2 Commits

Author SHA1 Message Date
Jakub Zych
a9a7a328e6 refactor: hard-fork GSD -> MSD (Make Software Done)
Mechanical rename produced by scripts/msd-rename.cjs: gsd/Gsd/GSD -> msd/Msd/MSD
across contents and paths, upstream package/repo coordinates -> @golem15/msd-core
and golem15com/msd-core. Deep links into upstream history, sibling upstream
packages, the GSD-2 import feature, CHANGELOG.md and .changeset/ are kept as-is.

Hand edits on top: MSD block-letter banner and logos, LICENSE copyright line,
package/plugin identity, regenerated lockfile, install-tree fixtures, derived
registries and benchmark baseline; migration checksum baseline re-locked
(MSD keeps its own install state, so no install had applied the old sums);
sort-order and regex-escaped expectations in tests adjusted.
2026-10-06 01:47:40 +02:00
Michel Moreira
1f84f45ed5 fix(#4259): fold shell continuations before the T6 docs-parity site scan (#4423)
* fix(#4259): fold shell continuations before the T6 docs-parity site scan

The scan is $-anchored with [^\n]* on both sides of --grep=, so `git log` and
`--grep=` had to share a physical line. A backslash-continued derivation —
the natural way to write a git log carrying a long ERE — produced zero hits
and T6 passed on it.

Both generations of the assertion were defeated. The current anti-revert ban
let a wrapped site through outright; at v1.12.0, where T6 instead asserted
pattern conformance, a wrapped site was silently exempted from the very
checks written to catch the macOS \b-no-op class, so it could have carried
exactly the malformed pattern T6 exists to reject. A real candidate
implementation for #3926 wrapped its derivation, passed T6, and was caught
only by later manual review.

Fold the continuations before matching rather than widening the regex: the
assertion's message and its PHASE_SCOPE_NUM filter both assume one site is
one string, and a [\s\S]*? would run the scan across unrelated statements.
Correcting the input repairs everything built on the scan at once.

The fold uses [ \t]* after the newline rather than \s* — the shell's own
rule, and it cannot swallow a blank line and glue two unrelated statements.
The scan is hoisted to findGrepSites so the controls can drive it directly:
the continued form is caught, the same-line form still is, a benign --grep
stays clean, a wrapped unrelated assignment is not glued, a continuation does
not cross a blank line, and the live workflow files still report nothing — so
this lands without editing any workflow to appease it.

* fix(#4259): honor shell continuation boundaries

* test(#4259): use canonical fenced-block scanner

* test(#4259): centralize shell continuation scanning

---------

Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
2026-09-11 10:40:37 -04:00