Summarize the problem:
When I merge a commit to Gitlab and run a pipeline, there is a script failure when trying to deploy. more specifically: WebpackError: ReferenceError: __WEBPACK_EXTERNAL_MODULE_firebase_app__ is not defined
-The goal of this post is to hopefully get a better understanding of what I am doing wrong here, as I am trying to follow the errors to find the solution on my own, but have not been successful thus far.
-The expected result is that, because the pipeline had previously worked up until this point, that it would continue to do so. The strange part about this pipeline failure is that I did not actually install any packages on the merge that caused the failure. I went ahead and updated packages and some previous errors went away (for instance, there was a failure involving the sharp library that has now gone away) but now firebase app has become the focus.
The only actual error I saw in the output was WebpackError: ReferenceError: __WEBPACK_EXTERNAL_MODULE_firebase_app__ is not defined followed by ERROR: Job failed: exit code 1
Describe what you've tried: as stated before, I initially thought if I just updated all dependencies, this would help get me a better idea of what was going wrong. I thought I was on the right track when I noticed more progress along the pipeline than before, and there was this issue error (node:260) [DEP_WEBPACK_EXTERNALS_FUNCTION_PARAMETERS] DeprecationWarning: The externals-function should be defined like ({context, request}, cb) => { ... } (Use `node --trace-deprecation ...` to show where the warning was created) so I followed the error message's advice and added curly brackets around context, request in my gatsby-node.js file. That error went away on the next pipeline attempt, but WebpackError: ReferenceError: __WEBPACK_EXTERNAL_MODULE_firebase_app__ is not defined remains
Show some code:
I believe the gatsby-node.js file is what is relevant here, so here it is:
exports.onCreateWebpackConfig = ({ stage, actions, getConfig }) => {
if (stage === 'build-html') {
actions.setWebpackConfig({
// Don't bundle modules that reference browser globals such as `window` and `IDBIndex` during SSR.
// See: https://github.com/gatsbyjs/gatsby/issues/17725
externals: getConfig().externals.concat(function ({ _context, request }, callback) {
// Exclude bundling firebase* and react-firebase*
// These are instead required at runtime.
if (/^#?(react-)?firebase(.*)/.test(request)) {
console.log('Excluding bundling of: ' + request);
return callback(null, 'umd ' + request);
}
callback();
}),
});
}
};
I can provide other details if needed, and would appreciate any advice that anyone has to offer. Thanks!
Downgraded Gatsby back to v2 (as well as all the relevant gatsby plugins) and the error went away
Related
I've upgraded to the most recent version of flutter and I've started going through now and upgrading the packages I use as well. I've noticed there are some breaking changes that have happened in several packages, but the breaking changes in the rxdart package have got me stumped.
the developers have kept up their changelogs and mentioned in 0.23.0 they removed the observable class as seen here and they even made a codemod package to try to make the migration easy, which can be seen here. this repaired only a couple of my issues, and left several more.
Epic<AppState> getConversationPreviewsEpic(ConversationRepository repo) {
return (Stream<dynamic> actions, EpicStore<AppState> store) {
return Stream(actions)
.whereType<SubscribeConversationPreviews>() // <-- subscribe to conversations stream
.where((action) => action.fromUser != null)
.switchMap((action) {
isConversationLoading = true;
return repo
.getConversationPreviewStream(action.fromUser, lastSnapshotAction: (lastTimeStamp, hasMore){
lastConversationTimeStamp = lastTimeStamp;
hasMoreConversations = hasMore;
})
.map((rooms) => ConversationPreviewsSnapshot(rooms)) //<-- when new conversations come in add them to the redux store
.takeUntil(actions.where((a) => a is UnsubscribeConversationPreviews)); //<-- unsubscribe from conversations stream
});
};
}
all of the issues I have are similar to this one here. I wish I could underline stuff in red like my editor does here. So on line 2 of the code block where it says
Stream<dynamic> actions
I get an error that Stream isn't a type
and then on the next line
return Stream(actions)
I get an error that Stream() doesn't take any arguments. I take away the arguments and the redline goes away but I can't test to see if the code still works because I can't figure out how to fix the first error.
One last thing, fromIterable and fromFuture don't work anymore and I don't see in the documentation how to adjust them to make them work again
RxDart 0.24.x based on extension method, call methods on Stream directly.
import 'package:rxdart/rxdart.dart';
actions.whereType< SubscribeConversationPreviews >();
Whenever I launch my app, and click on login on the first few tries, the login will attempt a POST http to the server. However $http always (everytime) returns NULL on first try. sometimes after several few tries still NULL if done fast. But subsequently, its all ok.
I dont get it, why is $http returning error response NULL initially ??
