CONTRIBUTING.md - sdk - Git at Google

 Want to contribute? Great! First, read this page (including [the small print](#the-small-print) and [AI contribution guidelines](#ai-contribution-guidelines)  at the end).

 ## Ways you can contribute

 You can help the Dart project in many ways, in addition to contributing code. For example, you can [report bugs](https://dartbug.com), ask and answer [Dart questions on StackOverflow](https://stackoverflow.com/questions/tagged/dart), and improve the documentation.

 If you'd like to improve the documentation, you have three options:

   * Give us feedback:
     * If you're looking at a page with a **bug icon** at the **upper right**,
       click that icon to report a bug on the page.
     * To report an API doc bug,
       [create an SDK issue](https://github.com/dart-lang/sdk/issues/new?title=API%20doc%20issue:).
   * Contribute to the Dart developer websites such as [dart.dev](https://dart.dev) (repo: [dart-lang/site-www](https://github.com/dart-lang/site-www)). For more information, see [Writing for Dart and Flutter websites](https://github.com/dart-lang/site-shared/blob/master/doc/writing-for-dart-and-flutter-websites.md).
   * Improve the API reference docs at [api.dart.dev](https://api.dart.dev) by editing doc comments in the [Dart SDK repo](https://github.com/dart-lang/sdk/tree/main/sdk/lib). For more information on how to write API docs, see [Effective Dart: Documentation](https://dart.dev/guides/language/effective-dart/documentation).

 ## Before you contribute

 Before we can use your code, you must sign the [Google Individual Contributor License Agreement](https://developers.google.com/open-source/cla/individual) (CLA), which you can do online.  The CLA is necessary mainly because you own the copyright to your changes, even after your contribution becomes part of our codebase, so we need your permission to use and distribute your code.  We also need to be sure of various other things—for instance that you'll tell us if you know that your code infringes on other people's patents.  You don't have to sign the CLA until after you've submitted your code for review and a member has approved it, but you must do it before we can put your code into our codebase.

 Before you start working on a larger contribution, you should get in touch with us first through the  [Dart Issue Tracker](https://dartbug.com) with your idea so that we can help out and possibly guide you. Coordinating up front makes it much easier to avoid frustration later on.

 All submissions, including submissions by project members, require review.  We use the same code-review tools and process as the chromium project.

 We occasionally take pull requests, e.g., for comment changes, but the main flow is to use the Gerrit review system as explained below.

 ## Submitting patches directly from GitHub

 This repository primarily uses [Gerrit for code reviews](https://dart-review.googlesource.com/), rather than GitHub PRs. However, you may submit [a GitHub PR ](https://github.com/dart-lang/sdk/pulls), e.g. to edit some API documentation, and it will be automatically converted into a Gerrit change list (a "CL") by a copybara-service bot. You can find the link to that CL as a comment left on the PR by the bot. Any changes made to the PR after opening will also be sync'ed by the bot into the Gerrit review. The PR will be automatically closed when the CL is reviewed and landed.

 ## Setting up the environment

 In order to submit a patch from a local workspace, you need to install the Chromium
 [depot\_tools](https://commondatastorage.googleapis.com/chrome-infra-docs/flat/depot_tools/docs/html/depot_tools_tutorial.html#_setting_up).

 ## Getting the code

 To work with the Dart code, you need to download and build the development branch. Active development of Dart takes place on the `main` branch, from which we push "green" versions that have passed all tests to `dev` branch. Complete instructions are found at [Getting The Source](docs/Building.md#getting-the-source). **You must use the `gclient` tool (`fetch`), using `git clone` will not get you a functional environment!**

 ## Starting a patch with git

 Create a new branch using `git new-branch` (this is a command added by the aforementioned depot_tools).

 You can be in any branch when you run `git new-branch`.

 ```bash
 git new-branch <feature name>
 <write code>
 git commit
 <write code...>
 git commit
 ...
 ```

 ## Keeping your branch updated with origin/main

 As you work, and before you send a patch for review, you should
 ensure your branch is merging cleanly to `origin/main`.

 There are multiple ways to do this, but we generally recommend
 using `git rebase-update` (another feature added by depot_tools):

 ```bash
 git rebase-update
 ```

 You can run this command from any branch.

 This command will fetch
 `origin/main`, rebase all your open branches, and delete
 cleanly merged branches.

 Your local workflow may vary.

