Skip to content

Commit

Permalink
Add predicate capturing results of test runs
Browse files Browse the repository at this point in the history
Signed-off-by: Aditya Sirish <[email protected]>
  • Loading branch information
adityasaky committed Mar 30, 2023
1 parent 5fa28eb commit 3ee93c9
Showing 1 changed file with 172 additions and 0 deletions.
172 changes: 172 additions & 0 deletions spec/predicates/test-results.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,172 @@
# Predicate type: Test Results

Type URI: https://in-toto.io/attestation/test-result/v0.1

Version: 0.1.0

Authors:

## Purpose

This predicate type defines a generic schema to express the result of running
tests in software supply chains. The schema may be extended to support different
types of testing or specific test harnesses.

## Use Cases

Software development processes include several types of tests. This attestation
can be used to express the results of running those tests. It can be used to
verify:

1. that all tests were in fact run, and
2. that all required tests passed

Therefore, each attestation corresponds to one invocation of a test suite, and
may include the results of several individual tests.

### Asserting Test Configurations Used

The supply chain owner creates a policy that records the expected test
configurations. During verification, the policy checks that the test attestation
used the right configurations. A custom inspection may optionally parse the
`testRun.link` field to verify attestation matches the test run.

### Asserting Test Results

In addition to the previous use case, the supply chain owner creates a policy
verifying that test results passed. In the simplest case, the policy applies to
all tests. Therefore, it asserts the contents of `result` and that
`testRun.failedTests` is empty. In more nuanced cases, a subset of tests may
matter. For example, the tested artifact may be an OS image that's to be
deployed to three types of devices. As such, the test harness tests the new
image on an instance of each device. When verifying the attestation prior to an
image being installed on a device, it only matters that the tests passed on the
corresponding test device and not necessarily the others.

As before, a custom inspection may examine the `testRun.link` contents to verify
the contents of the attestation.

## Prerequisites

Understanding of the
[in-toto attestation specification](https:/in-toto/attestation).

## Model

This predicate type includes one compulsory field, `result`, that describes the
result of the test run. The `testRun` object can be used to communicate a link
to the test run and list tests that passed, failed, and passed with warnings. It
also may optionally contain a reference to the configuration for the test run.
The expected `subject` for this are the source artifacts themselves or some
reference to them such as a Git commit.

## Schema

```json
{
"_type": "https://in-toto.io/Statement/v1",
"subject": [{...}],
"predicateType": "https://in-toto.io/attestation/test-result/v0.1",
"predicate": {
"result": "pass|fail",
"testRun": {
"link": "<URL>",
"configuration": ["<ResourceDescriptor>", ...],
"passedTests": ["<TEST_NAME>", ...],
"warnedTests": ["<TEST_NAME>", ...],
"failedTests": ["<TEST_NAME>", ...]
}
}
}
```

### Parsing Rules

This predicate follows the in-toto attestation parsing rules. Summary:

- Consumers MUST ignore unrecognized fields.
- The `predicateType` URI includes the major version number and will always
change whenever there is a backwards incompatible change.
- Minor version changes are always backwards compatible and “monotonic.” Such
changes do not update the `predicateType`.
- Producers MAY add extension fields using field names that are unlikely to
collide with names used by other producers. Field names SHOULD avoid using
characters like `.` and `$`.
- Fields marked _optional_ MAY be unset or null, and should be treated
equivalently. Both are equivalent to empty for object or array values.

### Fields

`result` _boolean_ , _required_

Indicates the result of the test run. If true, it indicates _all_ tests passed
in the corresponding run.

`testRun` _object_, _optional_

`testRun.link` _URI_, _optional_

Contains a link to the test run. Useful to point to information that cannot be
captured in the attestation.

`testRun.configuration` _list of ResourceDescriptor_, _required_

Reference to the configuration used for the test run.

`testRun.passedTests` _list_, _optional_

Each entry corresponds to a single test that passed.

`testRun.warnedTests` _list_, _optional_

Each entry corresponds to a single test that expressed a warning.

`testRun.failedTests` _list_, _optional_

Each entry corresponds to a single test that failed.

## Example

```json
{
"_type": "https://in-toto.io/Statement/v1",
"subject": [
{
"digest": {
"gitCommit": "d20ace7968ba43c0219f62d71334c1095bab1602"
}
}
],
"predicateType": "https://in-toto.io/attestation/test-result/v0.1",
"predicate": {
"result": "pass",
"testRun": {
"link": "https:/in-toto/in-toto/actions/runs/4425592351",
"configuration": [{
"name": ".github/workflows/ci.yml",
"downloadLocation": "https:/in-toto/in-toto/blob/d20ace7968ba43c0219f62d71334c1095bab1602/.github/workflows/ci.yml",
"digest": {
"gitBlob": "ebe4add40f63c3c98bc9b32ff1e736f04120b023"
}
}],
"passedTests": [
"build (3.7, ubuntu-latest, py)",
"build (3.7, macos-latest, py)",
"build (3.7, windows-latest, py)",
"build (3.8, ubuntu-latest, py)",
"build (3.8, macos-latest, py)",
"build (3.8, windows-latest, py)",
"build (3.9, ubuntu-latest, py)",
"build (3.9, macos-latest, py)",
"build (3.9, windows-latest, py)",
"build (3.10, ubuntu-latest, py)",
"build (3.10, macos-latest, py)",
"build (3.10, windows-latest, py)",
"build (3.x, ubuntu-latest, lint)"
],
"warnedTests": [],
"failedTests": []
}
}
}
```

0 comments on commit 3ee93c9

Please sign in to comment.