Skip to content

style(e2e): remove commit hashes from contracts-remove.sh #1523

style(e2e): remove commit hashes from contracts-remove.sh

style(e2e): remove commit hashes from contracts-remove.sh #1523

Workflow file for this run

name: Codium AI
on:
pull_request:
types: [opened, reopened, ready_for_review]
paths-ignore:
- ".github/workflows/deploy.yml"
- ".github/workflows/docs-release.yml"
- ".github/workflows/outdated.yml"
- ".github/workflows/comment-tag-report.yml"
- ".github/workflows/pr-agent.yml"
- ".github/CODEOWNERS"
- "config/**"
- "README.md"
- "LICENSE"
- "CONTRIBUTING.md"
- "utils/slack-notifiers/**"
issue_comment:
workflow_dispatch:
env:
EXTRA_INSTRUCTIONS: |
These rules apply mostly to Rust code:
* Most part of the codebase is written in Rust and formatted and analysed with clippy.
* Do not waste time on trivial things like null handling or formatting because the compiler and clippy can handle these things.
* When you detect a task spawned to background with tokio or threads, ensure to review the operations are properly logged and monitored.
* When you detect the use of tokio::spawn or tokio::spawn_blocking, suggest using spawn_named or spawn_blocking_named functions.
* When you detect the use of tokio::sleep or thread::sleep, suggest to use traced_sleep, so we can have visibility in tracing events.
* When a long-running blocking task is running inside Tokio async context, suggest using spawn_blocking_named to avoid Tokio thread locking.
* When a function is instrumented with tracing::instrument, suggest to record identifiers as span fields if none are present.
* Never suggest to use log crate. We use structured tracing here.
* When you detect a tracing event, do not allow the message to be formatted with dynamic fields. Instead suggest to add these dynamic fields as tracing fields.
* When an error is logged, ensure the original error is logged in a field called reason.
* When you detect an unwrap, suggest to use expect.
* When you detect a core component or service is being initialized, ensure its initialization is being logged with all relevant configurations.
* When too much cloning is happening, suggest to use references if possible.
* When you detect an infinite loop, ensure that tracing events inside the loop use `parent: None` to avoid unbounded Vec growth resulting in memory leak. Example: `tracing::info!(parent: None, "some message");`
jobs:
pr_agent_job:
if: ${{ github.event.sender.type != 'Bot' }}
runs-on: ubuntu-latest
permissions:
issues: write
pull-requests: write
contents: write
name: PR Agent
steps:
- name: Review
id: pr_agent
uses: Codium-ai/pr-agent@main
env:
OPENAI_KEY: ${{ secrets.OPENAI_KEY }}
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
github_action_config.auto_review: "true"
github_action_config.auto_improve: "true"
github_action_config.auto_describe: "true"
pr_description.publish_labels: "false"
github_action_config.enable_output: "false"
pr_reviewer.enable_review_labels_effort: "false"
pr_reviewer.enable_review_labels_security: "false"
pr_reviewer.extra_instructions: "${{ env.EXTRA_INSTRUCTIONS }}"
config.model: "anthropic/claude-3-5-sonnet-20240620"
config.model_turbo: "anthropic/claude-3-5-sonnet-20240620"
anthropic.key: ${{ secrets.ANTHROPIC_KEY }}