (result) {
deferred.resolve(result);
},
function (statusCode) {
console.log(statusCode);
deferred.reject(statusCode);
});
return deferred.promise;
});
When I try to run $httpBackend.whenAsync() the request just seems to 404. Checking the same request with $httpBackend.whenGET() I receive the promise containing the data I requested.
What am I doing wrong?
Currently, I use the built-in meteor http method (see http://docs.meteor.com/#http) for issuing http calls, on both my client and my server.
However, I'm experiencing two issues:
is it possible to cancel a request?
is it possible to have multiple query parameters which share the same key?
Are these just Meteor limitations, or are there ways to get both to work using Meteor?
I know I could you jquery on the clientside, and there must be a server-side solution which supports both as wel, but I'd prefer sticking with meteor code here.
"is it possible to cancel a request?"
HTTP.call() does not appear to return an object on which we could call something like a stop() method. Perhaps a solution would be to prevent execution of your callback based on a Session variable?
HTTP.call("GET", url, function(error, result) {
if (!Session.get("stopHTTP")) {
// Callback code here
}
});
Then when you reach a point where you want to cancel the request, do this:
Session.set("stopHTTP", true);
On the server, instead of Session perhaps you could use an environment variable?
Note that the HTTP.call() options object does accept a timeout key, so if you're just worried about the request never timing out, you can set this to whatever millisecond integer you want.
"is it possible to have multiple query parameters which share the same key?"
Yes, this appears to be possible. Here's a simple test I used:
Meteor code:
HTTP.call("GET", "http://localhost:1337", {
query: "id=foo&id=bar"
}, function(error, result) {
// ...
});
Separate Node.js server: (just the basic example on the Node.js homepage, with a console.log line to output the request URL with query string)
var http = require('http');
http.createServer(function(req, res) {
console.log(req.url); // Here I log the request URL, with the query string
res.writeHead(200, {
'Content-Type': 'text/plain'
});
res.end('Hello World\n');
}).listen(1337, '127.0.0.1');
console.log('Server running at http://127.0.0.1:1337/');
When the Meteor server is run, the Node.js server logged:
/?id=foo&id=bar
Of course, this is only for GET URL query parameters. If you need to do this for POST params, perhaps you could store the separate values as a serialized array string with EJSON.stringify?
I am building a registration form (passport-local as authentication, forms as form helper).
Because the registration only knows GET and POST I would like to do the whole handling in one function.
With other words I am searching after something like:
exports.register = function(req, res){
if (req.isPost) {
// do form handling
}
res.render('user/registration.html.swig', { form: form.toHTML() });
};
The answer was quite easy
exports.register = function(req, res) {
if (req.method == "POST") {
// do form handling
}
res.render('user/registration.html.swig', { form: form.toHTML() });
};
But I searched a long time for this approach in the express guide.
Finally the node documentation has such detailed information:
http://nodejs.org/api/http.html#http_http_request_options_callback
Now you can use a package in npm => "method-override", which provides a middle-ware layer that overrides the "req.method" property.
Basically your client can send a POST request with a modified "req.method", something like /registration/passportID?_method=PUT.
The
?_method=XXXXX
portion is for the middle-ware to identify that this is an undercover PUT request.
The flow is that the client sends a POST req with data to your server side, and the middle-ware translates the req and run the corresponding "app.put..." route.
I think this is a way of compromise. For more info: method-override