I am having a weird issue. I created an application on Google App Engine and have a Login with Facebook button on it, for which I am doing server side authentication.
I give the redirect_url, and facebook was calling the URL correctly with no issues. The session parameters that I set were being retrieved on the redirect call and everything was working fine.
Yesterday, I got a domain on godaddy and mapped it to my appspot account using google apps. Now when I click on Facebook login, I am getting two calls on the redirect uri, the first one carries the session varaibles and the session one doesn't. I am not very familiar with domain mapping and followed the steps on Google Apps.
Can anyone help me in the right direction on this.
it's a little out of date, but i documented some GAE to Facebook gotchas here:
http://javagwt.blogspot.com/2010/08/facebook-apps-on-app-engine-without-any.html
It may also help to read about naked domain mapping with godaddy, to make sure you're not getting bounced around. Even though you are mapped to your domain through google apps, you can try to put the redirect URL for facebook as yourappid.appspot.com - the redirect URL you provide, and the one in your facebook app settings must match.
My app, nimbits.com writes to facebook from GAE all of the time - the code is on github under server/facebook
https://github.com/bsautner/com.nimbits/tree/master/nimbits-tds/src/com/nimbits/server
Thanks for the answers bsautner and Michele. I finally figured out the issue. I have google ads on my website. The google ads was trying to parse the URL content and creating a second request for every request that I create. After removing the google ads, I get a single callback with session values retained. It all works now. The final output is this website - www.imagecrashers.com. I will be glad for any suggestions from the gurus here, regarding layout or api calls simplification. Thanks again to all.
Related
I have a GAE app and a custom domain registered on enom.com. The app is a static website that's configured by app.yaml.
I'm trying to use LetsEncrypt certs for ssl, so I want to have valid certs for both www.example.com and example.com. I can get the cert for www.example.com working fine.
However the problem is in my naked domain. Whenever a http request goes to http://example.com/, it gets redirected to http://www.example.com/, ok. But, if a http request goes to e.g. http://example.com/a.html, the request is still redirected to http://www.example.com/. So when LE servers come looking for their well-known acme-challenge, it fails because they see index.html.
I guess this isn't a common behavior because no one is mentioning this, not at https://code.google.com/p/googleappengine/issues/detail?id=10802, nor at https://github.com/certbot/certbot/issues/1480.
I've tried to dig into why this is happening, one error I can see is here:
If I select to overwrite, GAE says it "failed to insert mapping"
This whole project was started by another person and he claims he's not aware of example.com being assigned anywhere else. I've looked at his Google Cloud Console and it would seem that he's correct.
Maybe something of interest is that in the Domain page of admin.google.com, naked domain redirect is set up. It redirects example.com to www.example.com. I've not found a way to disable it.
On my dns registrar, I have input the four A records, four AAAA records, and a www for CNAME.
TL;DR: My LetsEncrypt acme-challenge is failing for my naked domain, help!
I started getting the same error in app_engine after I went to Google Apps account and added example.com to redirect to wwww.example.com. After I did this, in app_engine I got "is already mapped" error. And there was no way to undo the redirect in Google Apps, so my guess was that Google Apps had mapped it and so App Engine could not modify it or add it. I had to explain this to Support team, and btw Google Apps support is free to call, so contact them and then get transferred to App Engine support team.
There is no way to fix it yourself, you have to get Google Support on call and explain clearly and they can reset. I was bounced between Google Cloud and Google Suites (Apps) support teams 7 times and after 2 weeks finally resolved, each one blaming the other, until I found a guy who understood this issue and fixed it for me.
Firstly, just letting you know I have searched a fair bit here and I am aware of some of the other questions on this topic but none answer my question.
The authentication of the Local GAE differs from the appspot deploy and I need it not to with minimal work-around code.
I'm writing an HTML5 app and I can do the google authentication via a button and it updates all the correct tokens so I can access the profile in either GAE Launched apps or appspot deployed ones.
