How to handle cookies across domains in development - ruby-on-rails

I developed two rails apps, they will each go on site.com and something.site.com respectively. In development mode they are run on localhost:3000 and localhost:4000, you must be logged into site.com to access something.site.com.
I have the cookies set up to handle domains from :all, and the tests all pass in relation to creating and destroying sessions. How ever I am unfamiliar how to test the cookies with out actually deploying the apps to production. I don't have staging servers, their small basic apps and I don't want them public. Is their any ways, on my mac, I can test (by launching both apps) the domain specific cookies?

You can edit your hosts file located at /etc/hosts to point 127.0.0.1 to local.site.com and then access your rails apps at local.site.com:3000 and local.site.com:4000
127.0.0.1 localhost local.site.com

I bet you could update your hosts file to point something.localhost to 127.0.0.1
127.0.0.1 something.localhost in `/etc/hosts
I'm sure you've read that as long as the TLD is the same, this should work.

Look at using POW and Powder. Runs slower, but allows you to run concurrent apps on localhost.
Pow User's Manual

Related

Rails tenanted app with CloudFlare and subdomains

I'm building a tenanted Rails application on Heroku which uses subdomains for each tenanted instance. The problem is I can't seem to get CloudFlare to properly cache a subdomain.
I have my Application hosted on heroku at say whatever.herokuapp.com
And using CloudFlare I'm pointing my main application at heroku's CNAME (whatever.herokuapp.com) and that works fine.
I'm using a wild cards *.mydomain.com on Heroku and then using CloudFlare again to point at them.
So say I create an app myinstance.mydomain.com on CloudFlare -> whatever.herokuapp.com
because I'm using the *.mydomain.com on Heroku it works fine.
BUT when I try to turn on the Acceleration (CloudFlare caching) it immediately breaks and all I see is a CloudFlare error Error 1001 telling me:
Most likely: if the owner just signed up for CloudFlare it can take a few minutes for the website's information to be distributed to our global network.
From their help desk:
Fundamentally in a standard set up the domain needs to be added to
CloudFlare in order for us to understand how to route traffic for that
domain.
If you wanted a setup where anybody in the world could CNAME to your
CloudFlare configuration, that is possible but it's an Enterprise
feature. We call it "Managed CNAME" - if that's something you're
interested in do let me know and I can put you in touch.
FYI, CloudFlare Enterprise averages at $5,000 a month.
CloudFlare is great in that it just works without much customization starting at zero cost, but unless your multi-tenanted app is for Bank of America or other Fortune 500 company I'd suggest something like Fastly ($5-50/month) or some other DIY caching solution that allows you to setup caching as you need rather than a one-size-fits-all top down solution.
https://devcenter.heroku.com/articles/fastly
Works directly with Heroku but there might be other solutions that work equally well.
That error can also occur if you're pointing a domain/subdomain not on CloudFlare to a domain that is on CloudFlare (creates a DNS resolution lookup error). I would probably contact CloudFlare support for additional assistance and troubleshooting.

Is my Rails 3.2.13 app running on my development machine being hacked?

