Hi,
Someone called ech0came into the #rubyonrails channel on Freenode and asked a question that was regarding your library. In trying to help them debug their issue I came across theclasses.rb` file in your library and was... shocked. For three reasons.
One, and this my favourite thing: You've documented things. You have spent some of your time documenting things! Most people don't do this and it grinds my gears. Thank you thank you thank you for doing this. Awesome work!
Secondly: There is so much code. You had to have spent a lot of time writing and testing all of this. That's another thing: TESTS. There are _actual _real* tests* for this. Amazing.
Thirdy, and more seriously...
Throwing all the classes in the one file is a developmental nightmare in Ruby, for a number of reasons. The first of course being that finding anything in that file becomes hard. Like, I tried finding the Character class (it's here btw) to help our friend, thinking it would be somewhere else.
That somewhere else would be lib/reve/character.rb, and that file would define just the Reve::Character class. A grand total of 15 lines of code.
By doing this, any developer working on this project, or anybody wanting to know how exactly this thing works would be able to go into their Editor of Choice (read: Vim, obviously) and then find the lib/reve/character.rb file to find out how exactly this Reve::Character class works.
Added bonus! Things like this class accessor would go into this class. Like this:
module Reve
class Character
def self.url
"http://api.eve-online.com/account/Characters.xml.aspx"
end
end
endYou could go one step further. See that first part of the link? http://api.eve-online.com/. You could move that into the Reve module and reference that everywhere you need to, rather than having it in a million (ok, that's hyperbole, but you get the point!) different places.
Finally, this would also allow you to do things like this in your library, which I think is the coolest part:
lib/reve.rb
module Reve
autoload :Character, "reve/character"
...
end
While this seems innocuous, what it actually allows you to do is really cool. People only need to require 'reve' in their libraries / applications. When they attempt to reference Reve::Character for the first time, it will autoload the file specified, basically doing this:
Next time, it's there and loaded and so it won't be re-required. The benefit of this is that it will only load the classes when needed to, meaning that Reve won't need to load 1000+ lines of code to begin with and will only load the things it needs to load when it needs to load them.
So.
What you can do is this:
- Investigate how you can split the classes into separate files.
lib/[gem]/[class].rb is the format that all the cool kids are doing. I think we should stay with the cool kids, primarily because they're cool.
- Move anything to do with a class into the class itself. Things like this where you're retrieving a list of characters would be better off inside the
Character class. That way, people could do Character.all or something like that.
- Split your tests out as well. Currently it's all in one huge file. This just leads to better organisation. To add a new feature to the
Character class for example, I would have two buffers open in Vim: one for the Character class itself and one for the test. Then I would do my changes. As it is right now, it's very painful to do this.
I believe in your abilities enough to fix this. From what I've seen, you've not written horrible code, it's just unorganised and that's really easily fixed.
Please don't take this as a criticism of your code, it's not. It's just some friendly advice from another Ruby programmer who obsesses probably a bit too much over the organisation of everything. I'm just trying to help.
Congratulations if you've read this far, and good luck!
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent