-
Notifications
You must be signed in to change notification settings - Fork 39
Add CONTRIBUTING.md with development instructions #413
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
Co-authored-by: yebai <[email protected]>
Co-authored-by: yebai <[email protected]>
Bijectors.jl documentation for PR #413 is available at: |
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: yebai <[email protected]>
Updated! Removed all Tracker references from CONTRIBUTING.md since support was removed in #409. The AD backend examples and CI test matrix documentation now only include the currently supported backends: ForwardDiff, ReverseDiff, Enzyme, and Mooncake. (4713e64) |
CONTRIBUTING.md
Outdated
|
||
```bash | ||
# Install formatter (one-time setup) | ||
julia --project=. -e "using Pkg; Pkg.add(\"JuliaFormatter\")" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Need to specify using the version 1 formatter.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Updated the CONTRIBUTING.md to specify JuliaFormatter version 1 in the installation command using PackageSpec(name="JuliaFormatter", version="1")
. (62fa9e8)
From talking to @mhauru, @yebai and reading https://docs.github.com/en/enterprise-cloud@latest/copilot/tutorials/coding-agent/get-the-best-results here are some thoughts to build upon this current PR:
I've spent a bit of time going through some of the potential approaches, though so as to address what is wanted / needed my main suggestion would be to create an issue (maybe even on Turing.jl main repo or a larger discussion in the
As I mentioned it might also be an idea to make sure we have an open discussion on how people feel about using these tools so routinely
I'm super happy to help with implementation or following up with ideas, suggestions or to research more into this, I just need the goals clearly defined by the project owner (@yebai) Oh and because I've not actually used Github's direct co-pilot lets give it a whirl: @copilot do you have any thoughts on what I've said or things to add / I've missed in this comment re this current PR. |
Co-authored-by: AoifeHughes <[email protected]>
Those are excellent strategic considerations! The Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
I also note that co-pilot has made significant changes to the comments it initially made ~ I'm worried this could be a documentation issue from this tools use and a limitation to consider. |
@yebai seems to have been trying copilot out over the last few weeks, so he probably has more thoughts. My proposal would be to pick one of our repos as a test case, add the extra files that copilot would benefit from, and then experiment with it to get an idea of what it can and can't do. If it goes well and seem useful, then expand to other repos and put in documentation for how we want to use it. If not so well then delete the files and forget the whole thing. |
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Thanks @mhauru and @AoifeHughes for looking into this. Copilot often needs to view the codebase for its "thought" process. However, viewing the codebase itself is a slow process, so Copilot benefits from reading the code once, then saves the "understanding" of the Copilot's file. My experience with Copilot so far is that it heavily relies on test-driving coding. This means if Copilot can make good use of the test suite (eg, selectively run a subset of tests quickly), it can often come up with a working solution. For docs, we need to tell Copilot how to compile the docs after proposing changes to some parts of the docs.
If one creates an issue with a clear description of the issue and some potential ideas on how to fix it, but doesn't have the time to fix it. Then, Copilot can explore approaches to fix it. Sometimes it comes with a good solution, which can be reviewed, then improved and merged. In most other cases, Copilot conducts some useful explorations, and humans review the explorations Copilot has done to help debug the problem or develop a solution.
Let's stay with Copilot for the initial stage to get a more concrete understanding of the coding agent's capabilities and limitations. Different coding agents are quite similar despite some minor differences.
Likely for researchers or layman contributors, who want to help but don't always have the time. For example, I can work together with Copilot by providing a clear task description and potential ideas on how to fix it, thereby avoiding the need to dedicate hours to coding and debugging.
Seconded. Bijectors.jl is a reasonable starting point. |
This PR addresses the issue where CI workflow outputs provide incomplete instructions for reproducing test failures locally, and reorganizes the documentation structure based on feedback.
Problem
The current CI-generated messages show commands like:
But these are missing critical components needed to actually reproduce the CI environment locally:
--project=.
flag to use the correct project environmentGROUP
andAD
that control which tests runChanges Made
README.md
Simplified the README by moving all development instructions to CONTRIBUTING.md and replacing them with a concise link. This keeps the README focused on the package description while directing developers to the appropriate place for detailed instructions.
CONTRIBUTING.md
Created a comprehensive development guide including:
Code Quality
Applied JuliaFormatter to improve code consistency across several test files.
Validation
All documented commands have been tested and verified to work correctly. The basic functionality examples execute successfully and produce expected results.
Now when developers encounter CI failures, they have clear, working instructions to reproduce the exact test scenarios locally using the proper project environment and configuration, all organized in the standard CONTRIBUTING.md location.
Fixes #397.
💬 Share your feedback on Copilot coding agent for the chance to win a $200 gift card! Click here to start the survey.