-
Notifications
You must be signed in to change notification settings - Fork 677
Developer Guide: ThinkUp for Beginners by a Beginner
For a beginner developer, finding an open source project to work on can be difficult and daunting. Lots of the projects you know and love are already well established with large communities of developers and their own best practices that might seem alien to you. This is where ThinkUp starts to look really good. ThinkUp has a small developer base (at the time of writing this, 26 developers total) and a well designed and executed development process.
I'm going to try and walk you through all of the things you need to know before diving into the code. I am a beginner developer myself and have been contributing to ThinkUp for a few months so I know what it feels like at first!
This page is aimed at the beginners. People who have not done much open source work or group coding and people unfamiliar with the ThinkUp code base. Sections of this page may apply to some but not to others. That's okay, you're free to skip over parts you feel comfortable with. For example, if you're happy with how Git works then you can skip over that part. If you know how the MVC design pattern works you're probably cool to just breeze over that, too.
This page does, however, assume that you're familiar with object orientation, inheritance, abstraction, version control and the purpose of ThinkUp. If any of these concepts look alien to you, I recommend you look them up before continuing.
First things first, you need to get yourself a copy of the code. I won't go through this because there is already a fantastic guide written for it which can be found here. I will make a few notes for beginners, though.
When I first started developing for ThinkUp the scariest thing was the version control. I had never used Git before and it was totally alien to me (parts of it still are, but that's for the next section). I also didn't understand how open source software was developed so here is a brief run down for you.
You may be wondering: How does open source software get developed with lots of people making changes at the same time without descending into total chaos? The answer is very simple: version control and branching. Examples of version control packages are Subversion, Mercurial, CVS and, of course, Git. A branch is simply a part of the project that is being developed on by someone. For example, one of my branches is "429-wordpress-plugin-cleanup", the 429 is the issue number related to it and the rest is a description of what I'm doing on the branch.
If you're hopelessly confused as to what version control software is, it's quite easy to understand. It's literally just some software that helps you manage source code by keeping numerous copies of files so that you can roll back to certain points in time. It also allows you to create branches of code which are just exact copies of the code that you can develop on and, if you no longer like what you're doing, you can delete the branch safely without affecting the main code base.
A lot of open source software uses the branch model and the way it works is that there is a branch called "master" which is the main branch of the project (sometimes this is called the "trunk"). All changes that are made will eventually end up in there (provided they pass review). You, as a developer, create your own branch of development, write your own code on it and, when you think it's ready, you submit it to Gina for review and she provides you with feedback if changes need to be made. If it's all good and doesn't need changing, Gina will merge your changes into the master branch and you'll end up on the Contributor list. Cool, huh?
While version control sounds like a fantastic idea, it is the opposite of beginner friendly. It is best to do it from the command line which sounds scary if you haven't used the command line before but once you get used to it, it's really not that bad. I'm assuming at this point that you've read the guide to getting the ThinkUp source code and know some basic command line skills such as navigating directories. I'm also assuming that you have navigated to the ThinkUp directory as none of the following commands will work unless you're in the directory you fetched the Git source code to.
The key to developing on ThinkUp is branching. All of the commits you do will go onto your own branch that you create. What you have to do is create a branch, commit changes to it, send it to Gina. Here's how you do it.
Creating a branch is easy but there are a few checks you need to do before you do it. Most of your branches will branch off from the "master" branch. When you create a branch in git via the command line, the branch is created as an off-shoot from the branch you're currently on. This may or may not be the master branch. To check, use the git status command. This will display the branch you are on and what changes exist.
So make sure you're on the master branch using the git status command. In the event that you aren't on the master branch, execute the following command:
git checkout master
And that should switch you to the master branch (if you want to make sure, you know how!). Branches usually correspond to issues from the issues list. Every issue has a number and this number is what the name of your branch would start with. For example, if you wanted to tackle issue 338, broken links when install directory has a space in it you would create your branch like this:
git branch 338-fix-broken-link-when-install-dir-contains-space
Or something along those lines. Then you need to checkout (switch to) that branch to start work on it.
git checkout 338-fix-broken-link-when-install-dir-contains-space
Now you're ready. If you do some work on the code and you're not on the right branch don't worry, you can just checkout the branch you were meant to be working on and everything will be fine. The problem arises (and I assure you everyone has done this at one point) when you commit to the wrong branch but first we need to cover commits.
What's a commit? When you've written some code and you get to a point where you feel you're done or you've got to a point where you're happy and ready to have the code looked at by other people you do what's called "committing". This saves the changes you've made to the branch along with a little message that explains what you've done. Before you commit you need to "add" files to the commit. If you do a git status after making changes to the code, you should see a list of files that you've edited with "modified" to their left and under the heading "Changed but not updated". These files need to be added to what's called the "stage", to demonstrate this let's imagine we edited the "class.AppConfig.php" file in the "model" folder. Listed in our modified files would be webapp/_lib/model/class.AppConfig.php, so let's add that to the "stage":
git add webapp/_lib/model/class.AppConfig.php
That should do it. Now when you do git status the file should be listed under "Changes to be committed". This means that when you commit, this file will be part of the changes you submit to the main ThinkUp project. Make sure they work! When you're ready:
git commit -m "Issue 338: Fixed the broken flux capacitor."
Obviously your commit message won't include references to time travel but rather references to what the commit achieves. Now, what if you file your commit only to realise you accidentally did it on the master branch? Not to worry! There's a command for that:
git reset --soft HEAD^
That command will undo the last commit on the current branch in "soft" mode, meaning that all of the changes get put back into the "changes to be committed" list. You may be thinking, what's all this HEAD lark? And why is there a caret after it? It's easy to explain: HEAD is a keyword that refers to the latest commit on a given branch. The caret after means the commit before the HEAD commit. 2 caret characters would mean 2 commits before HEAD, 3 carets would mean 3 and so on. I think this ends at 3 because if you want to go back 4 commits you would do this:
git reset --soft HEAD~4
The tilde symbol ~ with a number after it means go back that number of commits. There are other parameters you can pass to reset but I've never actually had to use them before. One of them is --hard which sounds pretty brutal. I have no idea what it does. When I do I'll be sure to add it to this guide.
More or less everything else you need to know about Git with ThinkUp can be found in the brilliant Getting the Source Code and Keeping it Up to Date developer's guide. it covers thinks like squashing commits (which is on the pull request checklist), rebasing (which makes life easier for just about everyone) and adding remote Git repositories.
One last thing before I move on to the next section. You might be wondering how to send your changes to Gina. The way this is done is through a process called "pulling". What this means is that once you've finished writing your code and it passes all of the pull request checklist items, you send a "pull request" to ask Gina to merge your code into the master branch. This then opens a kind of thread and alerts Gina that a new pull request has been issued, she looks at this thread (which contains a nice diff of all of the changes you've made) and she'll either make comments on how to improve it if she feels there's more to be done or she'll merge it if she thinks it's all good. But how is this done?
First, you need to push your changes to GitHub (after you've finished all of your code and rebasing etc.). Doing this is quite simple:
git push origin 338-fix-broken-link-when-install-dir-contains-space
You can replace the "338-fix-broken-link-when-install-dir-contains-space" with the name of your branch. Leaving that argument blank will push the master branch to GitHub which isn't what we want. "Origin" refers to the place that you downloaded the Git source from (you can see this by executing a git remote -v command). Then your new branch will appear in your ThinkUp fork on your profile. To navigate to it, go to your ThinkUp fork on your profile, mouse over "Switch Branches" in the top left and select the branch you just pushed from the drop down. When you've loaded the branch page, click on "Pull Request" in the top right and you'll be taken to a page where you can type up the changes you have made and anything you feel Gina should know about your new branch.
That's about all I can say about Git. This section got pretty long pretty fast but it's all stuff that's well worth knowing. If you do get stuck on anything, the mailing list or IRC channel would be more than happy to try and help you out as best they can. There are tons of guides out there for Git, too. I recommend you take a peek at one or two if you're still not quite understanding what's going on.
Phew! That was a lot of reading. If you're feeling confident and want to learn some more nitty gritty stuff about Git, there are some fantastic screencasts at GitCasts.
ThinkUp follows a "test driven development" model. It's less scary than it sounds. Essentially all that it means is for every bit of code that is written, there is a corresponding test to make sure it works. For example, my first commit on this project was modifying the installer process so you don't need a database set up before installing it. The installer would check if the database existed and create it if it didn't.
Along with this code I needed to write tests for all possible test cases: existing database names, insufficient privileges, invalid database names, invalid database log in credentials, the works. But because these tests exist, the project is far easier to maintain. Every change that is made has the potential to break another part of the code base and the unit tests (the tests are called unit tests) will highlight what part is broken.
If you've used JUnit in Java or any other unit testing package in any other language they all follow more or less the same syntax and ideology. ThinkUp already has a good guide on how to write unit tests but I'll highlight the basic idea:
You create a test case. For example, with the installer I would test making a connection to the database with false credentials (I'm fairly sure I used "localcheese" instead of "localhost" as the database server). Then I would analyse the page that returned from that to search for a specific error string. For some simple example tests, check out the TestOfConfig.php in the tests/ directory.
Yeah, it was a pretty alien concept to me when I started, too. It makes a lot of sense, though, and you'll come to love it in time. ThinkUp has a brief page on their movel-view-controller implementation which links off to the Wikipedia page on MVC which isn't very helpful to a new developer so I'll do my best to explain.
The basic principal behind MVC is separating programming logic from the design and user interface. The models handle programming logic (the PHP part) and the views handle the user interface (the HTML part). The controllers handle deciding which models to use with which views and what information should be sent to the views.
The views are the part that really interested me. They're Smarty .tpl files which are really good at combining simple PHP with HTML. Here's a really cool crash course in how to use Smarty .tpl files. (Note: In that example they use the assign() method to send data to .tpl files but as far as my ThinkUp development goes, I've only seen the addToView() method used. They seem to do the same thing.)
The purpose of the controllers is to use the models to generate data to send to the views. Every exposed page (page that you access in ThinkUp such as index.php, install.php etc.) is generated by calling the go() method of a controller. That go() method will handle all the programming logic and send a bunch of key-value pairs to a .tpl file which will be parsed, the PHP in it will be evaluated, and it will be displayed on the page. Elegant separation of all programming logic from user interface. As a rule, there is no HTML in any file apart from .tpl files (though there is the occasional exception).
Getting started writing code is difficult. You first need to read and understand the existing code surrounding the issue you want to work on. I recommend picking a fairly simple issue, something that shouldn't take too much knowledge of the code base because it's very easy to get overwhelmed by the amount of code there is.
The first issue I worked on was to do with the install process. ThinkUp used to require you to have an existing database already set up and waiting for the installer to populate it with the ThinkUp tables. If the database you typed in during the installation did not exist, the install failed. I wrote code that created the database if it did not exist. On the surface, not too much of a challenge, however, this small issue still required a lot of time and effort to code due to the fact it was my first look at the code and I had no knowledge of the code base.
Reading code is as much of a skill as writing code is. I will not lie to you, it isn't easy. Every line, every method call and every variable needs your attention and needs you to understand what it's doing. This process will take you a while and I recommend you dedicate large blocks of time to it. It's not something that can be achieved in half an hour.
Don't let me put you off. In time, you will be able to read and understand code with relative ease. Like most skills, practice makes perfect and you will be up and running before you know it.
That's totally okay. ThinkUp has both a mailing list and IRC channel for this exact scenario! Here's a link to the mailing list and here is a link to the IRC channel. You may not get a reply straight away but anyone who sees your query will endeavor to help you as much as they can!