I am using 'Thin' as my rails server in my development environment. I noticed something that alarmed me. I was taking a break... no keystrokes on my machine. The terminal window where Thin is running produced the following:
Started GET "/controller/method" for 127.0.0.1 at ...
Processing by ...Controller#method as HTML
Completed 401 Unauthorized in 58ms
Started GET "/users/sign_in" for 127.0.0.1 at 2013-08-16 11:47:02 -0400
Processing by Devise::SessionsController#new as HTML
...
Completed 200 OK in 178ms (Views: 22.3ms | ActiveRecord: 3.7ms)
Somewhere my development machine was responding to a request for a specific page that needed authorization and presented a log in screen to someone. It wasn't on my machine. No one else in my office has the technical ability, or the interest, to do a hack on my machine. So it must be coming from my wireless router or through our internet connection... i'm guessing.
In production I have an access log which gives an ip address for all those accessing our site. Is there something I can do to provide that kind of information? What can I do in my rails configuration to confirm that it was indeed a hack attempt?
I am addressing the rails process here. If there is a StackExchange site to address the security issues related to my router, internet connection and development machine? Which one(s) might it be?
Thanks.
It definitely seems like some automated hacking attempt. If you were running an older version of rails you could get it to execute arbitrary controller methods like that, or if you had set your routes file up improperly.
Just a reminder to set all of the controller methods you don't want to be visible to the outside world directly as private or protected. Also double check your routes file to ensure that you are avoiding less safe route definitions like match, if possible.
While this shouldn't entirely assuage your security concerns, being proactive about limiting the possible access attackers can gain will serve your project for the best. Additionally, you could set up thin to bind to 127.0.0.1 instead of 0.0.0.0, thereby preventing external requests (if possible).
Oh man... it was so simple. Can't believe it took me so long to figure out. Safari has the "top sites" feature. Whenever I open a new tab, top sites is the page that is displayed. Some of my top sites are fixed by dragging a url to the top sites icon. The remainder are composed of other pages that are among the most frequently displayed. After I alter code I frequently view the page to make sure there are no errors. Top sites is trying to display some of those pages. Viewing those pages requires authorization. Top sites is trying to hack my development server. :=]

OAuth: how to test with local URLs?

