Skip to content

Latest commit

 

History

History
363 lines (307 loc) · 16.3 KB

CONTRIBUTING.md

File metadata and controls

363 lines (307 loc) · 16.3 KB

Contributing to Nektar++

Contents

This is a reasonably complete guide to help if you're interested in contributing to Nektar++, either in reporting bugs or, hopefully, trying to fix them! It's split up into a number of sections:

Issues and bug reports

Think you've found a bug or issue with Nektar++? We're very keen to hear about it!

  • In the first instance, you should raise an issue on the issue tracker -- be sure to do a quick search and see if anyone has reported the same thing first.
  • Alternatively you can join the mailing list for more advice.

It's really helpful if you can include a small session file that reproduces the error, and can give a good description of the problem you're having.

How to contribute

If you've got a patch or feature, please consider contributing it back to the project. It's a pretty simple process:

  1. Fork the Nektar++ repository in nektar/nektar into your username's space.
  2. Create a branch with the naming convention:
    • feature/myawesomebranch: a new feature that wasn't in Nektar++ already.
    • fix/mygreatfix: fixes an issue that isn't tracked in the issue tracker.
    • ticket/123-myfantasticpatch: fixes an issue that is tracked in the issue tracker (please include the issue number somewhere!)
    • tidy/mybrillianttidying: cosmetic fixes to bring existing files up to the Nektar++ code guidelines.
  3. Make sure you've gone through the checklist below.
  4. Submit a request to merge into master (Merge Request). If you just want to see the diff and are not quite ready to merge, use the Draft tag in the title, for instance Draft: Update session reader. This will prevent your code from being accidentally merged.
  5. Respond to any comments in the code review.

Reviewing Merge Requests can be a arduous and time-consuming process for reviewers. To allow for efficient code reviews, we ask that you take particular care of the following points:

  • Submit small Merge Requests. Whenever you can, split your work into multiple self-contained, stacked Merge Requests. Say for instance that you split your work into two branches MR1, and MR2, that branches off MR1. Stacked Merge Requests would look like this:

    master <- MR1 <- MR2
    

    In the above scenario, changes in MR1 are first reviewed, before changes in MR2. After code review, MR2 is first merged into MR1's branch, which is then merged into master.

  • Provide a detailed description of your changes by filling the Merge Request template. Please keep the structure of the template as-is: fill the section with "Non applicable" if needed. Don't be afraid of "stating the obvious": the more information you provide, the easier (and quicker) the code review is.

  • Give your submission a descriptive title. Avoid generic titles like "Update module X" or "Fix bug in module Y".

An example of a Merge Request is available here.

Keeping your branch up to date

Regularly updating your branch with the latest changes on the base branch (usually master) is essential to avoid large conflicts when merging your work into the main development line. There are two ways of updating your branch:

  1. If you haven't pushed your branch to the upstream remote (nektar/nektar.git), or if you are sure nobody based work on your changes, rebase your changes on the latest master
    git checkout feature/mybranch
    git pull --rebase master

This will rewrite the history of your local branch in order for your changes to appear on top of the latest changes in master. This leads to a clean, linear history. See Git Branching - Rebasing 2. If you pushed your changes and there is a high chance that somebody based work on top of your -- yet unmerged -- changes, don't rebase. Merge the latest master state into your branch instead:

git checkout feature/mybranch
git fetch origin
git merge origin/master

This will create a merge commit joining your branch's history and the master branch's history. The merge approach do not rewrite your branch's history, at the expanse of making it more complex.

Provided that you merge your work as soon as possible, i.e. contribute small Merge Requests, most of the developement can happen in a local branch that you would rebase on top of master. This is the best case scenario. Merging becomes necessary if a branch is published and remains unmerged for more than a couple of days.

Submission checklist

  • Did you add regression tests (for fixes) or unit tests and/or normal tests for new features?
  • Have you run your branch through GitLab CI and do all the tests pass?
  • Have you fixed any new compiler warnings your code has introduced into the compilation step for all of the Linux CI environments?
    • unused parameters: if these are genuinely needed (e.g. virtual functions in a derived class, please use boost::ignore_unused() to mark as such.
    • switch case may fall-through: for switch statements which intentionally exploit fall-through between cases, mark the end of such cases with the comment /* Falls through. */ to suppress the warning.
    • Avoid ASSERTL0(false, msg); instead use NEKERROR(ErrorUtil:efatal, msg).
    • Ensure variables are initialised with sensible default values.
  • Is there documentation in the user guide and/or developer guide?
  • Have you added a CHANGELOG entry, including the MR number?
  • Are there any massive files you might have added in the commit history? We try to keep test files as small as possible. If so you'll need to rebase or filter-branch to remove those from the commit history.
  • Is the code formatted correctly?

