package / ocaml-base-compiler.4.10.0 / CONTRIBUTING.md

How to contribute changes

:+1::tada: First off, thank you for taking time to contribute! :tada::+1:

The following is a set of guidelines for proposing changes to the OCaml distribution. These are just guidelines, not rules, use your best judgment and feel free to propose changes to this document itself in a pull request.

This document assumes that you have a patch against the sources of the compiler distribution, that you wish to submit to the OCaml maintainers upstream. See INSTALL.adoc for details on how to build the compiler distribution from sources. See HACKING.adoc for details on how to modify the sources.

Contribution

Modifying its sources is far from the only way to contribute to the OCaml distribution. Bug reports (in particular when they come with a reproducible example), simple typos or clarifications in the documentation also help, and help evaluating and integrating existing change proposals also help. Providing good answers on the discussion forums, or asking the good questions that highlight deficiencies in existing documentations, also help. We currently have more contributors willing to propose changes than contributors willing to review other people's changes, so more eyes on the existing change requests is a good way to increase the integration bandwidth of external contributions.

There are also many valuable ways to contribute to the wider OCaml ecosystem that do not involve changes to the OCaml distribution.

The rest of the document is concerned with the form of change proposals against the OCaml distribution. (Code changes, but also improvement to documentation or implementation comments, which are valuable changes on their own.)

Workflow

All changes to the OCaml distribution need to be processed through the GitHub Pull Request (PR) system. In order to propose a change, a contributor thus needs to have a GitHub account, fork the ocaml/ocaml repository, create a branch for the proposal on their fork and submit it as a Pull Request on the upstream repository. (If you are not yet familiar with GitHub, don't worry, all these steps are actually quite easy!)

The current rule is that a PR needs to get an explicit approval from one of the core maintainer in order to be merged. Reviews by external contributors are very much appreciated.

Since core maintainers cannot push directly without going through an approved PR, they need to be able to apply small changes to the contributed branches themselves. Such changes include fixing conflicts, adjusting a Changelog entry, or applying some code changes required by the reviewers. Contributors are thus strongly advised to check the Allow edits from maintainer flag on their PRs in the GitHub interface. Failing to do so might significantly delay the inclusion of an otherwise perfectly ok contribution.

Coding guidelines

You should not leave trailing whitespace; not have line longer than 80 columns, not use tab characters (spaces only), and not use non-ASCII characters. These typographical rules can be checked with the script tools/check-typo.

If you are working from a Git clone, you can automate this process by copying the file tools/pre-commit-githook to .git/hooks/pre-commit.

Otherwise, there are no strongly enforced guidelines specific to the compiler -- and, as a result, the style may differ in the different parts of the compiler. The general OCaml Programming Guidelines are good to keep in mind, and otherwise we strive for good taste and local consistency (following the code located around your change).

If you strongly feel that a style-related change would improve quality of the existing code (for example, giving more descriptive names to some variables throughout a module, factoring repeated code patterns as auxiliary functions, or adding comments to document a part of the code that you had trouble understanding), you can have code cleanup commits at the beginning of your patch series, or submit code cleanups as your change proposal. Those cleanups should remain separate commits from the functional changes in the rest of the patch series; it is easier to review commits that are specifically marked as exactly preserving the code semantics.

Test you must.

Whenever applicable, merge requests must come with tests exercising the affected features: regression tests for bug fixes, and correctness tests for new features (including corner cases and failure cases). For regression tests, testing other aspects of the feature (in particular, related edge cases) that are not currently covered is a good way to catch other instances of bugs -- this did happen several times in the past. Warnings and errors should also be tested.

Tests go in the sub-directories of testsuite/tests. Running make all in testsuite/ runs all tests (this takes a few minutes), and you can use make one DIR=tests/foo to run the tests of a specific sub-directory. There are many kind of tests already, so the easiest way to start is to extend or copy an existing test.

In general, running a test produces one (or several) .result file, that are compared to one (or several) .reference file present in the repository; the test succeeds if they are identical. If your patch breaks a test, diffing the .result and .reference file is a way to see what went wrong. Some reasonable compiler changes affect the compiler output in way that make those outputs differ (for example slight modifications of warning or error messages may break all tests checking warnings). If you are positive that the new .result file is correct (and that the change in behavior does not endanger backward compatibility), you can replace the old .reference file with it. Finally, when adding new tests, do not forget to include your .reference files (but not .result) in the versioned repository.

Testing is also a way to make sure reviewers see working (and failing) examples of the feature you fix, extend or introduce, rather than just an abstract description of it.

Run tests before sending a PR

You should run all the tests before creating the merge request or pushing new commits (even if Travis will also do it for you): make tests (this takes a few minutes).

Unfortunately some of the lib-threads test are non-deterministic and fail once in a while (it's hard to test these well). If they consistently break after your change, you should investigate, but if you only see a transient failure once and your change has no reason to affect threading, it's probably not your fault.

Benchmarking

If your contribution can impact the performance of the code generated by the native compiler, you can use the infrastructure that the flambda team put together to benchmark the compiler to assess the consequences of your contribution. It has two main accessible parts:

  • The website that hosts benchmarks results, at http://bench.flambda.ocamlpro.com/. It exposes two ways to compare compilers: the first, under the header Plot a given benchmark, allows to select a benchmark and see graphs plotting the evolution of the performance of the different compilers over time. The second, under Compare two runs, allows to get an overview of the differences between a reference compiler (selected using the ref button) and a compiler under test (using the tst button). Clicking on the Compare button at the bottom right of the page will create a new page containing summaries and raw data comparing the selected runs.

  • The git repository containing the data about which benchmarks to run, on which compilers, at https://github.com/OCamlPro/ocamlbench-repo. This needs to be a valid opam 2.0 repository, and contains the benchmarks as normal packages and the compilers as versions of the package ocaml-variants. To add a compiler to the list, you must have a publicly accessible version of your branch (if you're making a pull request again the compiler, you should have a branch on github that was used to make the pull request, that you can use for this purpose). Then, you should make a pull request against ocamlbench-repo that adds a repertory in the packages/ocaml-variants sub-folder which contains a single opam file. The contents of the file should be inspired from the other files already present, with the main points of interest being the url field, which should point to your branch, the build field that should be adapted if the features that you want to benchmark depend on configure-time options, and the setenv field that can be used to pass compiler options via the OCAMLPARAM environment variable. The trunk+flambda+opt compiler, for instance, both uses a configure option and sets the OCAMLPARAM variable. The folder you add has to be named ocaml-variants.%VERSION%+%DESCR%, where %VERSION% is the version that will be used by opam to check compatibility with the opam packages that are needed for the benchmarks, and %DESCR% should be a short description of the feature you're benchmarking (if you're making a pull request against ocaml, you can use the PR number in the description, e.g. +gpr0000). Once your pull request is merged, it will likely take a few hours until the benchmark server picks up the new definition and again up to a few hours before the results are available on the results page.

Description of the proposed change

In the merge request interface

The description of the merge request must contain a precise explanation of the proposed change.

Before going in the implementation details, you should include a summary of the change, and a high-level description of the design of the proposed change, with example use-cases.

In the patches

If some of the explanations you provide for the merge request would make sense as comments in the code, or documentation in the manual, you should include them there as well.

In-code comments help make the codebase more accessible to newcomers (many places in the compiler could benefit from a few extra explanations), and they are also useful to code reviewers. In particular, any subtlety in code that cannot be made self-explanatory should come with an explanation in comment. If you add some non-obvious code specifically to fix a bug, include the issue number in comments.

Do not assume that code reviewers are all experts in the existing codebase. If you use subtle code, add a comment, even if the same kind of code is used somewhere else in the same module. (If this is a common and useful domain-specific idiom that is already explained somewhere, pointing to this explanation in your commit message is better than adding redundant explanations.)

