-
Notifications
You must be signed in to change notification settings - Fork 11
signup does not work #205
Comments
Hi, |
could be the same issue as mention in #71 ? if so, who could help? |
This happens after I did confirm the popup dialog. Hence after I let your application have access to my github account. I did try this with firefox and chrome, as well as a chrome instance with all popup and script blockers deactivated. I do think there is something in my profile which you don't handle properly in the application. Just a wild guess which might be completely unrelated, but maybe you don't handle Github PRO accounts properly? Not that I think that they are much different from other accounts... |
Just noticed that I do seem to get a different stacktrace in chrome:
|
pretty sure that the error message / stack traces are just an artifact of some bodged error handling here. And therefore provide no useful information. They are basically hiding the real hiccup. Which is reviled when looking at the console: It seems like Which then returns with:
and Response body: {
"message": "tokens is undefined",
"stack": "Mirage: Your POST handler for the url /api/v1/token threw an error:\n\n_default/<@http://localhost:4200/assets/opensource-challenge-client.js:4095:9\nhandle@webpack://__ember_auto_import__/./node_modules/miragejs/dist/mirage-esm.js?:2813:19\n_getMirageResponseForRequest@webpack://__ember_auto_import__/./node_modules/miragejs/dist/mirage-esm.js?:3326:31\nhandle@webpack://__ember_auto_import__/./node_modules/miragejs/dist/mirage-esm.js?:3306:19\n_registerRouteHandler/<@webpack://__ember_auto_import__/./node_modules/miragejs/dist/mirage-esm.js?:7845:31\nhandleRequest@webpack://__ember_auto_import__/./node_modules/pretender/dist/pretender.es.js?:399:30\nsend@webpack://__ember_auto_import__/./node_modules/pretender/dist/pretender.es.js?:186:21\nfetch/<@http://localhost:4200/assets/vendor.js:86584:15\ninitializePromise@http://localhost:4200/assets/vendor.js:78674:15\nPromise@http://localhost:4200/assets/vendor.js:79196:35\nfetch@http://localhost:4200/assets/vendor.js:86521:16\nself.default@http://localhost:4200/assets/vendor.js:86623:27\nmakeRequest/<@http://localhost:4200/assets/vendor.js:141295:28\ninitializePromise@http://localhost:4200/assets/vendor.js:78674:15\nPromise@http://localhost:4200/assets/vendor.js:79196:35\nmakeRequest@http://localhost:4200/assets/vendor.js:141294:14\nauthenticate/<@http://localhost:4200/assets/opensource-challenge-client.js:196:14\ninitializePromise@http://localhost:4200/assets/vendor.js:78674:15\nPromise@http://localhost:4200/assets/vendor.js:79196:35\nauthenticate@http://localhost:4200/assets/opensource-challenge-client.js:190:14\nsuperWrapper@http://localhost:4200/assets/vendor.js:48090:22\nauthenticate@http://localhost:4200/assets/vendor.js:141782:28\nauthenticate@http://localhost:4200/assets/vendor.js:142689:22\n_toriiLogin@http://localhost:4200/assets/opensource-challenge-client.js:5984:35\n"
}
from mirage-esm.js:6928:19 My best guess is that this happens probably because the |
Tried using older versions of ember-simple-auth, sadly those didn't solve the problem.
Trying to find out what is happening. |
Everything related to mirage does not relate to any problem in production, as mirage is only a mock api to be used locally if you do not run the real backend. I don't think it is possible to login to the development build without running the real API aswell. |
Thanks for the hint @topaxi , do you have any ideas to where the problem might be or how to debug it more efficiently? |
As far as i can see it seems like this
Because the given model (in this case the variable err) is not correctly being serialized (https://api.emberjs.com/ember/3.12/classes/Route/methods/transitionTo?anchor=transitionTo). Thus ember dying. When i was testing locally i even had the value false in the err variable which does definitely not seem healthy. But this would also mean that there is an error happening before this one and therefore not solve the original problem... |
I'm also still getting this error message. But today I noticed that I'm actually logged in afterwards, so this issue is not that critical anymore - at least for me :) |
@felixfontein this is actually right. |
That's definitely not optimal, but works for me ;) Thanks! |
Sorry, this does not work for me. I will not grant an application access to my personal data which breaks when I do so. |
@Enteee it should work now. Feel free to test it and give us feedback if there are any persisting issues. As to the previous problem that required two logins: There were never any concerns to be had about your personal data. The flow of the code had an issue where it created the user (when logging in for the first time) but didn't find aforementioned user in the next step. When the user then tried a second login attempt it was able to find the user without an issue. This is also the reason why users that participated in previous years never had this issue. |
and which pull request now fixed the issue? |
@Enteee it was not one PR but in fact several commits over the last week. The open-source challenge has a client and an api and the problem was finally found in the elixir api. The last commit on this issue is this one: opensource-challenge/opensource-challenge-api@0c40439 |
When I try to register I do get:
and:
The text was updated successfully, but these errors were encountered: