This repo is primarily maintained by Konstantin Gredeskoul and Yuki "Yugui" Sonoda.
We are both very busy and would really love more contributors to join the core team.
If you are interested in developing Ruby Rules for Bazel, please submit a couple of PRs and then lets talk!
Tip
|
You can read or print this README in a proper PDF format by grabbing our README.pdf. |
Readiness | Types of Applications |
---|---|
ruby apps, ruby gems, micro-services, ideally in a mono-repo |
|
medium-sized Ruby on Rails apps, ideally in a mono-repo |
|
complex Ruby on Rails monoliths, single-repo |
Note
|
we have a short guide on Building your first Ruby Project on the Wiki. We encourage you to check it out. |
workspace(name = "my_ruby_project")
load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
load("@bazel_tools//tools/build_defs/repo:git.bzl", "git_repository")
#———————————————————————————————————————————————————————————————————————
# To get the latest ruby rules, grab the 'master' branch.
#———————————————————————————————————————————————————————————————————————
git_repository(
name = "bazelruby_rules_ruby",
remote = "https://github.com/bazelruby/rules_ruby.git",
branch = "master"
)
load(
"@bazelruby_rules_ruby//ruby:deps.bzl",
"rules_ruby_dependencies",
"rules_ruby_select_sdk",
)
rules_ruby_dependencies()
#———————————————————————————————————————————————————————————————————————
# Specify Ruby version — this will either build Ruby or use a local
# RBENV installation if the Ruby version matches.
#———————————————————————————————————————————————————————————————————————
load("@bazel_skylib//:workspace.bzl", "bazel_skylib_workspace")
bazel_skylib_workspace()
rules_ruby_select_sdk(version = "3.0.2")
#———————————————————————————————————————————————————————————————————————
# Now, load the ruby_bundle rule & install gems specified in the Gemfile
#———————————————————————————————————————————————————————————————————————
load(
"@bazelruby_rules_ruby//ruby:defs.bzl",
"ruby_bundle",
)
ruby_bundle(
name = "bundle",
# Specify additional paths to be loaded from the gems at runtime, if any.
# Since spec.require_paths in Gem specifications are auto-included, directory paths
# in spec.require_paths do not need to be listed in includes hash.
includes = {
"grpc": ["etc"],
},
excludes = {
"mini_portile": ["test/**/*"],
},
gemfile = "//:Gemfile",
gemfile_lock = "//:Gemfile.lock",
)
# You can specify more than one bundle in the WORKSPACE file
ruby_bundle(
name = "bundle_app_shopping",
gemfile = "//:apps/shopping/Gemfile",
gemfile_lock = "//:apps/shopping/Gemfile.lock",
)
# You can also install from Gemfile using `gemspec`.
ruby_bundle(
name = "bundle_gemspec",
srcs = ["//:lib/my_gem/my_gem.gemspec"],
gemfile = "//:lib/my_gem/Gemfile",
gemfile_lock = "//:lib/my_gem/Gemfile.lock",
)
Any of the project BUILD
files can now reference any gems included in the Gemfile
referenced by the ruby_bundle
rule, and defined in the project’s WORKSPACE
file.
Add ruby_library
, ruby_binary
, ruby_rspec
or ruby_test
into your BUILD.bazel
files.
#———————————————————————————————————————————————————————————————————————
# Define Ruby executable, test, spec and package a gem
#———————————————————————————————————————————————————————————————————————
load(
"@bazelruby_rules_ruby//ruby:defs.bzl",
"ruby_binary",
"ruby_library",
"ruby_test",
"ruby_rspec",
)
ruby_library(
name = "foo",
srcs = glob(["lib/**/*.rb"]),
includes = ["lib"],
deps = [
"@bundle//:activesupport",
"@bundle//:awesome_print",
"@bundle//:rubocop",
]
)
ruby_binary(
name = "bar",
srcs = ["bin/bar"],
deps = [":foo"],
)
ruby_test(
name = "foo-test",
srcs = ["test/foo_test.rb"],
deps = [":foo"],
)
ruby_rspec(
name = "foo-spec",
specs = glob(["spec/**/*.rb"]),
rspec_args = { "--format": "progress" },
deps = [":foo"]
}
Use ruby_gem
rule to package any number of ruby files or folders into a Ruby-Gem compatible ZIP archive.
load(
"@bazelruby_rules_ruby//ruby:defs.bzl",
"ruby_gem",
)
ruby_gem(
name = "awesome-sauce-gem", # name of the build target
gem_name = "awesome-sauce", # name of the gem
gem_version = "0.1.0",
gem_summary = "Example gem to demonstrate Bazel Gem packaging",
gem_description = "Example gem to demonstrate Bazel Gem packaging",
gem_homepage = "https://github.com/bazelruby/rules_ruby",
gem_authors = [
"BazelRuby",
"Konstantin Gredeskoul"
],
gem_author_emails = [
"[email protected]",
],
gem_runtime_dependencies = {
"colored2": "~> 3.1.2",
"hashie": "",
},
gem_development_dependencies = {
"rspec": "",
"rspec-its": "",
"rubocop": "",
},
srcs = [
glob("{bin,exe,lib,spec}/**/*.rb")
],
deps = [
"//lib:example_gem",
],
)
If you are using ASDF to manage your ruby installs, you can use them by adding .bazelrc
:
build --test_env=ASDF_DIR --test_env=ASDF_DATA_DIR build --action_env=ASDF_DIR --test_env=ASDF_DATA_DIR
You will have to be sure to export the ASDF_DATA_DIR
in your profile since it’s not set by default. e.g. export ASDF_DATA_DIR="$HOME/.asdf"
ruby_library(
name,
deps,
srcs,
data,
compatible_with,
deprecation,
distribs,
features,
licenses,
restricted_to,
tags,
testonly,
toolchains,
visibility)
Attributes | |||
---|---|---|---|
|
A unique name for this rule. |
||
|
List of At least |
||
|
List of targets that are required by the At least |
||
|
List of paths to be added to |
||
|
List of options to be passed to the Ruby interpreter at runtime.
|
||
And other common attributes. |
ruby_binary(
name,
deps,
srcs,
data,
main,
compatible_with,
deprecation,
distribs,
features,
licenses,
restricted_to,
tags,
testonly,
toolchains,
visibility,
args,
output_licenses
)
Attributes | |||
---|---|---|---|
|
A unique name for this rule. |
||
|
List of |
||
|
List of targets that are required by the |
||
|
The entrypoint file. It must be also in If not specified, |
||
|
List of paths to be added to |
||
|
List of options to be passed to the Ruby interpreter at runtime.
|
||
And other common attributes. |
ruby_test(
name,
deps,
srcs,
data,
main,
compatible_with,
deprecation,
distribs,
features,
licenses,
restricted_to,
tags,
testonly,
toolchains,
visibility,
args,
size,
timeout,
flaky,
local,
shard_count
)
Attributes | |||
---|---|---|---|
|
A unique name for this rule. |
||
|
List of |
||
|
List of targets that are required by the |
||
|
The entrypoint file. It must be also in If not specified, |
||
|
List of paths to be added to |
||
|
List of options to be passed to the Ruby interpreter at runtime.
|
||
And other common attributes. |
NOTE: This is a repository rule, and can only be used in a WORKSPACE
file.
