cleanup: 2026-11 audit plan (epic) #742
Labels
No labels
loop
queue/failed
queue/queued
ready
status/parked
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
torbscript/language#742
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
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?
Context
One audit of the whole repository (language, compiler, standard library, documentation, issue tracker) on 2026-11-13, after an accepted record (CLOSED-TRAITS, the leading dot) was found to have had no issue for six weeks. The audit lists what is wrong or missing, maps every design record to the issue that tracks its open part, and orders the cleanup so that the loop can take it item by item.
Source
docs/project/2026-11-cleanup-audit.md(PR #756). 14 findings: 1 blocker, 6 major, 7 minor.Steps (epic: no
readyon this issue)In plan order. Breaking changes come first, because after 1.0 each needs a deprecation path. owner marks a step only the owner takes. Steps without a dependency may run in parallel.
.Case, thebare-caselint rule (CLOSED-TRAITS 1a-1)?in a body whose result type is inferredwhereclause with a concrete subject,into()under anIntoboundMultiplyimplementations on one typetorb canon, move the formatter's rewrite modulesstd: remaining helpers and constructor namestorbSteps 15 to 17 are documentation only and have no dependency: they fit while step 2 waits for the seed.
Owner decisions
Listed with a recommendation each in section 6 of the report: the day of the seed for phase 1a,
staticagainsttype fn, the other open questions of CONCEPT, when the application framework is built, one suffix for error types, whichdocssubcommands are public.Dependencies
As written per step. Nothing here depends on an issue outside the list.
Each step is its own issue; this epic only orders them.
[loop:torbscript] audit checkpoint 1: done: tracker snapshot (142 open, 300 closed), design-record status lines, ROADMAP, CONCEPT open questions, leading-dot census, std error-type classification. In flight: tooling vs docs, compiler structure. Next: std API consistency, book and skill drift, then the report and the plan issues.
Checkpoint 2 (audit lead)
.Case, and thebare-caselint rule (CLOSED-TRAITS 1a-1) #743?in a body whose result type is inferred #747whereclause with a concrete subject, andinto()under anIntobound #748torb canonalias, and name the formatter's rewrite modules for what they are #751torbhas a page or a line that says it is internal #752Multiplyimplementations on one type are unreachable (LINEAR.md item 12) #753Checkpoint 3 (audit lead, final)
docs/project/index.mdis hand-written anddocs index --checkhas to confirm it.Owner decisions on the audit questions (Torben, WhatsApp 2026-11-13 13:57 Berlin):
staticvstype fn/type const: open, explanation sent; no decision yet.Errorfor what aResultcarries,Problemonly for the kind inside a record-shaped error; renameStorageFailure,PathFailure,JsonPointerProblem,PatchProblemin the 7.3 rounds. Accepted.torb upgradeanddocs checkget their own pages; the remainingdocssubcommands go on one contributor page. Accepted.Also open: a struct construction redesign (body form constructed with
Point { x: 5 }, head formtype Point(x: Int, y: Int)with()), in discussion with the owner; a design record follows once agreed. It supersedes the open question "Fields on one line".Owner follow-up (Torben, 2026-11-13 14:34 Berlin): point 1, Orion does the seed refresh as needed after #743. Point 2,
staticstays (recorded in #750). Struct redesign dropped: no head form, no struct literals, no migration (recorded in #750).