I’ve starting work on a rewrite for Feedjira and I wanted to try using a GitHub issue to discuss what I’d like to see in the new version. I’m very interested in what users think, so please chime in!!
In no particular order, here are some design goals for the new version:
Change HTTP Library
This gem depends on curb, but I’d like to switch that to something that will allow me to support JRuby. I did some experimenting and really liked working with Faraday, so that’s what I’m using at the moment.
Throwing an array of feed urls at Feedjira and watching it concurrently fetch and parse them extremely quickly is cool in READMEs and demos, but I don’t think it’s reality. I think what’s much more likely is a single url being passed and concurrency being dealt with in the application, workers being the likely approach.
Feedjira has a dirty little secret and it’s the
Feedjira::Feed class. If you
open that sucker up, you’ll see 19 class methods and 0 instance methods. These
class methods do everything from fetching and parsing to setting up curb and
unzipping raw xml. Not only is SRP off in the corner crying, its a mess to
test. I’d like to break this down into smaller objects that get instantiated and
passed around in normal OO fashion.
One of the things I like best about Feedjira is the ability for a user to define their own custom parsers and I think its under-used. What I’d like to see is Feedjira ship with a parser for Atom and a parser for RSS - the others are really just custom parsers.
Custom Parser Project
With version 1.2, I was able to spin off the benchmarks into their own project and I was very happy with how it turned out - I’d like to see a similar thing happen with the custom parsers mentioned above. You should be able to easily grab these parsers and throw them in your application. If you write a particularly cool custom parser, then you could contribute it back and others could benefit from it.
Raise Exceptions Instead of Callbacks
Feedjira almost never raises an exception and instead provides the ability to register callbacks for success and failure situations, but this has proven brittle and painful. With the new version, I want to see sensible exceptions raised when something goes wrong and that pretty much removes the need for callbacks, so those will be cut.
Feed Updating is Business Logic
The updating functionality of Feedjira leaves a lot to be desired - its inconsistent and there are lots of edges. I tend to steer users away and encourage them to write this kind of code in their application instead. See this discussion for more, but needless to say, I’d like to see this entire feature set removed.
I want configuration for Feedjira done in a proper config block like most gems. If you want to change the default parsers, that should be a config option. If you want to set the user agent, that should be a configuration option.
Do you have any other ideas to make Feedjira better? Please check out the GitHub issue and share what you’re thinking. I’m open to any feedback, so if I just threw your favorite feature on the chopping block, please tell me!