It's good to hear that you want to contribute to Pony!
There are a number of ways to contribute to Pony. As this document is a little long, feel free to jump to the section that applies to you currently:
Additional notes regarding formatting:
For any feature requests or enhancements to the Pony distribution, it is quite likely that you have to go through our RFC process. Before opening or submitting any feature requests, please make sure you are familiar with the RFC process and follow the process as required.
If you submit a pull request to implement a new feature without going through the RFC process, it may be closed with a polite request to submit an RFC first.
First of all please search existing issues to make sure your issue hasn't already been reported. If you cannot find a suitable issue — create a new one.
Provide the following details:
- short summary of what you was trying to achieve,
- a code causing the bug,
- expected result,
- actual results and
- environment details: at least operating system and compiler version (
ponyc -v
).
If possible, try to isolate the problem and provide just enough code to demonstrate it. Add any related information which might help to fix the issue.
We use a fairly standard GitHub pull request workflow. If you have already contributed to a project via GitHub pull request, you can skip this section and proceed to the specific details of what we ask for in a pull request. If this is your first time contributing to a project via GitHub, read on.
Here is the basic GitHub workflow:
- Fork the ponyc repo. you can do this via the GitHub website. This will result in you having your own copy of the ponyc repo under your GitHub account.
- Clone your ponyc repo to your local machine
- Make a branch for your change
- Make your change on that branch
- Push your change to your repo
- Use the github ui to open a PR
Some things to note that aren't immediately obvious to folks just starting out:
- Your fork doesn't automatically stay up to date with change in the main repo.
- Any changes you make on your branch that you used for the PR will automatically appear in the PR so if you have more than 1 PR, be sure to always create different branches for them.
- Weird things happen with commit history if you don't create your PR branches off of main so always make sure you have the main branch checked out before creating a branch for a PR
If you feel overwhelmed at any point, don't worry, it can be a lot to learn when you get started. Feel free to reach out via Zulip.
You can get help using GitHub via the official documentation. Some highlights include:
Before issuing a pull request we ask that you squash all your commits into a single logical commit. While your PR is in review, we may ask for additional changes, please do not squash those commits while the review is underway. Once everything is good, we'll then ask you to further squash those commits before merging. We ask that you not squash while a review is underway as it can make it hard to follow what is going on. Additionally, we ask that you:
- Write a good commit message
- Issue 1 Pull Request per feature. Don't lump unrelated changes together.
If you aren't sure how to squash multiple commits into one, Steve Klabnik wrote a handy guide that you can refer to.
We keep a CHANGELOG of all software changes with behavioural effects in ponyc. If your PR includes such changes (rather than say a documentation update), a Pony team member will do the following before merging it, so that the PR will be automatically added to the CHANGELOG:
- Ensure that the ticket is tagged with one or more appropriate "changelog - *" labels - each label corresponds to a section of the changelog where this change will be automatically mentioned.
- Ensure that the ticket title is appropriate - the title will be used as the summary of the change, so it should be appropriately formatted, including a ticket reference if the PR is a fix to an existing bug ticket.
- For example, an appropriate title for a PR that fixes a bug reported in issue ticket #98 might look like:
- Fixed compiler crash related to tuple recovery (issue #98)
Once those conditions are met, the PR can be merged, and an automated system will immediately add the entry to the changelog. Keeping the changelog entries out of the file changes in the PR helps to avoid conflicts and other administrative headaches when many PRs are in progress.
Please note, if your changes are purely to things like README, CHANGELOG etc, you can add [skip ci] as the last line of your commit message and your PR won't be run through our continuous integration systems. We ask that you use [skip ci] where appropriate as it helps to get changes through CI faster and doesn't waste resources that Appveyor and TravisCI are kindly donating to the Open Source community.
If your contribution contains changes to the grammar of the language, you should update the pony.g
file at the root of the repository with ponyc --antlr > pony.g
.
When contributing to documentation, try to keep the following style guidelines in mind:
- Wherever possible, try to match the style of surrounding documentation.
- Avoid hard-wrapping lines within paragraphs (using line breaks in the middle of or between sentences to make lines shorter than a certain length). Instead, turn on soft-wrapping in your editor and expect the documentation renderer to let the text flow to the width of the container.
- Apply code highlighting to identifier names, such as types, fields, variables and parameters via the customary markdown syntax of wrapping the name in backticks.
For code formatting guidelines please see The Style Guide.
For standard library file naming guidelines see The Style Guide.
To get C code coverage information for test runs or for calling ponyc, call make
with use=coverage config=debug
. This works both for clang and gcc. Make sure to configure CC
and CXX
environment variables both to either gcc
and g++
or clang
and clang++
.
-
Compile ponyc with
use=coverage config=debug
and environment variablesCC=gcc CXX=g++
-
Run ponyc or the test suite from
build/debug-coverage
-
generate the html coverage report:
# generate coverage report lcov --directory .build/debug-coverage/obj –zerocounters lcov --directory .build/debug-coverage/obj --capture --output-file ponyc.info genhtml -o build/debug-coverage/coverage ponyc.info
-
open the html report at
build/debug-coverage/coverage/index.html
-
Compile ponyc with
use=coverage config=debug
and environment variablesCC=clang CXX=clang++
-
Run ponyc or the test suite from
build/debug-coverage
with environment variable:LLVM_PROFILE_FILE="build/debug-coverage/coverage.profraw"
-
generate coverage data:
llvm-profdata merge -sparse -output=build/debug-coverage/coverage.profdata build/debug-coverage/coverage.profraw
-
show coverage data (only for
lexer.c
in this case):llvm-cov show ./build/debug-coverage/libponyc.tests -instr-profile=./build/debug-coverage/coverage.profdata src/libponyc/ast/lexer.c