Welcome! This guide serves as the guideline to contributing to the µRaiden Network codebase. It's here to help you understand what development practises we use here and what are the requirements for a Pull Request to be opened against µRaiden.
There are two ways you can contribute to the development. You can either open an Issue or if you have programming abilities open a Pull Request.
If you experience a problem while using µRaiden or want to request a feature then you should open an issue against the repository. All issues should contain:
For Feature Requests:
- A description of what you would like to see implemented
- An explanation of why you believe this would make a good addition to µRaiden
For Bugs:
- A short description of the problem
- Detailed description of your system, µRaiden version, geth version, solidity version e.t.c.
- What was the exact unexpected thing that occurred
- What you were expecting to happen instead
If you have some coding abilities and would like to contribute to the actual codebase of µRaiden then you can open a Pull Request(PR) against the repository.
All PRs should be:
- Self-contained
- As short as possible and address a single issue or even a part of an issue. Consider breaking long PRs into smaller ones.
In order for a Pull Request to get merged into the main repository you should have one approved review from one of the core developers of µRaiden and also all Continuous Integration tests should be passing and the CI build should be green.
Please refer to README.md for installation instructions.
Run the tests using
env/bin/pytest microraiden
Tests, especially integration tests, will take some time. If you want to run single tests simply specify them on the command line, like so:
env/bin/pytest microraiden/tests/<path-to-test(s)>
In this section we are going to describe the coding rules for contributing to the µRaiden repository. All code you write should strive to comply with these rules.
For an exhaustive guide read this guide. It's all really good advice. Some rules that you should always follow though are:
- A commit title not exceeding 50 characters
- A blank line after the title (optional if there is no description)
- A description of what the commit did (optional if the commit is really small)
Why are these rules important? All tools that consume git repos and show you information treat the first 80 characters as a title. Even Github itself does this. And the git history looks really nice and neat if these simple rules are followed.
µRaiden is written in Python and we follow the official Python style guide PEP8. It is highly recommended to use the flake8 tool in order to automatically determine any and all style violations. The customizable part of flake can be seen in the configuration file. For all the rest which are not configurable here is some general guidelines.
Line Length Flake8 will warn you for 99 characters which is the hard limit on the max length. Try not to go above it. We also have a soft limit on 80 characters but that is not enforced and is there just to encourage short lines.
Breaking function definitions when line is above 99 characters
Always put each argument into its own line. Look at the following examples to understand:
The following should be avoided
def function_with_many_args(argument1, argument2, argument3, argument4, argument5, argument6, argument7):
passand instead you should
def function_with_many_args(
argument1,
argument2,
argument3,
argument4,
argument5,
argument6,
argument7):
passBreaking function calls when line is above 99 characters
Much like in the above example the following should be avoided
function_call_with_many_arguments(argument1, argument2, argument3, argument4, argument5, argument6, argument7)and instead you should
function_call_with_many_arguments(
argument1,
argument2,
argument3,
argument4,
argument5,
argument6,
argument7
)Difference being that you can place the closing parentheses in the next line.
Usage of single and double quotes
All strings must use single quotes by default.
Bad:
s = "foo"Good:
s = 'foo'The only reason to use double quotes is to avoid escaping the single quote in a string. So this is okay:
s = "Augusto's computer is awesome"Naming Style
Use Snake Case for variable and function names and Camel Case for class names.
Naming Convention
Use descriptive variable names and avoid short abbreviations.
The following is bad:
mgr = Manager()
a = AccountBalanceHolder()
s = MicroRaidenService()While this is good:
manager = Manager()
balance_holder = AccountBalanceHolder()
service = MicroRaidenService()We try to follow a consistent naming convention throughout the codebase to make it easy for the reader of the code to understand what is going on. Thus we introduce the following rules:
For addresses:
<name>_address_hexfor hex encoded addresses<name>_addressfor binary encoded addresses
Lists of objects:
<name>s, e.g.channelsfor a listChannelobject instances.
Mappings/dicts:
If it is a simple one to one mapping
<name>_to_<name>, e.g. tokenaddress_to_taskmanager
If the mapped to object is a list then add an s
<name>_to_<name>s, e.g. tokenaddress_to_taskmanagers = defaultdict(list())
For solidity we generally follow the style guide as shown in the solidity documentation with a few notable exceptions:
Variable Names
All variable name should be in snake case, just like in python. Function names on the other hand should be mixedCase. MixedCase is essentially like CamelCase but with the initial letter being a small letter. This helps us to easily determine which function calls are smart contract calls in the python code side.
function iDoSomething(uint awesome_argument) {
doSomethingElse();
}Modifiers in long function declarations
This is how the solidity documentation suggests it:
function thisFunctionNameIsReallyLong(
address x,
address y,
address z,
)
public
onlyowner
priced
returns (address)
{
doSomething();
}This is the minor modification we make in order to make the code more readable when quickly skimming through it. The thinking here is to easily spot the start of the function's block when skimming and not have the modifiers appearing as if they are a block on their own due to the hanging parentheses.
function thisFunctionNameIsReallyLong(
address x,
address y,
address z)
public
onlyowner
priced
returns (address)
{
doSomething();
}When developing a feature, or a bug fix you should always start by writting a test for it, or by modifying existing tests to test for your feature. Once you see that test failing you should implement the feature and confirm that all your new tests pass.
Afterwards you should open a Pull Request from your fork or feature branch against master. You will be given feedback from the core developers of µRaiden and you should try to incorporate that feedback into your branch. Once you do so and all tests pass your feature/fix will be merged.
If you are a core developer of µRaiden with write privileges to the repository then you can add commits or rebase to master any Pull Request by other people.
Let us take this PR as an example. The contributor has everything ready and all is looking good apart from a minor glitch. You can wait until he fixes it himself but you can always help him by contributing to his branch's PR:
git remote add pcppcp [email protected]:pcppcp/microraiden.git
git fetch pcppcp
git checkout travis_ci
Right now you are working on the contributor's Pull Request. Make sure to coordinate to avoid any conflicts and always warn people beforehand if you are to work on their branch. Once you are done:
git commit -m "your additions"
git push pcppcp travis_ci
Congratulations, you have added to someone else's PR!