Skip to content

Latest commit

 

History

History
197 lines (158 loc) · 8.37 KB

CODING.md

File metadata and controls

197 lines (158 loc) · 8.37 KB

UHD Coding Standards

Note: This file applies to all code within the UHD repository, not just for code that is actually part of the UHD library.

Preamble

To quote R. W. Emerson: "A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines". Ignoring the little statesmen for a minute, these coding standards are here to make our life easier, not simply add additional rules. They are meant as additional guidance for developers, and are not meant to be interpreted as law.

So, ultimately, it is up to the developer to decide how much these guidelines should be heeded when writing code, and up to reviewers how much they are relevant to new submissions. That said, a consistent codebase is easier to maintain, read, understand, and extend. Choosing personal preferences over these coding guidelines is not a helpful move for the team and future maintainability of the UHD codebase.

General Coding Guidelines

  • Never submit code with trailing whitespace.
  • Code layout: We use 4 spaces for indentation levels, and never tabs.
  • Code is read more often than it's written. Code readability is thus something worth optimizing for.
  • Try and keep lines short (79 characters is nice, maximum is 90/100 characaters for C/C++ and Python, respectively), unless readability suffers. We often have to do fun things like SSH into machines and edit code in a terminal, and do side-by-side views of code on small-ish screens, so this is actually pretty helpful.
  • Go crazy with log messages. Trace-level log messages in particular can be used copiously and freely (unless in rare cases where they can interfere with performance). Note that in C++, we have the option of fully compiling out trace-level messages (and even higher levels).

C++-specific Guidelines

  • All C++ code must be formatted according to the .clang-format file in the root of the project.
  • If in doubt, consult the C++ Core Guidelines. If the guidelines have an answer, and it works for you, just pick that.
  • Use Doxygen doc-blocks copiously.
  • All things equal, prefer standard C++ constructs over Boost constructs (see also Boost guidelines).
  • Given the option, prefer C++ lambdas over std::bind, and just don't use boost::bind.
  • size_t is the correct container for all indexing of C++ structures (such as vectors). But keep in mind that the size of size_t is platform-dependent!
  • Use size-specific types whenever interacting with hardware (int32_t, etc.). It's very easy to get bitten by incorrect sizes.
  • Include include files in the following order: Local headers, other UHD headers, 3rd-party library headers, Boost headers, standard headers. The rationale is to include from most to least specific. This is the best way to catch missing includes (if you were to include the standard header first, it would be available to all include files that come later. If they need that standard header too, they should be including it themselves). Note that clang-format will do this for you. Example:
#include "x300_specific.hpp"
#include <uhd/utils/log.hpp>
#include <libusb/foo.hpp>
#include <boost/shared_ptr.hpp>
#include <mutex>
  • Feel free to use modern C/C++ features even if they were not used before. Make sure they work with the compilers and dependencies which are set for the version of UHD the commit will be made upon. Our continuous integration system will be able to confirm this. For interesting new features, use 'anchor commits', which describe clearly which new feature was used. They can be used as a reference for other developers. Example:
commit 9fe731cc371efee7f0051186697e611571c5b41b
Author: Andrej Rode <[email protected]>
Date:   Tue Nov 22 16:19:38 2016 -0800

    utils: uhd_find_devices display one device for each unique serial found
    
    Note: This also is the first precedent for the usage of the 'auto' keyword
    in UHD. Commits past this one will also be able to use 'auto'.
    
    Reviewed-By: Martin Braun <[email protected]>
  • Prefer .at() over [] for maps and vectors. Keep in mind that [] will invoke a default constructor of the value type, whereas .at() will throw an exception if the index doesn't exist -- which is usually the desired behaviour.

Boost-specific Guidelines

  • Avoid Boost where possible.
  • Don't use Boost's sleep functions. Prefer std::chrono functions.
  • When using boost::assign::list_of or boost::assign::map_list_of, make sure to explicitly cast to the appropriate container (even better: Avoid boost::assign, maybe use std initializer lists where possible):
std::vector<std::string> foo =
    boost::assign::list_of<std::string> // Explicitly list the type (in this
        ("string1")                     // case, std::string). Then list all
        ("string2")                     // the items (string1, string2, etc).
        ("etc")
    .to_container(foo); // Finally, call conversion routine explicitly.
// Same for maps:
std::map<std::string, std::string> bar =
    boost::assign::map_list_of<std::string, std::string>
        ("a", "b")
        ("c", "d")
        ("etc", "etc")
    .to_container(bar);

Python-specific Guidelines

  • Starting with UHD 4.0, Python 2 is no longer supported, and we don't need to accommodate for it any longer. Prefer Python 3 constructs.
  • We follow the NI Python Style Guidelines. Install ni-python-styleguide to help following the recommendations (e.g., by running pip install ni-python-styleguide. The tool is available on GitHub.

CMake-specific Guidelines

  • CMake commands are written in lowercase.

Revision Control Hygiene

  • In this repository, we almost always use fast-forward merges, and no merge commits.
  • Prefix all commit message subject lines with the section of code they apply to, and use the imperative mood (Example: "x300: Fix overflow at full moon"). Try and keep the subject line to 50 characters, but make 72 characters a hard limit.
  • Follow up in greater detail in the body of the commmit message. The body is separated from the subject line with one blank line. Consider the body of the git commit an email to the future reader of this changeset, so don't be sparse in the commit body, and use it to explain the what and why of this commit (the "how" part should be obvious from the code change). Lines should be limited to 72 characters.
  • Avoid refactoring, whitespace cleanup, or fixing code to match coding guidelines in the same commit as modifying behaviour. Prefer dedicated cleanup commits.
  • Remember that we ship git repositories, not just code. This means every commit is part of our product and should be treated as such.

Useful tooling

Autoformatters

  • Before checking in C/C++/Python code, make sure it was reformatted with an autoformatter. For C/C++, that's clang-format. For Python, use black and isort, or install ni-python-styleguide which incorporates all of those tools.

Linters

There are a variety of linters available, all of which provide valuable insight:

  • flake8 is a very, very nitpicky linter for Python. It is integrated into ni-python-styleguide with some additional configuration, so we recommend using the latter.
  • clang-tidy can do some static analysis on C and C++ files.
  • For HDL code, Verilator and iverilog can be integrated as syntax error checkers in some editors, but keep in mind they often can't find available modules, and not always support SystemVerilog.

Git hooks

The easiest way to use git hooks is to use pre-commit:

  • Install pre-commit (e.g., pip install pre-commit, or apt install pre-commit).
  • In your UHD repository, install the pre-commit hooks by running pre-commit install
  • Manually install ni-python-styleguide (pip install ni-python-styleguide)

Before attempting to commit changes, the files will be automatically scanned for style/formatting issues.

FPGA Guidelines

FPGA guidelines are stored in a separate file. See CODING.md in the FPGA repository.