 ## Uploading the patch for review

 Upload the patch to Gerrit for review using `git cl upload`:

 ```bash
 # Upload a CL, add reviewers from OWNERS files, and allow automatic submission.
 git cl upload --send-mail --r-owners --auto-submit
 ```

 The above command returns a URL for the review. Attach this review to your issue in https://dartbug.com.

 To update the CL, just commit your changes and run `git cl upload` for your branch again.

 More detailed instructions for the `git cl` tools available on https://commondatastorage.googleapis.com/chrome-infra-docs/flat/depot_tools/docs/html/depot_tools_tutorial.html#_creating_uploading_a_cl

 ## Getting your patch reviewed and landed submitted.

 Ensure that your change has at least two reviewers who are OWNERS on your CL:

 ```bash
 git cl web # opens your review on https://dart-review.googlesource.com
 ```

 Check the "Reviewers" on the CL, there should be at least two. If there aren't,
 use the "ADD OWNERS" button to add more. If you can't get enough reviewers, or
 the reviewers are not responsive, you may add reviewers found in the `OWNERS`
 file at the root of the repository to escalate.

 It's part of a code owner's responsibility to review changes in their area and
 help you to get your patch submitted. They may also provide comments, or reject
 your CL. They can run the presubmit tests on your behalf and submit your patch.

 ## For committers: Submitting a patch

 If you have commit access, when the review is done and the patch is good to go, submit the patch on https://dart-review.googlesource.com:

 *   Press "Submit to CQ" (CQ stands for "Commit Queue").
 *   You can follow the progress by looking at the "Tryjobs" panel in your review.
 *   Once the Commit Queue is green, the patch will be merged.
 *   If any of the try jobs is red, you will have to fix the errors and then "Submit to CQ" once more.

 If you do not have commit access, a Dart engineer will commit on your behalf, assuming the patch is reviewed and accepted.

 ## For committers: Merging contributions from non-members

 If the author of a patch is not a committer, they will need help landing the patch.
 Once a patch gets an LGTM, it's easy for a committer to merge it in.

 * Find and open the review on https://dart-review.googlesource.com.
 * Follow the instructions in the previous section to submit the patch.

 ## Coding style

 The source code of Dart follows the:

   * [Google C++ style guide](https://google.github.io/styleguide/cppguide.html)
   * [Dart style guide](https://dart.dev/guides/language/effective-dart/style)

 You should familiarize yourself with those guidelines.