Here is my login controller doing the http post
Login Controller (LoginCtrl)
https://gist.github.com/anonymous/771194bc5815e4ccdf38b57d6158853f
var req = {
method: 'POST',
url: baseURL,
data: postObject,
//timeout: 5000
};
err is NULL here:
}).error(function(err) {
I dont know if it is CORS but I'ved got this set in config.xml
<access origin="*" />
my config.xml
https://gist.github.com/anonymous/b2df3a857338d14ec3fcd6dda776e212
Any ideas ?
Im using ionic 1.7.14
on device iOS 9.3.1
UPDATE
I'ved put the problem code here. can logout first to goto login screen. enter in anything in username/password field, click login once failed, second or third try will be success.
https://github.com/axilaris/ionic_null_http_problem
some troubleshooting so far: i noticed the http post request is called twice. not sure why.
UPDATED the code using $http.post.then but still has the same effect
$http.post(baseURL, postObject).then(function successCallback(response)
response has NULL data --> Object {data: null, status: 0, config: Object, statusText: ""}
It is hard to diagnose having the above details only.
However the problem could be that your handler (login function) is triggered before digest cycle finished updating $scope.data.username and $scope.data.password and for the first tries it sends empty values for those to the server and works fine later.
You can run Safari web inspector to see what is sent to the server to prove this.
The fix may depend on how your view/template is coded. Can you please share it? Or, ideally, create a working sample at http://play.ionic.io/
Another option to fix could be to try to wrap your code related to http request into
$timeout(function() {
// your code goes here
});
or, consider using .$applyAsync() (see the docs for details)
This might help to fix the problem
You are probably getting this inconsistent behavior as you are using the 'success' promise method instead of 'then' (note that use of the success method has now been deprecated).
The key differences between these two methods are:
then() - full power of the promise API but slightly more verbose
success() - doesn't return a promise but offeres slightly more convienient syntax
as highlighted in this answer.
Hence in your scenario, instead of using 'success':
var req = {
method: 'POST',
url: baseURL + 'session/login',
data: postObject,
//timeout: 5000
};
$http(req).success(function(resp) {...
use 'then' along with angular's post shortcut method (you don't have to use this shortcut method, but I think it makes the code more succinct) e.g.:
$http.post(baseURL + 'session/login', postObject).then(function successCallback(response) {
// this callback will be called asynchronously
// when the response is available
}, function errorCallback(response) {
// called asynchronously if an error occurs
// or server returns response with an error status.
});
Using 'then' returns a promise resolved with a value returned from a callback, so it should give you a consistently valid result.
it was a timeout in app.js that caused it. was set to 1 second which gives it it arbitrary success rate.
config.timeout = 1000;
I have used the code below:
Iterable<Module> modules = ImmutableSet.<Module> of(
new SshjSshClientModule());
ContextBuilder builder = ContextBuilder.newBuilder(provider).endpoint(endpoint)
.credentials(identity, credential)
.modules(modules);
System.out.printf(">> initializing %s%n", builder.getApiMetadata());
ComputeService compute = builder.buildView(ComputeServiceContext.class).getComputeService();
System.out.println(compute1.listImages());
but I am getting the following error message.........
com.google.gson.JsonSyntaxException: java.lang.IllegalStateException: Expected a string but was BEGIN_ARRAY at line 1 column 787
at org.jclouds.json.internal.DeserializationConstructorAndReflectiveTypeAdapterFactory$DeserializeIntoParameterizedConstructor.read(DeserializationConstructorAndReflectiveTypeAdapterFactory.java:181)
at org.jclouds.json.internal.NullFilteringTypeAdapterFactories$IterableTypeAdapter.readAndBuild(NullFilteringTypeAdapterFactories.java:92)
The code was working... before...
You've definitely hit a bug somewhere between the version of jclouds you're using and whatever version of whatever cloud you're using. We'll need more information to fix this. Please go through the instruction on how to Report a Bug to Apache jclouds.
I am using Meteor 4.2 (Windows) and I am always getting the "update failed: 403 -- Access denied. Can't replace document in restricted collection" when I am trying to update an object in my collection. Strangely I had no problem inserting new ones, only updates are failing.
I tried to "allow" everything on my collection:
Maps.allow({
insert: function () { return true; },
update: function () { return true; },
remove: function () { return true; },
fetch: function () { return true; }
});
But still, this update fails:
Maps.update({
_id: Session.get('current_map')
}, {
name: $('#newMapName').val()
});
Is there something else I can check? Or maybe my code is wrong? Last time I played with my project was with a previous version of Meteor (< 4.0).
Thanks for your help.
PS: Just for information, when I do this update, the local collection is updated, I can see the changes in the UI. Then very quickly it is reverted along with the error message, as the changes has been rejected by the server-side.
Alright, the syntax was actually incorrect. I don't understand really why as it was working well before, but anyway, here is the code that works fine:
Maps.update({
Session.get('current_map')
}, {
$set: {
name: $('#newMapName').val()
}
});
It seems like it must be related to what you're storing in the 'current_map' session variable. If it's a db object, then it probably looks like {_id:<mongo id here>} which would make the update finder work properly.
I ran into the same issues, and found the following to work
Blocks.update {_id:block_id}, {$set: params}
where params is a hash of all the bits i'd like to update and block_id is the mongo object id of the Block i'm trying to update.
Your note about the client side update (which flashes the update and then reverts) is expected behavior. If you check out their docs under the Data and Security section:
Meteor has a cute trick, though. When a client issues a write to the server, it also updates its local cache immediately, without waiting for the server's response. This means the screen will redraw right away. If the server accepted the update — what ought to happen most of the time in a properly behaving client — then the client got a jump on the change and didn't have to wait for the round trip to update its own screen. If the server rejects the change, Meteor patches up the client's cache with the server's result.
I'm having a doozie of a time trying to get a basic http test to work with vows.
I think I've followed the async example from vows http://vowsjs.org/#-writing-asynchronous-tests and substitued the appropriate calls, but I must be missing something.
The test code looks like this:
var http = require('http'),
vows = require('vows'),
assert = require('assert');
vows.describe("homepage").addBatch({
"Get the home page": {
topic: function() {
http.get({'host': "127.0.0.1", 'port': 5000, 'path': '/'}, this.callback);
},
'should respond with 200 OK': function(res) {
assert.equal(res.statusCode, 200);
}
}
}).export(module);
I get the following error when I try to run the test for this:
/Users/<home_folder>/node_modules/vows/lib/vows.js:80
rrored', { type: 'promise', error: err.stack || err.message || JSON.stringify(
^
TypeError: Converting circular structure to JSON
at Object.stringify (native)
at EventEmitter.<anonymous> (/Users/<home_folder>/node_modules/vows/lib/vows.js:80:90)
at EventEmitter.emit (events.js:64:17)
at /Users/<home_folder>/node_modules/vows/lib/vows/context.js:31:52
at ClientRequest.<anonymous> (/Users/<home_folder>/node_modules/vows/lib/vows/context.js:46:29)
at ClientRequest.g (events.js:143:14)
at ClientRequest.emit (events.js:64:17)
at HTTPParser.onIncoming (http.js:1349:9)
at HTTPParser.onHeadersComplete (http.js:108:31)
at Socket.ondata (http.js:1226:22)
I can get a simple http example to work on it's own. I can get the vows example to work on it's own but I can't combine them for whatever reason. I'd really appreciate some help here. I've been trying to get this to work for a while now (including much googling).
UPDATE:
Apparently adding an error argument to the call back solves this problem, thanks to help from Alexis Sellier (creator of vows).
But I have no idea why. When writing out the http lib example on it's own no error argument is required. I can't find any documentation in vows to indicate why it's needed so I'm at a bit of a loss.
My new question is why is the error argument required when using the http lib in vows?
After checking vow's source code, I think I know why. Vows always ensure that when you call this.callback, the resulting receiver function's first argument is always an error object. Vows interpret the callbacks by these rules:
If the first argument of your originating callback is a boolean, use that to determine whether or not to append an error object to the receiving callback (e.g. path.exists(boolean) will emit callback(error, exists) instead)
If the first argument is an object, assume it's an error object and use that to determine whether to add the originating callback to the "error" or "success" list. The reason this list exists is to support promise based tests I guess?
While I can't confirm the above is correct, my experience is that vows' async style is made to support node-styled callbacks (e.g. err as the first arg), and 3rd party npm modules that don't conform to this standard will be hard to test.
Please don't take my answer as gospel, as this is my own experience. Another gotcha is when you have async operations inside the function that you want to test - unless you provide a callback, vows won't be able to handle it properly.
Personally, I think vows still make it hard to test async code. I wish it had some waitFor() or until() flow control functions though.
My suggestion? When dealing with async code, use Step. Don't let vows control your flow.
It is actually missing in the documentations which is still a bit short. But you can get a glimpse of it here in this page :
'when peeled *asynchronously*': {
topic: function (banana) {
banana.peel(this.callback);
},
'results in a `PeeledBanana`': function (err, result) {
assert.instanceOf (result, PeeledBanana);
}
}
As it was said by Morten Siebuhr and Ruben Tan, this is how vows works and that is why it works like that.