docs(design): receiver header for scripts (SCRIPT-RECEIVERS.md) #504
Loading…
Reference in a new issue
No description provided.
Delete branch "docs/adr-receiver-types"
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?
Design record only, no implementation. Decision: a script states its receiver in a first-line header
receiver T from "path";project.trbis reserved by name; load-site inference and "the first load decides" go away; LSP types scripts alone, browser contracttorbscript/wantFiles/torbscript/files.End goal: hover/highlighting/go-to-definition in Forgejo for project.trb and imported scripts (browser LSP: https://git.dmb.zone/orion/infra/issues/2329).
Note: no
torbis built in the authoring environment, sotorb docs skill docs skills --checkwas not run; CI decides. Open question (header vs file name only) is posed on the epic.Tracked by epic #505; slices #506-#511.
CI red on
8ac5e97: ci / tier A (linux-x64) (-> gates) -docs index --check(anddocs check) fail: "docs/design/index.md:1: The generated part of this index is not up to date. Runtorb docs index". The hand-edited index.md must be regenerated withtorb docs indexand committed; the rest of tier A (compiler tests, format, std tests, lock) was green. No other blockers seen; the next CI run brings the review back.8ac5e970c8b81b31f497Approved on
b81b31f.Checklist:
docs index --check, pdisc, bootstrap). The earlier red index (8ac5e97) is fixed:docs/design/index.mdlists the new record in both the frontmatter and the body.project.trbreserved by name, load-site conflict becomes a compile error, forward-only reads for the browser). Options A-D are weighed with a clear rationale.Non-blocking:
orion/infra#2329, an issue in a tracker that is not public; consider describing the browser LSP work in words so the public design stands alone.