Get Started with a Few Simple Pull Requests
Making a Pull Request doesn't have to be a complex or stressful process.
I have been posting my own code on GitHub for several years. Until recently I had never submitted a Pull Request (PR). There was a PR submitted to one of my repos that I reviewed but I never made an original contribution to another developer's repo.
One week over the Summer I was moving apartments and in my down-time submitted a PR at a rate of about one per day. PRs were soon demystified. It is easier than you might think and well worth getting started if you haven't already.
You may have been directed to submit "simple" PRs before but might have had difficulty finding examples. There are helpful repos that allow beginners to learn the mechanics of how to submit a PR but not necessarily what a real PR might contain. Here I give an overview of my PRs so far in the hope that they might be demystified for you too.
Where to Start#
You probably want to start with a project that interests you. The best way to find such a project is to use and enjoy open-source software. If you find an issue in software that you use then try to determine how you might fix it.
It is often recommended that beginners look into improving documentation. I think that this is good advice. Documentation can often be an afterthought in open-source projects. Highly-skilled developers might have limited interest, time or ability to devote to good documentation.
Most — though not all — of the examples in this post involve docs.
Many developers are not necessarily strong writers, but many users also do not read enough documentation! To improve docs you should first get in the habit of reading docs. It also helps if you have a pedantic nature, and here my academic background served me well.
At time of writing I have submitted six PRs in about three months. Even though I am a fan of the command line, and manage all my own repos from the terminal, these PRs were all generated directly through GitHub's web interface.
Doc and code files in GitHub have an
edit button. Clicking this will trigger a prompt to fork the repo. After you fork the repo you will be able to make edits to the code and commit the changes to your fork. A
Compare & Pull Request button will then appear — click it and you can give your PR a title and description before submitting.
For the small changes described below it is likely that the reviewer of your PR will merge the changes after a quick review. Anything more substantial may require further dialogue, additional justification and possibly some edits to the original PR, all of which is outside the scope of this post.
Lately I have been using DOOM Emacs for taking notes with the wonderful org-mode. I use a lovely theme (naysayer) that had a small issue with the colour of text being difficult to read over a particular window region in Emacs.
To attempt a fix I changed two lines in the theme definition, tested the changes on two devices (Doom and Vanilla Emacs) and submitted a PR. Here was the explanation I wrote on submission; it might seem like overkill, but if there were to be any problem subsequently then this might provide helpful context:
These changes aim to improve text legibility in the modeline for both the project directory and the file path in a modified buffer when using Naysayer with DOOM Emacs (a very popular Emacs framework). Previously the project directory colour defaulted to a blueish-green that was difficult to read against Naysayer's tan-coloured modeline; in addition, the file path in a modified buffer changed to a light pink that was particularly illegible against the tan colour.
The text colour will now be the same as the rest of the foreground text in the modeline, which is the theme's behavior in Vanilla Emacs, and the text weighting will change in a modified buffer to subtly signify any unsaved changes.
The revised theme was tested on both Doom Emacs (28.1) and Vanilla Emacs (28.1) - no issues were observed in either case. Some of these problems were discussed in issue #5.
This one is going to seem annoying and pedantic but I think it was worthwhile. If even one reader is anything like me they might be thrown by inconsistent phrasing. I was reading about the child-parent relationship for elements in the excellent Motion Canvas animation library. When reading the docs I became distracted by the phrase "relatively to". This seemed a little awkward and far less common than "relative to". The more common phrase was also the one used elsewhere in the docs.
For the sake of consistency, and to prevent the likes of me from getting distracted when learning programmatic animation, I recommended a change in phrasing:
This is a small change to phrasing in the docs that substitutes the phrase "relatively to" with "relative to".
The latter is more conventional/idiomatic and is also consistent with similar usage contexts in the rest of the docs.
The purpose of the change is to minimise any ambiguity for the user when interpreting the technical content.
