If you remember my story on my first two contributions to open source, I had begun implementing some enhancements in Crawler-Commons that would support an enhancement in the StormCrawler project. My changes to Crawler-Commons were accepted, merged, and a new version of the library, 1.1, was published with my enhancements (amongst many things).

Now that the library was published, the dependency on it, in StormCrawler, was updated to use the latest release and I am now able to finish the enhancement in StormCrawler. Here’s how it went:

  • I started on the implementation with a few notes in my journal (yes, I sometimes journal about code):
Some implementation notes
Some more implementation notes

  • A few days after that, I wrote the code that implements this feature

  • For the few remaining days before submitting a pull request, I focused on writing unit tests. I ended up writing more unit test code than implementation code, by a large margin (42 lines of implementation code, 220 lines of unit testing code).

  • Interestingly enough, while writing unit tests for this contribution, I discovered a really minor oversight in the code I submitted to Crawler-Commons. I explained my discovery in my pull request description:

Somewhat related, while implementing this, I noticed that crawler-commons isn't adding the duration attribute from the VideoExtension to the map generated by VideoExtension.asMap() (oversight on my part). I wrote this up as an issue, crawler-commons/300 and submitted a PR. A future release will include duration, however, it will be missing from StormCrawler until a new version of crawler-commons is published with the fix.

I opened an issue for this oversight in Crawler-Commons and submitted a pull request with the fix.

  • Next, I began the process of submitting a pull request. The pull request guidelines for this project were specified in the Pull Request Template:
Thanks for contributing to StormCrawler, your efforts are appreciated!

Before opening a PR, please check that: 

* You've applied the formatting used by the project with `mvn java-formatter:format`
* You've squashed your commits into a single one
* You've described what the PR does or at least point to a related issue

Thanks!

It was super straightfoorward, except I’ve never squashed multiple commits down into one, manually with Git. I thought to myself, “You gon’ learn today”. Anyway, a Google search, led to a StackOverflow post, which led to an article with two ways to accomplish commit squashing. I chose the following route (the exact Git I wrote to squash!):

git checkout master
git fetch
git pull
git merge sc-749-extension-parsing 
git reset origin/master
git add --all
git commit
git push

I, effectively, pulled all of my changes onto the latest code in the master branch and committed it as one commit, instead of several. The other documented method for doing squashes is to use git rebase. I started down this road first, but it felt a bit like defusing a bomb, to be honest.

  • I submitted the pull request. It was approved and merged about a half day later and I got some great feedback.
{< tweet 1283874215146520576 >}

Update: May the above tweet live on in the twitter afterlife (the original tweet was deleted)

I won’t be writing about every open source contribution, but I’ll be continuing my contributions.

🧇