User documentation

Changes affecting the compiler libraries should be reflected in the documentation comments of the relevant .mli files.

It is recommended to included changes to the OCaml Reference Manual (in particular for any change in the surface language), which is now part of the main repository (under manual/).

Finally, changes in command-line options should be integrated in the manual, but also in the man pages present in the man/ sub-directory of the OCaml distribution.

Changelog

Any user-visible change should have a Changes entry:

  • in the right section (named sections if major feature, generic "Bug fixes" and "Feature requests" otherwise)

  • using the label "*" if it breaks existing programs, "-" otherwise

  • with all relevant issue and PR numbers #{N}, in ascending numerical order (separated by commas if necessary)

  • maintaining the order: the entries in each section should be sorted by issue/PR number (the first of each entry, if more than one is available)

  • with a concise readable description of the change (possibly taken from a commit message, but it should make sense to end-users reading release notes)

  • crediting the people that worked on the feature. The people that wrote the code should be credited of course, but also substantial code reviews or design advice, and the reporter of the bug (if applicable) or designer of the feature request (if novel).

  • following the format

      {label} {issue number(s)}: {readable description}
              ({credits})
    
    note that the `{credits}` should be on their own line, aligned with the
    issue number for readability
    (`{readable description}` can be multiline to not overflow 80
    columns, and should be aligned with the issue number as well.)
    

This changelog can be included in the main commit, if the merge request is just one patch, or as a separate commit, if it's a patch series and no particular commit feels best suited to receive the Changelog entry.

(Do not under-estimate the importance of a good changelog. Users do read the release notes, and things forgotten from the changelog will cause pain or regrets down the line.)

Clean patch series

Clean patch series are useful, both during the review process and for code maintenance after it has been merged. Before submitting your request, you should rebase your patch series:

  • on top of the OCaml branch in which you want to merge (usually trunk), solving any conflicts.

  • into a few well-separated, self-contained patches (github PRs can generate gazillions of micro-changes)

  • erasing history that does not make sense after the issue is merged (back-and-forth between different designs, etc. The PR number allows interested people to go back to the original discussion if needed.)

  • bisectable: the distribution should be in a good state after the application of each patch (in particular, later commits that fix bugs in previous commits should always be squashed into the commit they fix)

  • with readable commit messages (this is for future developers needing to understand a change that happened in the past). Commit messages should not overflow 80 columns, with the following format:

      {one-liner header description (with issue number if applicable)}
      {blank line}
      {one or several paragraphs of explanation if needed}
    

During review, you may make many other changes to the patch series. You can rebase it on the fly (if you git push -f on the branch of the pull request in your personal clone, Github will update the pull request automatically; remember to always create a new branch for any) or wait until the discussion has converged, once we agree the request is ready for merging. Doing a good rebase is grunt work that takes some time and care (use git log -u to make sure the rebase patches make sense), but:

  • It is easier and faster to do for the author of the patch than for others (if rebasing against the current trunk creates a conflict with another change you don't understand well, feel free to ask).

  • Maintainers are usually short on time, and asking them to do a rebase means they have less time to review and merge other contributions.

  • The long-term benefits of keeping a clean, bisectable history cannot be overstated. Imagine that in three years, under the pressure of a coming release, a contributor ends up somewhere in the middle of your patch series, wondering if or why it is the cause of a specific issue. Wasting his or her time then (with a "yolo" commit message, a big ugly commit of unrelated changes, or an un-testable intermediary state) is a sure way to generate ill will.

Contributing to the standard library

Contributions to the standard library are very welcome. There is some widespread belief in the community than the stdlib is somehow "frozen" and that its evolutions are mostly driven by the need of the OCaml compiler itself. Let's be clear: this is just plain wrong. The compiler is happy with its own local utility functions, and many recent additions to the stdlib are not used by the compiler.

Another common and wrong idea is that core OCaml maintainers don't really care about the standard library. This is not true, and won't be unless one of the "alternative standard" libraries really gains enough "market share" in the community.

So: please contribute!