This rule installs gems defined in a Gemfile using Bundler, and exports individual gems from the bundle, as well as the entire bundle, available as a ruby_library
that can be depended upon from other targets.
ruby_bundle(
name,
gemfile,
gemfile_lock,
bundler_version = "2.1.4",
includes = {},
excludes = {},
srcs = [],
vendor_cache = False,
ruby_sdk = "@org_ruby_lang_ruby_toolchain",
ruby_interpreter = "@org_ruby_lang_ruby_toolchain//:ruby",
)
Attributes | |||
---|---|---|---|
|
A unique name for this rule. |
||
|
The |
||
|
The
|
||
|
List of additional files required for Bundler to install gems. This could usually include |
||
|
Symlink the vendor directory into the Bazel build space, this allows Bundler to access vendored Gems |
||
|
The Version of Bundler to use. Defaults to 2.1.4.
|
||
|
List of glob patterns per gem to be additionally loaded from the library. Keys are the names of the gems which require some file/directory paths not listed in the |
||
|
List of glob patterns per gem to be excluded from the library. Keys are the names of the gems. Values are lists of blob path patterns, which are relative to the root directories of the gems. The default value is |
||
And other common attributes. |
ruby_bundle
creates several targets that can be used downstream. In the examples below we assume that your ruby_bundle
has a name app_bundle
:
-
@app_bundle//:bundler
— references just the Bundler from the bundle. -
@app_bundle//:gems
— references all gems in the bundle (i.e. "the entire bundle"). -
@app_bundle//:gem-name
— references just the specified gem in the bundle, eg.@app_bundle//:awesome_print
. -
@app_bundle//:bin
— references to all installed executables from this bundle, with individual executables accessible via eg.@app_bundle//:bin/rubocop
load("@bazelruby_rules_ruby//ruby:defs.bzl", "ruby_bundle")
ruby_bundle(
name = "gems",
bundler_version = '2.1.4',
gemfile = "//:Gemfile",
gemfile_lock = "//:Gemfile.lock",
)
To use the vendor cache, you have to declare a managed_directory
in
your workspace. The name should match the name of the bundle.
load("@bazelruby_rules_ruby//ruby:defs.bzl", "ruby_bundle")
workspace(
name = "my_wksp",
managed_directories = {"@bundle": ["vendor"]},
)
ruby_bundle(
name = "bundle",
bundler_version = "2.1.2",
vendor_cache = True,
gemfile = "//:Gemfile",
gemfile_lock = "//:Gemfile.lock",
)
# Reference the entire bundle with :gems
ruby_library(
name = "foo",
srcs = ["foo.rb"],
deps = ["@gems//:gems"],
)
# Or, reference specific gems from the bundle like so:
ruby_binary(
name = "rubocop",
srcs = [":foo", ".rubocop.yml"],
args = ["-P", "-D", "-c" ".rubocop.yml"],
main = "@gems//:bin/rubocop",
deps = ["@gems//:rubocop"],
)
ruby_rspec(
name,
deps,
srcs,
data,
main,
rspec_args,
bundle,
compatible_with,
deprecation,
distribs,
features,
licenses,
restricted_to,
tags,
testonly,
toolchains,
visibility,
args,
size,
timeout,
flaky,
local,
shard_count
)
Attributes | |||
---|---|---|---|
|
A unique name for this rule. |
||
|
List of |
||
|
List of targets that are required by the |
||
|
The entrypoint file. It must be also in If not specified, |
||
|
Command line arguments to the If not specified, the default arguments defined in |
||
|
List of paths to be added to |
||
|
List of options to be passed to the Ruby interpreter at runtime.
|
||
And other common attributes. |
Used to generate a zipped gem containing its srcs, dependencies and a gemspec.
