Disqus - Latest Comments for dcmanges

Re: Introducing Braintree.js

Dan Manges — Fri, 08 Mar 2013 16:17:12 -0000

Braintree.js works everywhere we're available. You can see the countries we support at https://www.braintreepaymen.... Our full server-side API can be used in harmony with Braintree.js, and it supports authorize/capture functionality.

Re: Wrestling with Bundler

Dan Manges — Mon, 04 Feb 2013 10:46:38 -0000

Hi Benjamin -- thanks for the comment. I'm planning on using the recipe once I get to that step. The primary reason I didn't want to do it initially is because it makes rolling Bundler out more complicated since the deployment script change needs to move through environments (qa -> staging -> production) in conjunction with the code change. I'd also typically release Bundler to a single application server to make sure the application and background jobs work smoothly before rolling it out everywhere. If possible, I thought using existing deployment scripts would be best.

Re: How We Moved Our Data Center 25 Miles Without Downtime

Dan Manges — Mon, 16 Jan 2012 11:57:09 -0000

OmniGraffle

Re: How We Built the Software that Processes Billions in Payments

Dan Manges — Wed, 17 Aug 2011 12:04:55 -0000

We found a way to release a minimum viable product early on. I think the first release was within a couple months of when we began development, and included only credit card vault storage, no transaction processing.

Re: ActiveModel::Validations and Command/Query Separation - Dan Manges's Blog

Dan Manges — Thu, 11 Aug 2011 12:23:37 -0000

A new user with no attributes set isn't valid (assuming validates_presence_of), so why shouldn't it have errors? I understand that the UI shouldn't display those errors, but I think it would make sense for user.errors to return them.

Re: ActiveModel::Validations and Command/Query Separation - Dan Manges's Blog

Dan Manges — Wed, 10 Aug 2011 10:30:59 -0000

I agree there are a lot of considerations to make, and that many Rails plugins and conventions have been built on the current behavior. Personally, I think running validations only when calling a validate method is very clear and intentional and addresses your comment of "deciding on my own if I want to perform the validation check." It would be easy for form helpers to continue to work: if you pass them the errors object (as a result of calling validate) they can display inline errors and maintain their current behavior.

Re: ActiveModel::Validations and Command/Query Separation - Dan Manges's Blog

Dan Manges — Wed, 10 Aug 2011 10:26:29 -0000

Thanks for sharing your thoughts. I agree that it's not an improvement with regard to the elegance of the API (because of the local variable assignment), but I do think removing the side effect of calling valid? would be an improvement on usability. Personally, I don't like to accept "once you understand" as a defense for code. And I don't mean that just with Rails, but in general. If I've written some code trips up other developers, I try to avoid defending it that way. Ideally, the code would be easy to understand and intuitive to work with, and not require any special understanding or learning curve.

Re: Client-side Encryption - Braintree

Dan Manges — Fri, 01 Jul 2011 12:24:15 -0000

Any merchant accepting credit card payments has to be PCI compliant. The question is which requirements are in scope for achieving compliance. End-to-end encryption should eliminate many requirements from being in scope. I say "should" because unfortunately much of PCI compliance is subjective to the QSA who is performing the assessment. We've been talking to various QSAs to determine where companies stand on performing an assessment of a merchant using end-to-end encryption.

Re: How We Built the Software that Processes Billions in Payments

Dan Manges — Thu, 30 Jun 2011 19:12:21 -0000

PostgreSQL. Maybe we'll write another blog post to expand on that. We're using Mongo for log data.

Re: We Fab Braintree Payments

Dan Manges — Wed, 29 Jun 2011 17:41:18 -0000

This is Dan from Braintree. Thank you for the kind words. We're happy to be working with Fab.com and are excited about your growth and success.

Re: Braintree’s $34 Million Accel Investment: Yesterday, Today & Tomorrow

Dan Manges — Wed, 29 Jun 2011 16:36:58 -0000

Thanks for the note, Andy. It looks like there was a caching problem that caused both domains to serve the same page. We've fixed it.

Re: Braintree’s $34 Million Accel Investment: Yesterday, Today & Tomorrow

Dan Manges — Wed, 29 Jun 2011 15:49:37 -0000

Thanks Daniel! Likewise, we love using Disqus. I've been using it on my personal blog for a long time as well; I think I was one of your first users (account name dcmanges).

