logos-blockchain-pocs/mistral/blockscout/CONTRIBUTING.md

Contribution Guidelines: What We Are Looking For

We welcome contributions that enhance the project and improve the overall quality of our codebase. While we appreciate the effort that goes into making contributions, we kindly ask that contributors focus on the following types of changes:

• Feature Enhancements: Substantial improvements or new features that add significant value to the project.
• Bug Fixes: Fixes for known bugs or issues that impact functionality.
• Documentation Improvements: Comprehensive updates to documentation that clarify usage, installation, or project structure.
• Performance Improvements: Changes that enhance the performance or efficiency of the application.

Contributing

1. Fork it ( https://github.com/blockscout/blockscout/fork )
2. Create your feature branch (git checkout -b my-new-feature)
3. Write tests that cover your work
4. Commit your changes (git commit -am 'Add some feature')
5. Push to the branch (git push origin my-new-feature)
6. Create a new Pull Request. The title of Pull Request should follow Conventional Commits specification and should start with feat:, fix:, chore:, doc:, perf:, refactor: prefix.

General

• Commits should be one logical change that still allows all tests to pass. Prefer smaller commits if there could be two levels of logic grouping. The goal is to allow contributors in the future (including your own future self) to determine your reasoning for making changes and to allow them to cherry-pick, patch or port those changes in isolation to other branches or forks.
• If during your PR you reveal a pre-existing bug:
  1. Try to isolate the bug and fix it on an independent branch and PR it first.
  2. Try to fix the bug in a separate commit from other changes:
     1. Commit the code in the broken state that revealed the bug originally
     2. Commit the fix for the bug.
     3. Continue original PR work.

Enhancements

Enhancements cover all changes that make users lives better:

• feature requests filed as issues that impact end-user contributors and developers
• changes to the architecture that make it easier for contributors (in the GitHub sense), dev-ops, and deployers to maintain and run blockscout

Bug Fixes

For bug fixes, whenever possible, there should be at least 2 commits:

1. A regression test commit that contains tests that demonstrate the bug and show as failing.
2. The bug fix commit that shows the regression test now passing.

This format ensures that we can run the test to reproduce the original bug without depending on the new code in the fix, which could lead to the test falsely passing.

Incompatible Changes

Incompatible changes can arise as a side-effect of either Enhancements or Bug Fixes. During Enhancements, incompatible changes can occur because, as an example, in order to support showing end-users new data, the database schema may need to be changed and the index rebuilt from scratch. During bug fixes, incompatible changes can occur because in order to fix a bug, the schema had to change, or how certain internal APIs are called changed.

• Incompatible changes should be called out explicitly, with any steps the various user roles need to do to upgrade.

• If a schema change occurs that requires a re-index add the following to the Pull Request description:

  **NOTE**: A database reset and re-index is required
  

Pull Request

There is a PULL_REQUEST_TEMPLATE.md for this repository, but since it can't fill in the title for you, please follow the following steps when opening a Pull Request before filling in the template:

• Title
  • Prefix labels if you don't have permissions to set labels in the GitHub interface.
  • (bug) for bug fixes
  • (enhancement) for enhancements
  • (incompatible changes) for incompatible changes, such a refactor that removes functionality, changes arguments, or makes something required that wasn't previously.
  • Single sentence summary of change
    • What was fixed for bugs
    • What was added for enhancements
    • What was changed for incompatible changes

See #255 as an example PR that uses GitHub keywords and a Changelog to explain multiple changes.

Basic Naming Convention

When contributing to the codebase, please adhere to the following naming conventions to ensure clarity and consistency:

• Use full names for entities. Avoid abbreviations or shorthand.
  • Instead of "tx" or "txn", use "transaction".
  • Instead of "txs", use "transactions".
  • Instead of "tx_hash" or "txn_hash", use "transaction_hash".
  • Instead of "block_num", use "block_number".
• Ensure that variable names are descriptive and convey the purpose or content clearly.
• Consistent naming helps in maintaining readability and understanding of the code, especially for new contributors.

By following these conventions, we can maintain a clean and understandable codebase.

API V2 Naming Convention

When contributing to the API v2, please adhere to the following naming conventions for response fields to ensure clarity and consistency:

• The block number should be returned as a number in the property with the name which ends with block_number.
• All hashes (transaction, block address etc.) should be returned as a hex string in the property which ends with _hash.
• Property name for aggregations like counts and sums should contain plural form of entity and _count, _sum suffix respecively, e.g. transactions_count, blocks_count, withdrawals_sum.
• All fields that contain the "index" suffix should be returned as numbers.