Git cheatsheet

Although Gitlab gives a nice interface to view the diff between a branch and master, for large merges, it can be slow. The following git aliases can provide a quicker alternative. You can use these by inserting them into the .gitconfig file in your home directory, or inside the nektar++/.git/config file.

[alias]
branch-name = "!git rev-parse --abbrev-ref HEAD"
diff-nows = diff --color -w
log-branch = log --pretty='%C(green)%h %C(red)%an %C(reset)(%C(blue)%ad%C(reset))%n%s' master..
diff-branch = diff -U5 --minimal --color -w master...

This gives you four commands:

  • git branch-name displays the current branch name
  • git diff-nows shows a diff of your current commit in colour, without whitespace changes.
  • git log-branch shows a minimised log of all the commits on the current branch that are not in master.
  • git diff-branch shows a diff of the current branch against master, without showing changes from master that aren't present in the branch (i.e. git diff master...branch), without whitespace changes. (This should be roughly equivalent to Gitlab's diff).

If you prefer a graphical interface to see the files that have changed in your commit, you can additionally use the git gui command to bring up a simple interface. git difftool can also be used in combination with a GUI diff viewer, to graphically view the output of git diff.

Testing and GitLab CI

Your new features or fixes should include tests that cover the code you've added. There are numerous examples within the various Tests directory lying within the source trees, and there is an example of writing .tst files for our Tester executable in the tests/Examples directory. Once you've written your tests, add them to the CMakeLists.txt file for the relevant solver, or to the appropriate demos directory for library features in whatever directory you are working in.

You should also test your branch on the Nektar++ GitLab CI, which will compile and test the code against a number of Linux, Mac and Windows operating systems. If your tests don't pass, we can't merge the code into master.

When you submit a merge request testing on GitLab CI will happen automatically, unless you have marked the merge request as a work-in-progress (WIP: prefix). Each time you push commits to a non-WIP merge request branch, it will also trigger a build.

Documentation

Nektar++ has a fairly comprehensive user guide and a developer guide that is presently very incomplete. The following are rough guidelines for what you should provide:

  • If you are writing user-exposed features, you should add some documentation to the user guide on how to use them.
  • Any functions/classes should include Doxygen documentation.
  • Generally, code should be well-commented using regular C++ comments to explain its function to help in reviewing it.

Nektar++ also has a growing number of tutorials to help introduce users and developers to the use of the library and the range of application solvers. These are stored in a separate repository, but are available from the main repository through a git submodule. To populate the docs/tutorial directory run git submodule init followed by git submodule update --remote. The latter command will ensure you have the latest master branch of the tutorials within your source tree.

Code review and merging

All merge requests will be reviewed by one of the senior developers. We try to stick to the following process:

  • Senior developer will be assigned, MR will be assigned a milestone to target a release.
    • If the branch is deemed to be minor and passes the checklist above, senior developer will handle the request by themselves.
    • Otherwise, senior developer will ask one or more other developers to review the code.
  • Submission checklist will be checked by the reviewers.
  • Where appropriate, reviewers will comment on regions of code that need further development and/or improvement.
  • In addition to any coding comments/suggestions, reviewers are asked to check the branch passes the regression tests and appropriate documentation has been added.
  • Once feedback received from the branch author (if necessary) and reviewers are happy, the branch will be merged.

Release branches

Nektar++ releases are versioned in the standard form x.y.z where x is a major release, y a minor release and z a patch release:

  • major releases are extremely infrequent (on the order of every 2-3 years) and denote major changes in functionality and the API;
  • minor releases occur around twice per year and contain new features with minor API changes;
  • patch releases are targeted on roughly a monthly basis and are intended to fix minor issues in the code.

The repository contains a number of release branches named release/x.y for each minor release, which are intended to contain fixes and very minor changes from master and which form the next patch release. This allows us to use master for the next minor release, whilst still having key fixes in patch releases.

