Stripeinvalidtransit

Stripe.js Bank Accounts: Invalid transit number. The number should be in the format xxxxx-yyy or xxxxxyyy.

coding software

This is going to be a quick one. I was working on a client project that uses Stripe Connect. We were using Stripe.js V2 to generate tokens for bank accounts. However, everytime I tried to use the test account numbers, I kept getting this error:

Invalid transit number. The number should be in the format xxxxx-yyy or xxxxxyyy.

After looking around a bit, I realized it was a really stupid error. I was trying to test a Canadian account, but when I went to look at the test accounts that Stripe provides, "United States" was selected. So lesson of the day, when testing Stripe Test Bank Accounts, make sure you select the proper country.

Hopefully that helps out some peeps.

Caseyli
Casey Li
CEO & Founder, BiteSite
Swiftputpatch

Swift PUT and PATCH requests not sending httpBody

coding software

Hey everyone, this will be a quick one.

I'm finally coding a Swift iOS app to interface with my Rails app and I'm now at the point where I'm starting to implement PUT or PATCH requests for my record updates.

For my requests, I have a basic APIController that runs this code to build the request:

urlRequest = URLRequest(url: url) urlRequest.httpMethod = "PATCH" urlRequest.httpBody = bodyString.data(using: .utf8)

This method generally worked perfectly fine for my POST requests, but for some reason PUT and PATCH requests were not sending the httpBody.

Turns out, for PUT and PATCH you need to be explicit about the content type. Since I was using form urlEncoded data (so something that looks like "key=value&key=value&key=value" I just had to add this line:

urlRequest.addValue("application/x-www-form-urlencoded", forHTTPHeaderField: "Content-Type")

And then everything worked.

So final code looked like this:

urlRequest = URLRequest(url: url) urlRequest.httpMethod = "PATCH" urlRequest.addValue("application/x-www-form-urlencoded", forHTTPHeaderField: "Content-Type") urlRequest.httpBody = bodyString.data(using: .utf8)

Hope that helps out some peeps!

Caseyli
Casey Li
CEO & Founder, BiteSite
Amazons3googledocspreview

Google Docs Viewer with Amazon S3 Private Documents and Presigned URLs

coding software

So recently one of our developers discovered an awesome service that Google Docs provides called the Google Docs Viewer. It's super powerful and super easy to use. Basically, it's a service that allows you to pass a URL of a document to this service, and it will render a preview of the document that's usable in the DOM. This is amazing for providing previews of common documents like PDF, DOCs and more. We saw Trello use it so we thought it's gotta be a good solution.

Here's an example:

<iframe src='https://docs.google.com/viewer?url=http://www.bitesite.ca/Signing%20up%20for%20Amazon%20S3.pdf&embedded=true'></iframe>

You could slap that iFrame into your HTML and boom, you have a preview of a PDF document embedded in your website.

However, the app that we were building was using private, protected documents hosted on Amazon S3. So to pass a URL to the Google Docs Viewer, we had to get a presigned URL. So it would look something like this:

<iframe src='https://docs.google.com/viewer?url=AMAZON_S3_PRESIGNED_URL&embedded=true'></iframe>

As soon as we did this, we kept getting a "No Preview Available" error from Google Docs. After trying lots of solutions, it turns out the error had to do with the fact that we were passing a URL with Query String params as a query string param to another URL. That is, the Amazon S3 Presigned URL contains query string parameters, so it contains a "?" and "&" which confuses the outer URL.

So long story short, you have to Encode the S3 Presigned URL before passing it to the Google Doc Viewer.

If you're using Presigned URLs, my guess is that you're dynamically setting the URL. We do it with Javascript, so our code looks something like this:

var encodedUrl = encodeURIComponent(amazon_presigned_url);
var iFrameUrl = 'https://docs.google.com/viewer?url=' + encodedUrl + '&embedded=true';

We did run into an issue where we tried encodeURI instead of encodeURIComponent, but that was just our misunderstanding of those methods. If you're doing server-side encoding, make sure you're it's correctly removing "/", "?", and "&".

Alright, that's it for now. Hopefully this helps out some peeps.

Caseyli
Casey Li
CEO & Founder, BiteSite
Reactnativeallwoarbitraryloads

React Native fetch only working with localhost and NOT IP addresses

coding software

Alright, this will be a quick one. So I was using 'fetch' to make a web request in my React Native app. My server was a Rails app. When I used 'localhost' in my request, the request would work perfectly, but if I used my IP address, the network request would fail.

fetch('localhost:3000')   /* this would work fine */
fetch('192.168.1.41:3000')    /* this would not work */

I double-checked to make sure my Rails server wasn't the issue and ensured it was bound to 0.0.0.0

rails s -b 0.0.0.0

So that was good. In fact, if I went into my iOS simulator and launched Safari, Safari could hit the localhost:3000 and IP Address:3000 no problem. So my iOS Simulator could hit my Rails app, but not my React Native app using the IP Address. Again, everything worked fine using localhost.

Turns out this is an iOS 9 issue where by default, HTTP requests are blocked (vs HTTPS requests). So to allow this you need to enable "Allow Arbitrary Loads" in your Info.plist in your XCode project.

For more details, check out this link https://facebook.github.io/react-native/docs/running-on-device-ios.html#app-transport-security

Caseyli
Casey Li
CEO & Founder, BiteSite
Bangoperators

Careful with Bang Operators in your Rails Models and ActiveRecord Callbacks (Frozen String error)

coding software ruby on rails

Alright here's a weird one that would have been easy to solve if I thought more about how Ruby works (and how it's not like Java:)).

So here was the issue, we were looping through a set of keys on a hash. Now, you should know that the keys on this hash were strings. So we had something like this

my_hash = {"color" => "blue", "size" => "medium" }

Now, on top of that we had a model that we'll call Person that had a "name" attribute and looked something like this:

class Person < ActiveRecord::Base
  before_validation upcase_name

  private
    def upcase_name
      name.upcase!
    end
end

Ok, for those who don't know, the bang version of upcase, upcase!, uppercases a string in place - it modifies the actual string itself.

So here is the setup, we were looping through keys of the hash:

my_hash.keys.each do |name|
  ...
end

Now, first thing to know is that you CAN'T modify a key in a hash. Anytime you try to modify the keys of a hash, you get an error. So something like this would produce an error:

my_hash.keys.each do |name|
  name.upcase!    # produces frozen string error
end

That all makes sense.

What we were surprised to see was that the following code also produces a similar error:

my_hash.keys.each do |name|
  Person.create(name: name)    # produces frozen string error
end

What's going on here? Well it turns out that the "before_validation" callback tries to upcase the name. But unlike languages like Java, the string that is passed in is the exact same object that exists in the "my_hash.keys" array. (I'm purposely avoiding by-value and by-reference wording here, because I feel it can lead to confusion)

So the hash holds a frozen lock on that string, and then when that string gets passed into the Person constructor, and the callback tries to modify it - you get the frozen string error.

How do you fix it? Change your callback:

class Person < ActiveRecord::Base
  before_validation upcase_name

  private
    def upcase_name
      name = name.upcase
    end
end

This version doesn't modify the original string but rather produces a new string.

Special thanks to d3chapma for helping me figure this one out. Check out his Gist here for more info.

Hope this helps out some peeps!

Caseyli
Casey Li
CEO & Founder, BiteSite