I am trying to test OAuth buttons, but they all (Facebook, Twitter, LinkedIn) come back with errors that seem to signal that I can not test or use them from a local URL.
How do people usually work in development with OAuth stuff if they all seem to require a non-dev and non-local connections environments?
Update October 2016: Easiest now: use lvh.me which always points to 127.0.0.1, but make sure to verify that this is still true every time you need to invoke it (because domains can expire or get taken over, and DNS poisoning is always a concern)
Previous Answer:
Since the callback request is issued by the browser, as a HTTP redirect response, you can set up your .hosts file or equivalent to point a domain that is not localhost to 127.0.0.1.
Say for example you register the following callback with Twitter: http://www.publicdomain.com/callback/. Make sure that www.publicdomain.com points to 127.0.0.1 in your hosts file, AND that twitter can do a successful DNS lookup on www.publicdomain.com, i.e the domain needs to exist and the specific callback should probably return a 200 status message if requested.
EDIT:
I just read the following article: http://www.tonyamoyal.com/2009/08/17/how-to-quickly-set-up-a-test-for-twitter-oauth-authentication-from-your-local-machine, which was linked to from this question: Twitter oAuth callbackUrl - localhost development.
To quote the article:
You can use bit.ly, a URL shortening service. Just shorten the [localhost URL such as http//localhost:8080/twitter_callback] and register the shortened URL as the callback in your Twitter app.
This should be easier than fiddling around in the .hosts file.
Note that now (Aug '14) bit.ly is not allowing link forwarding to localhost; however Google link shortener works.
PS edit: (Nov '18): Google link shortener stopped giving support for localhost or 127.0.0.1.
You can also use ngrok: https://ngrok.com/. I use it all the time to have a public server running on my localhost. Hope this helps.
Another options which even provides your own custom domain for free are serveo.net and https://localtunnel.github.io/www/
Or you can use https://tolocalhost.com/ and configure how it should redirect a callback to your local site. You can specify the hostname (if different from localhost, i.e. yourapp.local and the port number). For development purposes only.
For Mac users, edit the /etc/hosts file. You have to use sudo vi /etc/hosts if its read-only. After authorization, the oauth server sends the callback URL, and since that callback URL is rendered on your local browser, the local DNS setting will work:
127.0.0.1 mylocal.com
Set your local domain to mywebsite.example.com (and redirect it to localhost) -- even though the usual is to use mywebsite.dev. This will allow robust automatic testing.
Although authorizing .test and .dev is not allowed, authorizing example.com is allowed in google oauth2.
(You can redirect any domain to localhost in your hosts file (unix/linux: /etc/hosts))
Why mywebsite.example.com?
Because example.com is a reserved domain name. So
there would be no naming conflicts on your machine.
no data-risk if your test system exposes data
to not-redirected-by-mistake.example.com.
You can edit the hosts file on windows or linux
Windows : C:\Windows\System32\Drivers\etc\hosts
Linux : /etc/hosts
localhost name resolution is handled within DNS itself.
127.0.0.1 mywebsite.com
after you finish your tests you just comment the line you add to disable it
127.0.0.1 mywebsite.com
Google doesn't allow test auth api on localhost using http://webporject.dev or .loc and .etc and google short link that shortened your local url(http://webporject.dev) also bit.ly :). Google accepts only url which starts http://localhost/...
if you want to test google auth api you should follow these steps ...
if you use openserver go to settings panel and click on aliases tab and click on dropdown then find localhost and choose it.
now you should choose your local web project root folder by clicking the next dropdown that is next to first dropdown.
and click on a button called add and restart opensever.
now your local project available on this link http://localhost/
also you can paste this local url to google auth api to redirect url field...
This answer applies only to Google OAuth
It is actually very simple and I am surprised it worked for me (I am still sceptical of what my eyes are seeing).
Apparently you can add localhost as a trusted domain on the Google Developer Console, since localhost is an exception for most rules as you can see here.
This can be done on this page under OAuth 2.0 Client IDs. Click edit and then add http://localhost:8000 or similar ports, and hit save.
It is crucial that you include http or https in the input box.
HTTP or HTTPS?
I am once again surprised that Google allows http, although do note that there is a minor security risk if your application has been released to production.
If you want to be extra cautious, you can choose to stick with https. This will require you to set up an SSL certificate on your localhost server.
This is easier than you think, since the SSL certificate needs not be valid. Many http servers should give you this option. You will have to click the "proceed anyway" button anyway in your browser to bypass the big red warning.
This is more secure than http since either a) users will see a big red warning if hackers are trying something phishy, or b) the only time they won't see this warning is if the user intentionally set up a self-hosted SSL certificate, in which case they probably know what they are doing (I suppose a virus could technically do this as well, but at that stage they've already gotten enough control of a user's system to do anything they want).
I ran into some issues with the tools mentioned in other answers such as http://tolocalhost.com not forwarding query parameters (not to mention you have to visit the page and configure it first, same case with https://thomasmcdonald.github.io/Localhost-uri-Redirector/) and http://lvh.me not being useful to me because I run a proxy on my local machine and need the public URL to point to a private URL like http://mywebsite.dev.
So I made my own tool that filled my needs and may fill yours:
https://redirectmeto.com
Examples:
https://redirectmeto.com/https://www.google.com/search?q=puppies
http://redirectmeto.com/http://localhost:4000/oauth/authorize
http://redirectmeto.com/http://client.dev/page
Another valuable option would be https://github.com/ThomasMcDonald/Localhost-uri-Redirector. It's a very simple html page that redirects to whatever host and port you configure in the UI.
The page is hosted on Github https://thomasmcdonald.github.io/Localhost-uri-Redirector, so you can use that as your OAuth2 redirect url and configure you target host and port in the UI and it will just redirect to your app
If you have a domain, you can create a subdomain that redirects to your local entry point, it works for me
I created a public subdomain : oauth-test-local.alexisgatuingt.fr that redirects you to http:127.0.0.1:8000/oauth/callback/google with the returned data
Taking Google OAuth as reference
In your OAuth client Tab
Add your App URI example(http://localhost:3000) to Authorized JavaScript origins URIs
In your OAuth consent screen
Add mywebsite.com to Authorized domains
Edit the hosts file on windows or linux Windows C:\Windows\System32\Drivers\etc\hosts Linux : /etc/hosts to add 127.0.0.1 mywebsite.com (N.B. Comment out any if there is any other 127.0.0.1)

