Testing Commands - RTK (Rust Token Killer)

Overview

RTK provides token-optimized test runners for multiple ecosystems. All test commands achieve 90%+ token savings by showing only failures while preserving:

Exit codes for CI/CD
Failure details (stack traces, assertions)
Test counts and duration
Full output via tee recovery

Supported Test Runners

Cargo

rtk cargo test - Rust tests via cargo

Vitest

rtk vitest - JavaScript/TypeScript via Vitest

Pytest

rtk pytest - Python tests via pytest

Go

rtk go test - Go tests with NDJSON streaming

rtk test (Generic)

Generic test runner wrapper that filters any test command.

Usage

rtk test <command>

Examples

$ cargo test
running 15 tests
test utils::test_parse ... ok
test utils::test_format ... ok
test git::test_status ... ok
test git::test_log ... ok
test grep::test_search ... ok
test ls::test_tree ... ok
test read::test_filter ... ok
test runner::test_err ... ok
test runner::test_test ... ok
test tracking::test_db ... ok
test filter::test_rust ... FAILED
test filter::test_python ... FAILED
test json::test_parse ... ok
test config::test_load ... ok
test init::test_setup ... ok

failures:

---- filter::test_rust stdout ----
thread 'filter::test_rust' panicked at 'assertion failed: `(left == right)`
  left: `42`,
 right: `43`', src/filter.rs:142:5

---- filter::test_python stdout ----
thread 'filter::test_python' panicked at 'assertion failed: expected Python filter to strip docstrings', src/filter.rs:178:5

failures:
    filter::test_rust
    filter::test_python

test result: FAILED. 13 passed; 2 failed; 0 ignored; 0 measured; 0 filtered out
# 40+ lines, ~2500 tokens

$ rtk test cargo test
✗ FAILED: 2/15 tests

filter::test_rust (src/filter.rs:142)
  assertion failed: `(left == right)`
  left: `42`, right: `43`

filter::test_python (src/filter.rs:178)
  assertion failed: expected Python filter to strip docstrings

[full output: ~/.local/share/rtk/tee/1707753600_cargo_test.log]
# 10 lines, ~250 tokens (90% savings)

Features

Failures only: Skips passing tests
Stack traces: Preserves failure details
Counts: Shows passed/failed/skipped
Tee recovery: Full output saved to file

Implementation

From src/runner.rs:run_test():

Executes command via shell
Parses stdout/stderr for test results
Filters out passing tests
Preserves exit codes

rtk cargo test

Rust test runner via cargo.

Usage

rtk cargo test [options] [testname] [-- test-options]

Options

--lib: Test library only
--bins: Test binaries
--tests: Test integration tests
--doc: Test documentation
--all-features: Enable all features
--no-fail-fast: Don’t stop on first failure

Examples

rtk cargo test              # All tests
rtk cargo test test_parse   # Specific test
rtk cargo test --lib        # Library tests only
rtk cargo test -- --nocapture  # Show stdout

Token Savings

Tests	Standard	RTK	Savings
10 passing	1,000	50	-95%
8 pass, 2 fail	2,500	250	-90%

rtk vitest

JavaScript/TypeScript test runner via Vitest.

Usage

rtk vitest [command] [options]

Commands

run: Run tests once (default)
watch: Watch mode (passthrough)
dev: Dev mode (passthrough)

Examples

$ vitest run

 RUN  v0.34.0 /path/to/project

 ✓ src/utils.test.ts (5)
 ✓ src/components/Button.test.tsx (8)
 ✗ src/hooks/useData.test.ts (3)
 ✗ src/api/client.test.ts (2)

 Test Files  4 failed | 2 passed (6)
      Tests  5 failed | 13 passed (18)
   Start at  10:00:00
   Duration  1.23s

 FAIL  src/hooks/useData.test.ts > fetches data
AssertionError: expected { status: 404 } to equal { status: 200 }
...
# 60+ lines, ~3000 tokens

$ rtk vitest run
✗ 5 failed, 13 passed (18 tests, 1.23s)

src/hooks/useData.test.ts
  ✗ fetches data
    AssertionError: expected { status: 404 } to equal { status: 200 }

src/api/client.test.ts
  ✗ handles timeout
    Error: Timeout exceeded
...

[full output: ~/.local/share/rtk/tee/1707753600_vitest.log]
# 12 lines, ~300 tokens (90% savings)

Features

JSON parsing: Parses --reporter=json output
ANSI stripping: Removes color codes
Package manager detection: Auto-detects pnpm/npm/yarn
Tee recovery: Saves full output on failure

Implementation

From src/vitest_cmd.rs:VitestParser:

Forces --reporter=json for structured output
Parses JSON line-by-line (handles pnpm prefixes)
Extracts failures with stack traces
Falls back to regex if JSON parsing fails

rtk pytest

Python test runner via pytest.