 All files in the Dart project must start with the following header. If you add a new file please also add this. The year should be a single number (not a range; don't use "2011-2012", even if the original code did).  If you edit an existing file you don't have to update the year.

 ```dart
 // Copyright (c) 2017, the Dart project authors. Please see the AUTHORS file
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 ```

 ## AI contribution guidelines

 PRs prepared using AI tools must follow these requirements:

 1. You must review all AI-generated code before uploading it for review, and before requesting
    re-review of any updates to the uploaded change.
    * You are responsible for ensuring that code you submit meets coding standards.
    * Unmodified AI output generally does _not_ meet those standards.
 2. You must ensure that uploaded code compiles and passes at least the tests you have yourself
    added as part of the change.
 3. You must understand and be able to discuss the code you are uploading for review.
    * Non-trivial changes require discussion and iteration during review. If you do not understand the
      code, you cannot meaningfully respond to review feedback.
    * In our experience, simply feeding review feedback into an AI agent and uncritically reposting
      its output will not lead to a constructive review.
 4. You must verify the accuracy of any AI-generated text you include in the change description or
    review discussion comments.
    * If an AI provides you incorrect information, it is just hallucinating; if you choose to paste
      that text into GitHub, you are misrepresenting your change to your reviewer.
    * In particular, do not tell a reviewer that you have addressed their feedback just because
      AI output says so. It is your responsibility to make sure that review feedback has actually
      been addressed.

 ### Reviewer guidelines

 Because the Dart team’s time is limited, and the capacity for people outside the team to generate
 plausible-looking code is unlimited, be mindful as a reviewer about what code you choose to review.
 Consider immediately closing changes that have any of the following red flags:

 * The change description has entirely replaced our template with AI-generated output.
   * If the contributor did not follow our process from the outset, they are unlikely to
     understand what is expected of them during review.
 * The change description does not match the changes.
   * If the contributor did not review both the changes and the description enough to notice this,
     they have not followed the AI contribution policy.
 * The change contains irrelevant AI-generated files, such as agent planning .md files.
   * If the contributor did not review the changes enough to notice and remove these files, they
     have not followed the AI contribution policy.

 As always when closing a change, explain why and provide next steps.

 As a guiding principle, if at any point in the process you feel that you are getting unfiltered
 or minimally filtered AI output as code and/or comment responses, ask yourself:

 * If this change weren’t here, would I choose to spend my time fixing this issue?
 * Would I choose to fix it using an AI agent that took hours or days to respond to every prompt?

 Unless the answer to both questions is yes, the review is not a good use of your time.

 This applies even if you are multiple rounds into the review: if the contributor closed the change,
 would you take it over using an extremely high-latency agent? Beware the sunk cost fallacy.

 ### Philosophy

 A common question in discussions around AI policies for open source projects is: “Why have an
 AI policy at all? Why does it matter how the code was created; shouldn’t the code speak for itself?”

 In general, we agree, which is why our policy focuses on behaviors rather than tools. The behaviors
 that led to the creation of this policy are problematic regardless of whether they are AI-generated
 or human-generated. However, it is useful to highlight these policies in the context of AI because
 we have seen that these behaviors are orders of magnitude more common when AI is involved. For example:

 * Submitting a change containing hundreds of lines of code that the contributor doesn't understand is
   always a problem. This was rare before the widespread use of AI agents, but is more common with
   AI-assisted development.
 * Having several rounds of exchanges where a reviewer asks for a change, and the contributor says
   that they have made that change but haven't, is always a waste of reviewer time. It’s very rare
   for a contributor to deliberately and obviously lie to a reviewer, but unfortunately common
   for contributors to uncritically repeat AI agent hallucinations.
 * Changes that ignore our process have always been problematic, because standardizing the change process
   is an important part of how we keep our review load manageable. However, a contributor who has
   spent hours or days on a change is much more likely to take some time to learn and follow our process
   to avoid having that effort be wasted than someone who spent a few minutes generating the change.


 ## The small print

 Contributions made by corporations are covered by a different agreement than the one above, the [Software Grant and Corporate Contributor License Agreement](https://cla.developers.google.com/about/google-corporate).

 We pledge to maintain an open and welcoming environment. For details, see our [code of conduct](https://dart.dev/code-of-conduct).
	Want to contribute? Great! First, read this page (including [the small print](#the-small-print) and [AI contribution guidelines](#ai-contribution-guidelines) at the end).

	## Ways you can contribute

	You can help the Dart project in many ways, in addition to contributing code. For example, you can [report bugs](https://dartbug.com), ask and answer [Dart questions on StackOverflow](https://stackoverflow.com/questions/tagged/dart), and improve the documentation.

	If you'd like to improve the documentation, you have three options:

	* Give us feedback:
	* If you're looking at a page with a bug icon at the upper right,
	click that icon to report a bug on the page.
	* To report an API doc bug,
	[create an SDK issue](https://github.com/dart-lang/sdk/issues/new?title=API%20doc%20issue:).
	* Contribute to the Dart developer websites such as [dart.dev](https://dart.dev) (repo: [dart-lang/site-www](https://github.com/dart-lang/site-www)). For more information, see [Writing for Dart and Flutter websites](https://github.com/dart-lang/site-shared/blob/master/doc/writing-for-dart-and-flutter-websites.md).
	* Improve the API reference docs at [api.dart.dev](https://api.dart.dev) by editing doc comments in the [Dart SDK repo](https://github.com/dart-lang/sdk/tree/main/sdk/lib). For more information on how to write API docs, see [Effective Dart: Documentation](https://dart.dev/guides/language/effective-dart/documentation).

	## Before you contribute

	Before we can use your code, you must sign the [Google Individual Contributor License Agreement](https://developers.google.com/open-source/cla/individual) (CLA), which you can do online. The CLA is necessary mainly because you own the copyright to your changes, even after your contribution becomes part of our codebase, so we need your permission to use and distribute your code. We also need to be sure of various other things—for instance that you'll tell us if you know that your code infringes on other people's patents. You don't have to sign the CLA until after you've submitted your code for review and a member has approved it, but you must do it before we can put your code into our codebase.

	Before you start working on a larger contribution, you should get in touch with us first through the [Dart Issue Tracker](https://dartbug.com) with your idea so that we can help out and possibly guide you. Coordinating up front makes it much easier to avoid frustration later on.

	All submissions, including submissions by project members, require review. We use the same code-review tools and process as the chromium project.

	We occasionally take pull requests, e.g., for comment changes, but the main flow is to use the Gerrit review system as explained below.

	## Submitting patches directly from GitHub

	This repository primarily uses [Gerrit for code reviews](https://dart-review.googlesource.com/), rather than GitHub PRs. However, you may submit [a GitHub PR ](https://github.com/dart-lang/sdk/pulls), e.g. to edit some API documentation, and it will be automatically converted into a Gerrit change list (a "CL") by a copybara-service bot. You can find the link to that CL as a comment left on the PR by the bot. Any changes made to the PR after opening will also be sync'ed by the bot into the Gerrit review. The PR will be automatically closed when the CL is reviewed and landed.

	## Setting up the environment

	In order to submit a patch from a local workspace, you need to install the Chromium
	[depot\_tools](https://commondatastorage.googleapis.com/chrome-infra-docs/flat/depot_tools/docs/html/depot_tools_tutorial.html#_setting_up).

	## Getting the code

	To work with the Dart code, you need to download and build the development branch. Active development of Dart takes place on the `main` branch, from which we push "green" versions that have passed all tests to `dev` branch. Complete instructions are found at [Getting The Source](docs/Building.md#getting-the-source). You must use the `gclient` tool (`fetch`), using `git clone` will not get you a functional environment!

	## Starting a patch with git

	Create a new branch using `git new-branch` (this is a command added by the aforementioned depot_tools).

	You can be in any branch when you run `git new-branch`.

	```bash
	git new-branch <feature name>
	<write code>
	git commit
	<write code...>
	git commit
	...
	```

	## Keeping your branch updated with origin/main

	As you work, and before you send a patch for review, you should
	ensure your branch is merging cleanly to `origin/main`.

	There are multiple ways to do this, but we generally recommend
	using `git rebase-update` (another feature added by depot_tools):

	```bash
	git rebase-update
	```

	You can run this command from any branch.

	This command will fetch
	`origin/main`, rebase all your open branches, and delete
	cleanly merged branches.

	Your local workflow may vary.

	## Uploading the patch for review

	Upload the patch to Gerrit for review using `git cl upload`:

	```bash
	# Upload a CL, add reviewers from OWNERS files, and allow automatic submission.
	git cl upload --send-mail --r-owners --auto-submit
	```

	The above command returns a URL for the review. Attach this review to your issue in https://dartbug.com.

	To update the CL, just commit your changes and run `git cl upload` for your branch again.

	More detailed instructions for the `git cl` tools available on https://commondatastorage.googleapis.com/chrome-infra-docs/flat/depot_tools/docs/html/depot_tools_tutorial.html#_creating_uploading_a_cl

	## Getting your patch reviewed and landed submitted.

	Ensure that your change has at least two reviewers who are OWNERS on your CL:

	```bash
	git cl web # opens your review on https://dart-review.googlesource.com
	```

	Check the "Reviewers" on the CL, there should be at least two. If there aren't,
	use the "ADD OWNERS" button to add more. If you can't get enough reviewers, or
	the reviewers are not responsive, you may add reviewers found in the `OWNERS`
	file at the root of the repository to escalate.

	It's part of a code owner's responsibility to review changes in their area and
	help you to get your patch submitted. They may also provide comments, or reject
	your CL. They can run the presubmit tests on your behalf and submit your patch.

	## For committers: Submitting a patch

	If you have commit access, when the review is done and the patch is good to go, submit the patch on https://dart-review.googlesource.com:

	* Press "Submit to CQ" (CQ stands for "Commit Queue").
	* You can follow the progress by looking at the "Tryjobs" panel in your review.
	* Once the Commit Queue is green, the patch will be merged.
	* If any of the try jobs is red, you will have to fix the errors and then "Submit to CQ" once more.

	If you do not have commit access, a Dart engineer will commit on your behalf, assuming the patch is reviewed and accepted.

	## For committers: Merging contributions from non-members

	If the author of a patch is not a committer, they will need help landing the patch.
	Once a patch gets an LGTM, it's easy for a committer to merge it in.

	* Find and open the review on https://dart-review.googlesource.com.
	* Follow the instructions in the previous section to submit the patch.

	## Coding style

	The source code of Dart follows the:

	* [Google C++ style guide](https://google.github.io/styleguide/cppguide.html)
	* [Dart style guide](https://dart.dev/guides/language/effective-dart/style)

	You should familiarize yourself with those guidelines.

	All files in the Dart project must start with the following header. If you add a new file please also add this. The year should be a single number (not a range; don't use "2011-2012", even if the original code did). If you edit an existing file you don't have to update the year.