I need the google account details of the logged in user within the app
I am writing (for API calls to calendar and contacts for example)
, and I'd rather not have to write a login handler only for my local development platform - automated for simplicity or otherwise.
I've read that adding login:required forces a login, and on appspot this works perfectly. Locally it does nothing useful.
I've read that you can write a Python decorator to use #login_required - but I'm not writing in Python (It's php generating an HTML5 page). I could write a bit of a PHP wrapper to handle it, or automate a call in Javascript on page load - but this is the workaround I don't want to write because it's handled in the production environment for me.
I want the login:required option as everything is handled for me in
production
. I have googled the options for the login tag and nothing there suggests I can force a google login in the locally launched app. I have googled the launcher and settings, but nothing seems apparent.
I suppose I could live with the dev workaround, and the app could assume I'm authenticated and the JSON request handlers in my app would just use the login:required with the correct google tokens being passed once I am authenticated.
Do I have any other options?
This sounds like it could be a PHP runtime bug. login:required works fine on the python local dev server. Have you checked the issues page to see if it's been reported?
https://code.google.com/p/googleappengine/issues/list
Otherwise it's possible that it's bypassing the login on your dev server because you have some cookie in your browser indicating that you're already logged in. You might try clearing cookies
Alternatively (at least on the python devserver), you can go to your login page via http://localhost:8000/_ah/login to force a logout (obviously fix the hostname and port number)
There appears to be no way round this other than to write the whole OAuth handler yourself (or get one elsewhere) - significant overkill for a development environment only 'issue'.
I have written the app to handle the getting of the google profile details as it starts and force an authenticate if they are not present.
This means that the login:required will work as expected in the production world and force you to authenticate to google before you even get to the application... then the app just gets the profile details because the tokens are already present.
login:required in the dev environment just puts up a screen which you just 'ok', then the app attempts to gets the profile details but forces the authentication itself because there are no authentication tokens present.
It's unfortunate, but it's a single step in a development that users will not have to use, but it works.
I am trying to access google app engine https://appengine.google.com/ from my newly created google apps account.
It shows an error "This webpage has a redirect loop" after login to my new google apps account. I am not sure how to fix this, i tried clearing cache & cookies, but it repeats the same.
I am not sure if it is the right place to ask the question. Please let me know if there is any different forum to post this question.
Please let me know if you need my email address.
Thanks,
Ramesh.V
Maybe it's the browser. Have you tried to authenticate in another browser? Or do you have a automatic login plug-in as last pass f. e. that's causing the loop?
There are also other people with this issue [1].
[1] http://code.google.com/p/googleappengine/issues/detail?id=8142
I had this same problem. It seems to be an issue with Federated Login. When I switched the app from Federated Login to Google Account API everything worked properly.
Federated login is still experimental and should not be used in a production environment in my view.
Cheers,
Kjartan S.
We have added a custom domain to our appengine app. We followed the instructions when we made our changes, but apparently something went wrong and we can't find the way to fix it.
In our google apps appengine tab, the main URL specified is https://appid.appspot.com. That means, however, that all traffic from the domain mappings will be sent to the https url, and of course this won't work. I don't know how this https url ended up there as in the app engine admin console, the app url is http://appid.appspot.com.
We haven't find the way to change this url. We have tried to disable this app in google apps but it didn't work, it stays there.
in your app.ymal
-secure: optional
for more details:
http://code.google.com/appengine/docs/python/config/appconfig.html
scroll to Secure URLs
Some of my clients are having problems accessing my Google App Engine website from typing in the URL. Despite it being written down - it's on a printed invite - as http://subdomain.domain.com (which works), people insist on putting in http://www.subdomain.domain.com
Is there some way of adding another mapping to make www.subdomain.domain.com point at subdomain.domain.com??
Help greatly appreciated.
You should be able to simply add that subdomain in Google Apps (following the instructions here). Failing that, you could use a third-party redirection service to send a 302 to your 'real' subdomain.