Usage

rtk pytest [options] [file::test]

Options

-v: Verbose (passthrough)
-k <pattern>: Run tests matching pattern
-x: Stop on first failure
--tb=short: Short traceback (default)

Examples

$ pytest
============================= test session starts ==============================
platform linux -- Python 3.11.0, pytest-7.4.0, pluggy-1.0.0
rootdir: /path/to/project
collected 12 items

tests/test_utils.py ........                                             [ 66%]
tests/test_api.py .F.                                                    [100%]

=================================== FAILURES ===================================
________________________________ test_fetch_data _______________________________

    def test_fetch_data():
>       assert response.status_code == 200
E       AssertionError: assert 404 == 200

tests/test_api.py:42: AssertionError
=========================== short test summary info ============================
FAILED tests/test_api.py::test_fetch_data - AssertionError: assert 404 == 200
========================= 1 failed, 11 passed in 1.23s =========================
# 25+ lines, ~800 tokens

$ rtk pytest
✗ 1 failed, 11 passed (1.23s)

tests/test_api.py::test_fetch_data
  AssertionError: assert 404 == 200
  at tests/test_api.py:42

[full output: ~/.local/share/rtk/tee/1707753600_pytest.log]
# 7 lines, ~80 tokens (90% savings)

Features

State machine parser: Parses text output (no JSON needed)
Auto-flags: Forces --tb=short -q for compact output
Import errors: Preserves stderr for import failures
Tee recovery: Saves full output on failure

Implementation

From src/pytest_cmd.rs:filter_pytest_output():

State machine: Header → TestProgress → Failures → Summary
Extracts failure sections with traceback
Filters out passing tests (. dots)
Preserves exit codes

rtk go test

Go test runner with NDJSON streaming.

Usage

rtk go test [packages] [flags]

Flags

-v: Verbose (passthrough)
-run <regexp>: Run tests matching pattern
-short: Run short tests only
-timeout <duration>: Test timeout

Examples

$ go test ./...
=== RUN   TestParseConfig
--- PASS: TestParseConfig (0.00s)
=== RUN   TestFilterData
--- PASS: TestFilterData (0.00s)
=== RUN   TestFetchUser
--- FAIL: TestFetchUser (0.02s)
    api_test.go:42: Expected status 200, got 404
=== RUN   TestValidateInput
--- PASS: TestValidateInput (0.00s)
FAIL
FAIL    github.com/user/project/pkg/api    0.023s
ok      github.com/user/project/pkg/utils  0.008s
# 15+ lines, ~600 tokens

$ rtk go test ./...
✗ 1 failed, 3 passed (0.02s)

github.com/user/project/pkg/api
  TestFetchUser
    api_test.go:42: Expected status 200, got 404

[full output: ~/.local/share/rtk/tee/1707753600_go_test.log]
# 7 lines, ~60 tokens (90% savings)

Features

NDJSON parsing: Parses -json output line-by-line
Streaming: Handles interleaved package events
Build errors: Captures build failures separately
Tee recovery: Saves full output on failure

Implementation

From src/go_cmd.rs:filter_go_test_json():

Forces -json flag for structured output
Parses GoTestEvent NDJSON structs
Groups by package (handles multiple packages)
Tracks pass/fail/skip counts per package
Preserves exit codes

Tee Output Recovery

All test commands use tee recovery to save full output on failure.

How It Works

Test runs and produces output
RTK filters to failures only
If exit code != 0, saves raw output to ~/.local/share/rtk/tee/
Prints hint: [full output: ~/.../1707753600_cargo_test.log]
LLM reads file instead of re-running test

Configuration

From ~/.config/rtk/config.toml:

[tee]
enabled = true          # default: true
mode = "failures"       # "failures" (default), "always", or "never"
max_files = 20          # max files to keep (oldest rotated out)
max_file_size = 1048576 # 1MB per file max

Environment Overrides

RTK_TEE=0 rtk cargo test           # Disable tee
RTK_TEE_DIR=/tmp/tee rtk pytest    # Custom directory

CI/CD Integration

All test commands preserve exit codes:

# .github/workflows/ci.yml
name: CI

on: [push, pull_request]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Run tests
        run: |
          rtk cargo test --all-features
          rtk pytest tests/
          rtk go test ./...
          rtk vitest run

Tests fail the CI pipeline if any test fails (exit code preserved).

Token Savings Summary

Test Runner	Standard Tokens	RTK Tokens	Savings
cargo test (15 tests, 2 fail)	2,500	250	-90%
vitest (18 tests, 5 fail)	3,000	300	-90%
pytest (12 tests, 1 fail)	800	80	-90%
go test (10 tests, 1 fail)	600	60	-90%

​Overview

​Supported Test Runners

Cargo

Vitest

Pytest

Go

​rtk test (Generic)

​Usage

​Examples

​Features

​Implementation

​rtk cargo test

​Usage

​Options

​Examples

​Token Savings

​rtk vitest

​Usage

​Commands

​Examples

​Features

​Implementation

​rtk pytest

​Usage

​Options

​Examples

​Features

​Implementation

​rtk go test

​Usage

​Flags

​Examples

​Features

​Implementation

​Tee Output Recovery

​How It Works

​Configuration

​Environment Overrides

​CI/CD Integration

​Token Savings Summary

​Next Steps

JavaScript

Python

Overview

Supported Test Runners

rtk test (Generic)

Usage

Examples

Features

Implementation

rtk cargo test

Usage

Options

Examples

Token Savings

rtk vitest

Usage

Commands

Examples

Features

Implementation

rtk pytest

Usage

Options

Examples

Features

Implementation

rtk go test

Usage

Flags

Examples

Features

Implementation

Tee Output Recovery

How It Works

Configuration

Environment Overrides

CI/CD Integration

Token Savings Summary

Next Steps