	```dart
	// Copyright (c) 2017, the Dart project authors. Please see the AUTHORS file
	// for details. All rights reserved. Use of this source code is governed by a
	// BSD-style license that can be found in the LICENSE file.
	```

	## AI contribution guidelines

	PRs prepared using AI tools must follow these requirements:

	1. You must review all AI-generated code before uploading it for review, and before requesting
	re-review of any updates to the uploaded change.
	* You are responsible for ensuring that code you submit meets coding standards.
	* Unmodified AI output generally does _not_ meet those standards.
	2. You must ensure that uploaded code compiles and passes at least the tests you have yourself
	added as part of the change.
	3. You must understand and be able to discuss the code you are uploading for review.
	* Non-trivial changes require discussion and iteration during review. If you do not understand the
	code, you cannot meaningfully respond to review feedback.
	* In our experience, simply feeding review feedback into an AI agent and uncritically reposting
	its output will not lead to a constructive review.
	4. You must verify the accuracy of any AI-generated text you include in the change description or
	review discussion comments.
	* If an AI provides you incorrect information, it is just hallucinating; if you choose to paste
	that text into GitHub, you are misrepresenting your change to your reviewer.
	* In particular, do not tell a reviewer that you have addressed their feedback just because
	AI output says so. It is your responsibility to make sure that review feedback has actually
	been addressed.

	### Reviewer guidelines

	Because the Dart team’s time is limited, and the capacity for people outside the team to generate
	plausible-looking code is unlimited, be mindful as a reviewer about what code you choose to review.
	Consider immediately closing changes that have any of the following red flags:

	* The change description has entirely replaced our template with AI-generated output.
	* If the contributor did not follow our process from the outset, they are unlikely to
	understand what is expected of them during review.
	* The change description does not match the changes.
	* If the contributor did not review both the changes and the description enough to notice this,
	they have not followed the AI contribution policy.
	* The change contains irrelevant AI-generated files, such as agent planning .md files.
	* If the contributor did not review the changes enough to notice and remove these files, they
	have not followed the AI contribution policy.

	As always when closing a change, explain why and provide next steps.

	As a guiding principle, if at any point in the process you feel that you are getting unfiltered
	or minimally filtered AI output as code and/or comment responses, ask yourself:

	* If this change weren’t here, would I choose to spend my time fixing this issue?
	* Would I choose to fix it using an AI agent that took hours or days to respond to every prompt?

	Unless the answer to both questions is yes, the review is not a good use of your time.

	This applies even if you are multiple rounds into the review: if the contributor closed the change,
	would you take it over using an extremely high-latency agent? Beware the sunk cost fallacy.

	### Philosophy

	A common question in discussions around AI policies for open source projects is: “Why have an
	AI policy at all? Why does it matter how the code was created; shouldn’t the code speak for itself?”

	In general, we agree, which is why our policy focuses on behaviors rather than tools. The behaviors
	that led to the creation of this policy are problematic regardless of whether they are AI-generated
	or human-generated. However, it is useful to highlight these policies in the context of AI because
	we have seen that these behaviors are orders of magnitude more common when AI is involved. For example:

	* Submitting a change containing hundreds of lines of code that the contributor doesn't understand is
	always a problem. This was rare before the widespread use of AI agents, but is more common with
	AI-assisted development.
	* Having several rounds of exchanges where a reviewer asks for a change, and the contributor says
	that they have made that change but haven't, is always a waste of reviewer time. It’s very rare
	for a contributor to deliberately and obviously lie to a reviewer, but unfortunately common
	for contributors to uncritically repeat AI agent hallucinations.
	* Changes that ignore our process have always been problematic, because standardizing the change process
	is an important part of how we keep our review load manageable. However, a contributor who has
	spent hours or days on a change is much more likely to take some time to learn and follow our process
	to avoid having that effort be wasted than someone who spent a few minutes generating the change.


	## The small print

	Contributions made by corporations are covered by a different agreement than the one above, the [Software Grant and Corporate Contributor License Agreement](https://cla.developers.google.com/about/google-corporate).

	We pledge to maintain an open and welcoming environment. For details, see our [code of conduct](https://dart.dev/code-of-conduct).