ruby_gem(
name,
gem_name,
gem_version,
gem_summary,
gem_description,
gem_homepage,
gem_authors,
gem_author_emails,
gem_runtime_dependencies,
gem_development_dependencies,
require_paths = ["lib"],
srcs = srcs,
deps = deps,
data = data
)
Attributes | |
---|---|
|
A unique name for this build target. |
|
The name of the gem to be generated. |
|
The version of the gem. Is used to name the output file, which becomes |
|
One line summary of the gem purpose. |
|
Single-line, paragraph-sized description text for the gem. |
|
Homepage URL of the gem. |
|
List of human readable names of the gem authors. Required to generate a valid gemspec. |
|
List of email addresses of the authors. |
|
List of At least |
|
List of targets that are required by the At least |
|
List of paths to be added to the Ruby LOAD_PATH when using this gem. Typically this value is just |
|
This is a dictionary where keys are gem names, and values are either an empty string or a gem version specification. For instance, the pessimistic version specifier |
|
Similar to the above, this specifies gems necessary for the development of the above gem, such as testing gems, linters, code coverage and more. |
And other common attributes. |
[check square] Using various versions of Ruby installed locally
[square] Building native extensions in gems with Bazel
[square] Releasing your gems with Bazel (Coinbase fork might have this feature, worth checking)
We welcome contributions to RulesRuby. Please make yourself familiar with the code of conduct, which basically says — don’t be an a-hole.
You may notice that there is more than one Bazel WORKSPACE inside this repo. There is one in examples/simple_script
for instance, because
we use this example to validate and test the rules. So be mindful whether your current directory contains WORKSPACE
file or not.
You will need Homebrew installed prior to running the script.
After that, cd into the top level folder and run the setup script in your Terminal:
❯ bin/setup
This runs a complete setup, shouldn’t take too long. You can explore various script options with the help
command:
❯ bin/setup -h
USAGE
# without any arguments runs a complete setup.
bin/setup
# alternatively, a sub-setup function name can be passed:
bin/setup [ gems | git-hook | help | main | os-specific | rbenv | remove-git-hook ]
DESCRIPTION:
Runs full setup without any arguments.
Accepts one optional argument — one of the actions that typically run
as part of setup, with one exception — remove-git-hook.
This action removes the git commit hook installed by the setup.
EXAMPLES:
bin/setup
Or, to run only one of the sub-functions (actions), pass
it as an argument:
bin/setup help
bin/setup remove-git-hook
We provided a handy script bin/show-env
to display where your dependencies are coming from. Here is an example of running it on a Mac OS-X system:
❯ bin/show-env
Besides making yourself familiar with the existing code, and Bazel documentation on writing rules, you might want to follow this order:
bin/test-suite
OR, you can run individual Bazel test commands from the inside.
-
bazel test //…
-
cd examples/simple_script && bazel test //…
-
Open a pull request in Github, and please be as verbose as possible in your description.
-
In general, it’s always a good idea to ask questions first — you can do so by creating an issue.
After running setup, and since this is a bazel repo you can use Bazel commands:
bazel build //...:all
bazel query //...:all
bazel test //...:all
But to run tests inside each sub-WORKSPACE, you will need to repeat that in each sub-folder. Luckily, there is a better way.
This script runs all tests (including sub-workspaces) when ran without arguments:
bin/test-suite
Run it with help
command to see other options, and to see what parts you can run individually. At the moment they are:
# alternatively, a partial test name can be passed:
bin/test-suite [ all | bazel-info | buildifier | help | rspec | rubocop | simple-script | workspace ]
On a MacBook Pro it takes about 3 minutes to run.
We are using RuboCop for ruby and Buildifier for Bazel. Both are represented by a single script bin/linter
, which just like the scripts above runs ALL linters when ran without arguments, accepts help
commnd, and can be run on a subset of linting strategies:
bin/linter
The following are the partial linting functions you can run:
# alternatively, a partial linter name can be passed:
bin/linter [ all | buildifier | help | rubocop ]
To regenerate, first you may need to grab an API token and export the GITHUB_TOKEN
variable:
export GITHUB_TOKEN=....
Then use the make
target:
make update
Or, manually:
gem install github_changelog_generator
github_changelog_generator -u bazelruby -p rules_ruby -t your-github-token
© 2018-2021 BazelRuby Contributors.
Core Team:
Core Team (Emeritus):
Licensed under the Apache License, Version 2.0 (the "License").
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.