Obviously, the proposals to evolve the standard library will be evaluated with very high standards, similar to those applied to the evolution of the surface langage, and much higher than those for internal compiler changes (optimizations, etc).

A key property of the standard library is its stability. Backward compatibility is not an absolute technical requirement (any addition to/of a module can break existing code, formally), but breakage should be limited as much as possible (and assessed, when relevant). A corollary is that any addition creates a long-term support commitment. For instance, once a concrete type or function is made public, changing the exposed definition cannot be done easily.

There is no plan to extend dramatically the functional domain covered by the standard library. For instance, proposals to include support for XML, JSON, or network protocols are very likely to be rejected. Such domains are better treated by external libraries. Small additions to existing modules are much simpler to get in, even more so (but not necessarily) when:

  • they cannot easily be implemented externally, or when
  • they facilitate communication between independent external libraries, or when
  • they fill obvious gaps.

Of course, standard guidelines apply as well: proper documentation, proper tests, portability (yes, also Windows!), good justification for why the change is desirable and why it should go into stdlib.

So: be prepared for some serious review process! But yes, yes, contributions are welcome and appreciated. Promised.

Contributing optimizations

Contributions to improve the compiler's optimization capabilities are welcome. However, due to the potential risks involved with such changes, we ask the following of contributors when submitting pull requests:

  • Explain the benefits of the optimization (faster code, smaller code, improved cache behaviour, lower power consumption, increased compilation speed).

  • Explain when the optimization does and does not apply.

  • Explain when, if ever, the optimization may be detrimental.

  • Provide benchmark measurements to justify the expected benefits. Measurements should ideally include experiments with full-scale applications as well as with microbenchmarks. Which kinds of measurements are appropriate will vary depending on the optimization; some optimizations may have to be measured indirectly (for example, by measuring cache misses for a code size optimization). Measurements showing clear benefits when combined with some other optimization/change are acceptable.

  • At least some of the measurements provided should be from experiments on open source code.

  • If assistance is sought with benchmarking then this should be made clear on the initial pull request submission.

  • Justify the correctness of the optimization, and discuss a testing strategy to ensure that it does not introduce bugs. The use of formal methods to increase confidence is encouraged.

A major criterion in assessing whether to include an optimisation in the compiler is the balance between the increased complexity of the compiler code and the expected benefits of the benchmark. Contributors are asked to bear this in mind when making submissions.

Contributor License Agreement

We distinguish two kind of contributions:

  • Small changes that do not bear a specific mark of their authors (another developer recreating the change without access to the original patch would write an indistinguishable patch), and are thus not protected by copyright, do not require any particular paperwork. This is convenient for everyone, and of course does not mean that those contributions are of lesser importance. (For example a bugfix can be obvious once a bug is understood, reported and reproduced, and yet invaluable for users.)

  • Larger changes that are covered by copyright. For them, we require contributors to sign a Contributor License Agreement (CLA), which gives INRIA (Institut National de Recherche en Informatique et en Automatique) the rights to integrate the contribution, maintain it, evolve it, and redistribute it under the license of its choice. This is not a copyright assignment (as requested by the Free Software Foundation for example), contributors retain the copyright on their contribution, and can use it as they see fit. The OCaml CLA is lightly adapted from the CLA of the Apache Foundation, and is available in two versions: for individual contributors and for corporations.

You must understand that, by proposing a contribution for integration in the OCaml distribution, you accept that it be considered under one of those regimes. In particular, in all cases you give INRIA the permission to freely re-license the OCaml distribution including the contribution.

This ability to re-license allows INRIA to provide members of the Caml Consortium with a license on the Caml code base that is more permissive than the public license.

How to sign the CLA

If your contribution is large enough, you should sign the CLA. If you are contributing on your own behalf, you should sign the individual CLA. For corporate contributions, if your employer has not already done so, they should sign the corporate CLA. Review the CLA, sign it, and send it -- scanned PDF by email, or postail mail -- to Xavier Leroy (contact info).