Record Source
The record source is the input interface — the role every supplier of node and edge records fills, whether a stream describes today’s content or brings back what the case recorded. Its requirements bind what any supplier must put on the records it supplies, so a datum the rest of the engine depends on cannot be present in one stream and missing from another.
The record source shall supply, in every test outcome record, the result of the test execution that outcome records. |
What the gate judges by
The record source shall supply, on every waiver record and every test outcome record, each datum the gate evaluator judges that record’s validity or currency by. |
The record source shall supply, in every waiver record, the date on which the waiver expires. |
The record source shall supply, in every waiver record, the name of the approver who granted it. |
The record source shall supply, in every test outcome record, the revision of the implementation repository the test execution ran against. |
Anchors
Where the anchored content was read from a repository, the record source shall name that repository in the content anchor by its configured name, never by a path. |
The record source shall supply, with every content hash it supplies, an anchor from which the content the hash covers can be found again. |
Content behind a hash
A record source that reads content can also hand that content back, so an operator can compare it with what the case recorded. The outcome reader supplies no content: the record it hashes is built from a run bundle.
Where a record source reads the content that a content hash it supplies covers, the record source shall supply, for a node and the name of one of its content hashes, the bytes from which it computed that hash. |