Both depend on test passing but run independently of each other.
2. Built-in parallel mode
Add a --parallel flag or multi-reviewer config so a single invocation dispatches both reviews concurrently using goroutines:
review-bot --config reviewers.yaml
Where reviewers.yaml defines multiple reviewer profiles (model, token, name).
Benefits
Cuts wall-clock time in half
Reviews post roughly simultaneously
Better resource utilization
Considerations
Approach 1 is simpler but requires building the binary twice (or downloading from release)
Approach 2 is more elegant but needs careful error handling (one reviewer failing should not block the other)
Both approaches need separate Gitea tokens per reviewer to avoid auth conflicts
## Problem
Currently the CI workflow runs Sonnet and GPT reviews sequentially in the same job. Total review time is additive (30-60s per model = 60-120s total).
## Proposed Solution
Two approaches (not mutually exclusive):
### 1. Parallel CI jobs
Split into separate jobs that run concurrently:
```yaml
jobs:
review-sonnet:
runs-on: ubuntu-24.04
...
review-gpt:
runs-on: ubuntu-24.04
...
```
Both depend on `test` passing but run independently of each other.
### 2. Built-in parallel mode
Add a `--parallel` flag or multi-reviewer config so a single invocation dispatches both reviews concurrently using goroutines:
```bash
review-bot --config reviewers.yaml
```
Where `reviewers.yaml` defines multiple reviewer profiles (model, token, name).
## Benefits
- Cuts wall-clock time in half
- Reviews post roughly simultaneously
- Better resource utilization
## Considerations
- Approach 1 is simpler but requires building the binary twice (or downloading from release)
- Approach 2 is more elegant but needs careful error handling (one reviewer failing should not block the other)
- Both approaches need separate Gitea tokens per reviewer to avoid auth conflicts
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Problem
Currently the CI workflow runs Sonnet and GPT reviews sequentially in the same job. Total review time is additive (30-60s per model = 60-120s total).
Proposed Solution
Two approaches (not mutually exclusive):
1. Parallel CI jobs
Split into separate jobs that run concurrently:
Both depend on
testpassing but run independently of each other.2. Built-in parallel mode
Add a
--parallelflag or multi-reviewer config so a single invocation dispatches both reviews concurrently using goroutines:Where
reviewers.yamldefines multiple reviewer profiles (model, token, name).Benefits
Considerations
rodin referenced this issue2026-05-12 16:59:57 +00:00