Cherry-picking process

Any branches that are marked with the Proposed patch label should follow the following additional steps to cherry pick commits into the release/x.y branch.

  1. If the branch is on a remote other than nektar/nektar, make sure that's added to your local repository.

  2. On a local terminal, run git fetch --all to pull the latest changes. It's important for the commands below that you do this before you merge the branch into master.

  3. Merge the branch into master as usual using GitLab.

  4. Switch to the appropriate branch with git checkout release/x.y and update with git pull.

  5. Now check the list of commits to cherry-pick.

    git log --oneline --no-merges --reverse origin/master..REMOTE/fix/BRANCHNAME

    where REMOTE is the remote on which the branch lives and BRANCHNAME is the fix branch. If the list is empty, you probably did a git fetch after you merged the branch into master; in this case use origin/master^.

  6. If you're happy with the list (compare to the MR list on the GitLab MR if necessary), cherry-pick the commits with the command:

    git cherry-pick -x $(git rev-list --no-merges --reverse origin/master..REMOTE/fix/BRANCHNAME)
  7. It's likely you'll encounter some conflicts, particularly with the CHANGELOG. To fix these:

    • git status to see what's broken
    • Fix appropriately
    • git commit -a to commit your fix
    • git cherry-pick --continue
  8. If everything becomes horribly broken, git cherry-pick --abort.

  9. Once you're happy, git push to send your changes back to GitLab.

Steps 5 and 6 can be simplified by creating a script

#!/bin/bash
src=$1

logopts="--oneline --no-merges --reverse"
commits=`git log $logopts master..$1 | cut -f 1 -d " " | xargs`

echo "Will cherry-pick the following commits: $commits"
echo "Press ENTER to continue..."
read

cherryopts="-x --allow-empty --allow-empty-message"
git cherry-pick $cherryopts $commits

which accepts the name of the source branch as the sole argument.

Formatting guidelines

Nektar++ uses C++, a language notorious for being easy to make obtuse and difficult to follow code. To hopefully alleviate this problem, there are a number of fairly simple formatting guidelines you should follow.

To help with this, we now require the use of clang-format to ensure all code in the library is formatted consistently (see below). When adding new contributions, or making changes to existing code, please ensure you run clang-format on these files to ensure the formatting complies with our guidelines. There is a CI job which will verify that no further formatting changes are required before merging.

Basic rules

  • All code should be wrapped to 80 characters.

  • Indentation should be 4 spaces with no tabs. Namespaces should not be indented to give more room in the 80 character width.

  • Please comment your code with Doxygen and inline comments wherever possible -- but don't use trailing inline comments to save the 80 character limit!

  • All code blocks (even one-line blocks) should use braces, and braces should be on new lines; for instance

    if (someCondition)
    {
        myAwesomeFunction();
    }
  • Don't use preprocessor directives and macros unless there is no viable alternative.

  • However, please make sure you do have a header guard inside your .h files, which you should be sure to include in any headers you contribute.

  • Use one .cpp and .h file per C++ class, and try to keep inline header code to a minimum (unless performance is a major factor).

  • Put spaces around binary operators and constants.

  • Put spaces after if, while, etc., but not after function names (see the example above).

Variables and naming

  • Please use sensible names and use camelCase as a broad naming convention.
    • Variables should start with a lowercase letter, e.g. myAwesomeVariable.
    • Function, class, struct and typedef names should begin with capital letters, e.g. MyAwesomeFunction.
  • Inside classes, member variables should be prefixed with m_, e.g. m_myAwesomeVariable.
    • Global constants used throughout the library should be prefixed with k (e.g. kGeometricTolerance), and enumerations should be prefixed with e (e.g. eGeometry).
  • Use all uppercase letters with underscores between words for pre-processor definitions and macros.

Using clang-format

Code formatting is reasonably boring, so Nektar++ comes with a .clang-format file to allow for automatic code formatting.

Installing it is straightforward on most package managers. Nektar++'s source code is formatted using clang-format 11.

There are a number of instructions on how to use clang-format inside a number of text editors on the CLang website. However at a minimum, you should consider downloading the git-clang-format script into one of your $PATH locations. You can then run the command

git clang-format

before you do a git commit, and clang-format will automatically format your diff according to the guidelines.