StoneCypher / jssm / 1a8b9609889078528a2c6041276df2f94e383bf3 / 6

100%

master: 100%

Pull #973

github

docs(specs): decide npm ownership — self-publish now, co-publish as the FSL org later

Widened the maintainer survey from 5 packages to 21 (15 published), which
sharpened §2 and produced the answer.

Ownership, actually surveyed:

  9  purely self-published    apex, cypher, structured-text, cshtml-razor,
                              luau, func, curl, 4d, vue
  3  author + an org account  solidity, sap-abap, supercollider
  3  org-owned core           highlight.js, cdn-assets, vue-plugin
  6  no npm package at all    bbcode, raku, terraform, tsql, liquid, rdflang

Two findings. An org account is never SOLE owner of a grammar -- always
alongside the author, so it reads as adoption, not policy. And six org
repos never shipped to npm at all, several starred and non-trivial:
"repo exists in the org" and "grammar is installable" are separate
states, and a third of the sample never reached the second. That is what
"we will publish it later" looks like after a few years, and it is why
fsl#1978 bundles publish with SUPPORTED_LANGUAGES.md registration.

New 2b records John's decision: publish highlightjs-fsl from his own
account now, then add the FSL org's npm account as a co-maintainer once
that org exists (~Sept 2026). That is the solidity / sap-abap /
supercollider pattern -- the ecosystem's own answer to succession -- and
it beats the scoped form we rejected on every axis: no membership ask,
no name change, no deprecation, no split download stats, no scoped-path
work in the collector. `npm owner add` needs no republish.

Claude-Session: https://claude.ai/code/session_01EBLDd3nmRPcpC2weqD2TD3

3869 of 3869 branches covered (100.0%)

Branch coverage included in aggregate %.

6361 of 6361 relevant lines covered (100.0%)

5563.0 hits per line