These are some are thoughts to keep in mind when we're contributing to a software project.
Git is maybe the most extended tool to manage version control in software and as a daily tool we should invest some time in knowing how to use it "properly".
This helps to identify who make a change while we're browsing git log history. Don't use initials or nicknames, in a few months or maybe years eventually you will be looking at the git log and if you see something like this
Author: XYZ <email@example.com> you won't have a clue who make that change. We can setup this information with:
git config --global user.name "name and last name" git config --global user.email "work-or-personal-email"
We can use it without
--global to make these changes only in the current repository.
A easy to remember "rule" could be "Put what you did in the title and why you did it in the body". Memory is fragile and it could sound useless now but in a few months when you'll be looking at the log and see a commit and wanted to know why some change was made and the commit message just says "Fix some error" you probably will get angry with your past self.
There is an article that explains this so much better than me, you could read it at https://chris.beams.io/posts/git-commit/
An easy "hack" to force yourself to follow this is create a "commit template", it will be used when you make a commit and it will remind you about this rule. An example one could be:
# If applied, this commit will... # Explain why this change is being made # Provide links to any relevant tickets, articles or other resources
These lines are comments and won't be added to the final commit message.
Put this in
~/.git_commit_message.txt and then configure you
~/.gitconfig like this:
[commit] template = ~/.git_commit_message.txt
After configuring this every time you make a commit the content of
~/.git_commit_message.txt will be filled into the commit message so that will remind you how to write a proper message. The editor/IDE you use to write a message will recognize this file.
Friends don't let friend use
git add .
The tool you use to stage changes and then commit should allow you to stage chunks of code one by one. This way you can review what are you adding. You don't want to commit those debug messages or commented code that you were using while your working on the project. Or even worse you definitively don't want to commit credentials files.
Also if you like to use
git from the terminal, you can use
git add -p to use an interactive mode, this will go through all the unstaged changes and will ask you one by one if you want to add them.
Most projects already have a
CONTRIBUTING.md file. This file has some guidelines and conventions used in the project. Take your time to read and understand it. Those guidelines exist to keep consistency across the project. They're not fixed rules and of course they could be improved but always discuss that with your team.
OK you're done with a feature/fix/whatever your were working on. It's time to submit a pull request. This is a simple checklist that could be followed:
Make sure your changes are in sync with the latest changes in
upstream. Maybe you're been working on something for a few days but some new changes were made to the project since the last sync.
Make sure you ran the whole test suite, run checkers and others pre-commit workflow that your project have.
Check the commits you've made. Maybe you can clean a little bit the log, remove some
WIPcommits or improving some messages.
After you submit the pull request be your first reviewer before asking for more reviews. Once the pull request is created go to the pull request page on Github/Gitlab/etc and make a last check of your changes. Then assign the reviewers.
Use tags, this way you are adding extra information that will help to understand and categorize the pull request, by looking at the tags we can know if the pull request is a fix, improvement, documentation, etc.
Link your pull request with an issue. Github|Gitlab|etc have a nice feature to link pull requests with issues, for example if you're fixing a reported issue with number
fixes #10 in the pull request description and this will link it to that issue, this way we can have traceability about the work done and the issues reported.
Improve pull request description, when you open a pull request Github|Gitlab|etc takes the first commit title and message and use them as title and description for the pull request so if we have made good commit messages most of the work is already done. Make review process a little easier by explaining what have you done. Maybe link another issues in different projects or bug reports related the changes you have made. Some extra context is always welcome :)
Keep in mind that comments made in your pull request are made about the code and not about you.
Don't take a comment as a fact that should be followed immediately, reviewers can make mistakes too, if you think some suggestion doesn't apply, explain your point of view and try to understand what is the reason behind that suggestion.
Don't make changes just to make reviewers happy and get your pull request merged. A code review is an excellent opportunity to get more knowledge about the project and to learn more about the code review process itself.
If someone rewrite or delete some code that you wrote it isn't mean necessary is was "bad code" or something like that, maybe the requirement was different back then. Code has to evolve and at the end what matters the most is that the application does what is supposed to do. Also in some cases you'll be the one who is refactoring or fixing your own code after a while it was made ;)