Environment Configuration Best Practices

Runtime vs. Compile-time Configuration

We strongly favor runtime configuration over compile-time configuration whenever possible. This approach:

• Reduces the number of Docker images needed
• Increases deployment flexibility
• Simplifies maintenance and testing

When adding new configuration options, chain types, or refactoring existing ones, please follow the decision tree below to determine the appropriate approach:

flowchart TD
    A[Add/Modify Configuration Option or Chain Type] --> B{Is it feature-specific behavior of a function?}
    B -->|Yes| C[Use RuntimeEnvHelper or Application.get_env/3 and pattern matching]
    B -->|No| D{Does it need new database tables?}
    D -->|Yes| E[Create new Ecto.Repo and handle it at runtime in config_helper.ex]
    D -->|No| F{Is it an API endpoint?}
    F -->|Yes| G[Use chain_scope macro or CheckFeature plug]
    F -->|No| H{Does it modify existing database schema?}
    H -->|Yes| I[Use Compile-time configuration]
    H -->|No| J[Contact us to discuss this case further]
    I -->|Future Work| O[Refactor toward Runtime configuration]

Use runtime configuration and pattern matching

Anti-pattern:

# AVOID THIS
use Utils.CompileTimeEnvHelper,
  chain_type: [:explorer, :chain_type]

if @chain_type == :optimism do
  def foo, do: :bar
else
  def foo, do: :baz
end

Better approach:

# DO THIS INSTEAD
use Utils.RuntimeEnvHelper,
  chain_type: [:explorer, :chain_type]

def foo, do: chain_type() |> do_foo()

defp do_foo(:optimism), do: :bar
defp do_foo(_), do: :baz

New database tables

If your feature or chain-specific functionality requires new database tables:

1. Define a new repository module in apps/explorer/lib/explorer/repo.ex.
2. Add the repository to config/config_helper.exs in the repos/0 function.
3. Include a runtime check to load this repo conditionally:

# In config_helper.ex
ext_repos = [
  {parse_bool_env_var("MY_FEATURE_ENABLED"), Explorer.Repo.MyFeature},
  # other feature repos...
]
|> Enum.filter(&elem(&1, 0))
|> Enum.map(&elem(&1, 1))

This approach ensures migrations are automatically detected and applied at runtime without requiring recompilation.

API endpoints

For feature-specific or chain-specific API endpoints, use one of the following runtime approaches:

1. For chain-specific routes, use the chain_scope macro in your router:

scope "/v2", as: :api_v2 do
  chain_scope :polygon_zkevm do
    get("/zkevm-batch/:batch_number", V2.TransactionController, :polygon_zkevm_batch)
  end
end

2. For feature-toggle endpoints, use CheckFeature plug in pipelines:

pipeline :my_feature do
  plug(BlockScoutWeb.Plug.CheckFeature, feature_check: &my_feature_enabled?/0)
end

scope "/my-feature" do
  pipe_through(:my_feature)
  
  get "/data", MyFeatureController, :index
end

Both approaches return appropriate 404 responses when the feature is disabled or chain type doesn't match.

Modifying existing database schema

If your functionality requires modifying existing database schema structures (adding columns to shared tables, changing constraints, etc.), you currently must use compile-time configuration. This is the only case where compile-time configuration is still recommended.

# Current approach for schema modifications
use Utils.CompileTimeEnvHelper,
  chain_type: [:explorer, :chain_type]

if @chain_type == :optimism do
  # Schema modifications specific to Optimism
end

To prepare for future runtime refactoring, isolate these schema-specific changes as much as possible.

This limitation stems from Ecto schemas being defined at compile-time. When different chain types need variations in shared tables (additional fields, different constraints), these schema differences cannot be modified at runtime. We're currently researching approaches for dynamic schema adjustment based on runtime configuration.

For reference on which chain types still require compile-time configuration, see the Chain-Specific Environment Variables documentation.

Compile time Environment Variables

Before using compile-time configuration, ensure you've exhausted all runtime alternatives by following the decision tree above. If after careful consideration you still need to work with compile-time environment variables, follow these guidelines:

• Always use the Utils.CompileTimeEnvHelper module instead of direct Application.compile_env/2 calls:

# DO use this approach
use Utils.CompileTimeEnvHelper,
    attribute_name: [:app, :test]

# Access the value using the module attribute
@attribute_name

# DON'T use this approach
Application.compile_env(:app, :test)  # avoid direct compile_env calls

This approach provides faster compilation time and simplifies development and maintenance.