Ycechung Open Source Blog: November 2021

Friday 26 November 2021

Introduction

This week we are looking at software releases and getting our code into the hands of users. We are working with git tags, package managers and user testing.

Step 1:

I am using Node.js so I picked NPM as my release method as suggested.

Step 2: Implementation

The only tricky part of this lab is needing the --access public in order to publish the package publicly. This is the only part that I got stuck and needed to search on Google.

Other things such as having the correct version number for each iteration of release in package.json and making sure the package installs locally and works as intended.

Step 3: Tagging Releases

This is relatively straight forward. I created tags on main and used git to push --follow-tags.

Step 4: Do A Release

My release is located at https://www.npmjs.com/package/@ycechung/cmd-ssg, and I finalized on version 1.0.2.

Step 5/6: Update Documentation

I made sure my documentation is updated and pushed on Github. In addition, I made sure others know about my release and tested installing my project using npx.

It is near the end. Labs are done.

Thanks for Reading,

Eugene Chung

Friday 19 November 2021

Introduction

After doing the Docusaurus Lab, I was keen on doing an issue related to it. At the time, issue-10 was one of the few issues in IPC144 website. I have mixed feelings after seeing how complicated this implementation could get, but I am glad I have taken on this challenge. This is definitely one of the hardest issues I have worked on for Github to date, because the documentation on Facebook Algolia website is not very clear on the entire process of implementating search and that it is outdated if you used their service. Therefore I found out that I have to build my own index using a custom scrapper application. I broken the issue down into just building the index for my pull request and branching to two more issues, #69 and #70 for the future.

Steps to Creating Index

Based on the CONTRIBUTION.md file I modified, I added the following

1. sign up for an account on https://www.algolia.com/users/sign_in

2. create a project for the purpose of creating the index

3. add an 'INDEX_NAME' called 'docusaurus'

4. Add API_KEY and give rights (`search, addObject, deleteObject, editSearch)` for the key

5. SITE_URL is the deploy site, but siteindex needs to match it in order to crawl it

Implementation

The Algolia scrapper is run on the Docker platform on Ubuntu linux and I made a bash script for it called scrape.sh, in which the Github Actions CI is run to call it.

The bash script is as follows:

Just running the script locally gives the following result.

The Experience

The practice of working with Git on this repository is a good experience especially reviewing squash and re-basing over twenty commits is a unique experience. The order to get things merged and reviewing other's pull request is also interesting.