This PR is the most simple of all! It's often easier to spot typos when you haven't been working closely on the code or documentation for a project. Some people think typos signal a lack of professionalism, which I personally think is unfair. Nonetheless, typos bother some people and it's wortwhile fixing them. If you have a GitHub account and some time on your hands there's really no excuse for not lending a hand in these cases. I spotted this typo for the Manim Slides project, a really convenient application for making slides in the style of 3Blue1Brown's explanatory mathematics videos.
Change link titled "HTML Presenetation" (typo) to "HTML Presentation".
Description: Proposes a fix to a small typo in the docs.
This next PR attempted to reduce ambiguity due to a lack of explicitness. I was using the excellent Quarto publishing system to create material for a university class. In a set of notes I had included code for a Mermaid flow diagram but now I wanted to use the same diagram in a slidedeck. Quarto allows Mermaid diagrams to be defined in a source file that can then be referenced in other files, preventing duplication. This would allow me to code the diagram once and then refer to it as many times as I want.
I could not tell from the relevant section of the Diagram docs (
File Include) what the filetype was supposed to be when including Mermaid diagrams in a different file. I tried using a
.mmd filetype, which is mentioned elsewhere in the Quarto docs, and it worked. In my view, this should be made clear and obvious in the section of the docs that describes how to include externally-defined Mermaid diagrams:
Makes more explicit that an .mmd file can be used as an external source when including a Mermaid diagram in a .qmd document, as this is made more explicit for .dot files in the current version of the docs.
Mermaid and Graphviz diagrams can be included in .qmd documents from external files.
In the docs an example is given for how to achieve this using a Graphviz diagram defined in a .dot file.
This minor change makes it more explicit that an appropriate filetype that would achieve the same for a Mermaid diagram is .mmd.
The .mmd filetype is mentioned two sections previous in the context of specific editors (RStudio, VS Code). Readers of the docs are likely to benefit from seeing the relevant filetypes close to their implementation in the File Include section of the docs, especially if they are not users of RStudio or VS Code and may miss those previous sections.
My Website's Theme#
This blog is built using the Zola static site generator and a modification of the lovely Serene theme. While working on the website I was following instructions from Serene's repo and noticed that there was an error in the file paths that were cited. My PR description:
USAGE.mdinstructions are provided to copy the theme's
The instructions indicate that
/highlight_themes/is found in
myblog/themes/serene/static/but it is found one directory up in the theme root.
USAGE.mdso that the instructions are to copy
myblog/highlight_themes, which is presumably the intention.
Many of these PRs have involved documentation and this last one relates to documentation in the form of in-line code comments. A new programming language that I really like is Odin. I like it so much I made a Pong clone with it. There's a lot of really smart people working on Odin. So smart in fact that I would hesitate to mess with their pristine code until I understand the language better.
However, I did notice that there was some inconsistency in the commenting style in the Overview of the language. This Overview is the first encounter many people will have with the language, so it's better not to introduce any subtle confusion in the explanatory comments.
When writing in markdown, in-line code blocks are conventionally signified with backticks, but here a mixture of backticks and quotes was being used. Quotes are more appropriate when signifiying strings, which was how they were used elsewhere in the comments. In addition, I spotted some other issues that were too small to include in a separate PR. The PR description:
Code comments used single quotes, double quotes and back-ticks interchangeably to signify an in-line code block. Periods were infrequently used to terminate a comment.
These changes use back-ticks for all in-line code blocks and remove all terminating periods to improve consistency of the commenting style.
In addition, a few phrases were given slight rewrites to increase clarity or remove typos.
Most of my experience with open-source software has involved making my own projects and using software made by others. After getting a few PRs merged I feel like more of a well-rounded contributor to the open-source ecosystem. It is easy to get started if you don't set the bar too high.
Use some software with a narrow scope (a theme) and note down any issues you have. Pay attention to anything in software documentation that trips up or confuses you. Look at the closed PRs in a repo and try to find small ones that you can understand.
Helping people out with these tasks can be very satisfying and could help prepare you for bigger contributions in the future.