Re: m2node - a mongrel2 handler for node.js - Dan Manges's Blog

Dan Manges — Mon, 06 Jun 2011 12:55:07 -0000

Thanks. The FakeSocket is complex, and I'm not sure if there are any edge cases or scenarios that I'm not handling with it yet, but I thought that it was worth it to have a drop-in implementation.

Re: m2node - a mongrel2 handler for node.js - Dan Manges's Blog

Dan Manges — Mon, 06 Jun 2011 05:07:25 -0000

Whoops, that was sloppy of me. Thanks for letting me know, I fixed the link.

Re: m2node - a mongrel2 handler for node.js - Dan Manges's Blog

Dan Manges — Sun, 05 Jun 2011 20:18:32 -0000

Testing comments to make sure they still work since this is the first blog post I've written in two and a half years.

Re: Blue Ridge Pro Tip: Trust the Wiring, Test the Behavior

Dan Manges — Wed, 02 Sep 2009 13:26:08 -0000

What is your opinion on how to test that the event handlers are wired up properly? You can trust jQuery to do the right thing, but you can still make a mistake with the selector or delegation when setting up the event handler. Leave that up to manual testing?

Re: Dan Manges's Blog - Asset Versioning in Rails

Dan Manges — Sat, 08 Aug 2009 22:16:58 -0000

IIRC, FilesMatch doesn't support the query string.

Re: Make your test suite UNCOMFORTABLY FAST!

Dan Manges — Wed, 29 Jul 2009 03:32:00 -0000

There’s also DeepTest: http://github.com/qxjit/dee...

It works with test/unit and RSpec. In addition to running tests in parallel to take advantage of multiple cores, it can run tests distributed to multiple computers.

Re: Make your test suite UNCOMFORTABLY FAST!

Dan Manges — Wed, 29 Jul 2009 03:32:00 -0000

There’s also DeepTest: http://github.com/qxjit/dee...

It works with test/unit and RSpec. In addition to running tests in parallel to take advantage of multiple cores, it can run tests distributed to multiple computers.

Re: Dan Manges - Ruby Pattern: Extend through Include

Dan Manges — Sun, 25 Jan 2009 18:04:16 -0000

How are you running the 2nd example? If in IRB, you should see the output. If as a script, you might want to put a "puts" before "Person.verbal_object_id"

When you extend as part of the class definition itself, the methods in the module become methods on the class.

Re: Install Multiple Versions of Ruby on Leopard - Dan Manges's Blog

Dan Manges — Tue, 06 Jan 2009 11:14:56 -0000

I'm not sure. You could probably put /usr/local/ruby in your path first and be fine, but it might be good to delete all signs of ruby from /usr/local/lib to prevent running into any weird problems in the future.

Re: Dan Manges - Rails: UnitRecord - Test without the Database

Dan Manges — Fri, 12 Dec 2008 15:53:27 -0000

You should still have some tests that hit the database. I usually set up unit tests for models that use UnitRecord, and then separate functional tests for models that hit the database.

Re: Action Dependent Validations and Why :on => :update is Bad - Dan Manges's Blog

Dan Manges — Thu, 04 Dec 2008 13:25:23 -0000

That's definitely a good solution if the validation corresponds to a state.

Re: Action Dependent Validations and Why :on => :update is Bad - Dan Manges's Blog

Dan Manges — Thu, 04 Dec 2008 10:20:27 -0000

Password wasn't a good example. The point was more about running validations on update in general. Another example might be accepting new terms and conditions on update. You can't do validates_acceptance_of :terms_and_conditions, :on => :update, because not every single update to the object will need that. You would only want to validate it on the actual edit screen for the user.

Re: Action Dependent Validations and Why :on => :update is Bad - Dan Manges's Blog

Dan Manges — Thu, 04 Dec 2008 10:17:05 -0000

Password wasn't a good example. Of course, I wouldn't store plain text passwords anyway. There could be attributes you want to collect on update that you can't randomly generate. I think filling in a temporary value would add a lot of complexity considering you wouldn't want to display those temporary values anywhere.