Two audit findings where the report overstated a gap in orgo's own documentation.
A construct named inside verbatim is not a use of it (#1). The entity check
skipped =verbatim= and ~code~ spans; the other inline checks did not, so a page
documenting org syntax reported constructs it does not use. orgo audit docs reported
OUT macro call at guide/05-org-support.org:186 — the table row naming macros in
order to say they are unsupported. scan_inline now blanks the literal spans once and
runs the timestamp, macro call, radio/internal target, LaTeX and entity checks against
that. Emphasis pairs still see the raw line, since = and ~ are themselves markers
being counted.
An unlisted keyword is pass-through, not a blind spot (#2). ??? claimed a name
the implementation does not recognize at all, and marked LEDE — read by the shipped
docs template, styled by themes/docs.css, documented in the configuration guide.
Reaching templates as page.keywords.<name> is the designed behaviour, so no keyword
name is a blind spot. Which keywords have dedicated handling is still worth reporting,
so the distinction stays and the marker changes: — for pass-through, ??? reserved
for drawers and link schemes, where an unlisted name has no such fallback. This is
option 1 from the issue — the one that makes the legend true.
KEYWORDS
DESCRIPTION 14 14 guide/01-cli.org:2
— LEDE 14 14 guide/01-cli.org:3
TITLE 14 14 guide/01-cli.org:1
orgo audit docs now reports 1022 in-scope uses (100.0%), 0 out-of-scope, and no
???. Two fixtures and two tests cover both, and the full suite passes.