plan: one registry for every code, or the guard is a coin toss #1658
No reviewers
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
jaylouisw/packetforge!1658
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "plan/1116-1210"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Forward plan only, no code.
docs/plans/error-code-registry.md.Why now: #1654 is open and its symptom is already gone from main � #1647 took the fixture from 2202 to 2203 specs, seed 65 moved, #1655 moved it again. Measured on
abbab2e8:validation.RULESholds 24 codes; the cabling path (validation.js,media.js,optical.js,topology.js,link.js) emits 56, so 32 are unregistered;engine/as a whole emits 84.E_BIDI_MISMATCH(engine/optical.js:225) is one of the 32.The three guards that claim this ground all miss it:
test/validation.test.js:370is titled "RULES documents every code the engine can emit" but only shape-checks the table against a hardcoded 16-name list;docs/PHYSICAL_LAYER.md:181-196tabulates those same 16 of 24;test/property-validator.test.js:125gates onRULESand exemptsE_PORT_IN_USEby name.The plan is eight steps, each one PR: an inventory script, the 32 registrations, a completeness test that reads the sources, the property gate pointed at the same list, flip to a gate, generate the doc table, the runtime-code family (labelled needs-Jay), and decoupling the property test sample from the catalogue size so #1654 cannot recur on someone else's device PR.
Closes #1654.