Need to link WP Blog with Rails App on Heroku

I have a client who wants to migrate his Rails app to Heroku. However the client also has a blog associated with his domain that runs on WordPress. Currently, the WordPress blog is running happily alongside the Rails app, but once we migrate to Heroku, that clearly won't be possible.
The url for the app is like http://mydomain.com, and the url for the blog is like http://mydomain/blog.
I realize that the best long-term solution is to redo the blog in a Rails format like Toto or Jekyll. But in the short term, what is the best way to continue hosting the WP blog where it is (or somewhere) but use Heroku to run the app? The client doesn't want the blog to be on a subdomain, but to remain at mydomain/blog for SEO reasons and also since there is traffic to the blog. I have two ideas:
Use rack_rewrite or refraction (or just a regular old 301 and Apache mod_rewrite) on the old (non-Heroku) server to redirect the main url from the old site to Heroku. In this case, I can just leave the Wordpress blog running happily where it is. I think?? Is there a reason to choose one of those options (rack_rewrite, refraction, or mod_rewrite) over the others if I do it this way?
Switch the DNS info to point to the Heroku site, and then use a 301 redirect from the blog to the old site. But then I'll have to get the old (non-Heroku) site on a subdomain and use some kind of rewrite rules anyway so it looks like it isn't a subdomain.
Are either of these approaches preferable, or is there another way to do it that's easier that I'm missing?
The only tenable long term/scalable solution would be to host the blog permanently on a sub-domain or different domain and add a redirect from mydomain.com/blog to the new location (ie: blog.mydomain.com).
You would need a single server running a front-end like Apache/nginx on mydomain.com to serve up mixed back-ends like Rails and Wordpress and that is not possible on Heroku.
Sadly, this is where you need to dig in as a consultant and be stern with your client about the technical limitations.
Why does you client want to migrate to Heroku? Is there a larger goal behind that you could accomplish with different hosting where you control the front-end and can mix in different back ends?
Another solution is to set heroku to http://app.example.com, and Wordpress to http://example.com. You put your Wordpress-landing page in the root , and blog on /blog. When the user click "login" or "signup" on the landing-page, they're linked to the heroku-app.
This will be optimal in a SEO-perspective, but require some DNS-knowledge.
Winfield's answer is not correct. You can run a reverse proxy on your rack server (via Heroku) to direct to the blog, wherever it may be.
See https://github.com/jaswope/rack-reverse-proxy
After you install the gem and set up your app according to the docs, your ./config.ru file will have something like this:
use Rack::ReverseProxy do
reverse_proxy(/^\/blog(\/.*)$/,
'http://<app-name>.herokuapp.com$1',
opts = {:preserve_host => true})
end

Subdomains and locally installed Rails app

I can't figure out what I'm overlooking, perhaps it's obvious or lack of understanding.
The app I'm working with uses subdomains which on the hosting server work properly. I figured locally installing would kick up some issues around routing, so I read up on making changes to /etc/hosts and using the Ghost gem. Both seem to work fine i.e. localhost:3000/ becomes myapp.local:3000 but I don't understand how to go about logging into a subdomain account. Here's an example...
myapp.local:3000/session/new = the default login page for the app
myapp.local:3000/signup = default signup page
I can create an account here e.g. Sub1
The thank you page is shown w/ the reference to sub1.myapp.com which points to the hosted app (the local db shows this domain as well)
sub1.myapp.local manually added to /etc/hosts and dscacheutil -flushcache
sub1.myapp.local:3000/session/new is the subdomain
login attempts return that this isn't a valid domain. This seems to make sense because the local db shows the url as sub1.myapp.com on the hosting server.
So my question is whether there's a local workaround that I can use for development or have I totally missed a fundamental concept along the way?
you might just want to try putting the actual dot com in your /etc/hosts file.
ie:
127.0.0.1 sub1.myapp.com
127.0.0.1 myapp.com
127.0.0.1 anyothersubdomains.myapp.com
what this usually does is trick your computer into thinking it is the host of all of those, so you can't go to the real site anymore in a web browser.
if you do want it to be .local, presumably so that you can refer to the real online site while working on a local copy, you should probably take a look in app/controllers/application_controller.rb (sometimes application.rb) and look for logic in there that helps determine what to do depending on the subdomain. maybe its hard coded to only look for a .com or something.
If you are using the webrick server or something like Puma for development you can use lvh.me to access your subdomains. e.g.
http://sub.lvh.me:3000/
http://lvh.me:3000/ is equal http://localhost:3000/

Resources