Skip to content

dream(content-integrity): href="#…" anchors all resolve to id="…" targets in #14

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jjohare merged 1 commit into from
Oct 6, 2026
Merged

dream(content-integrity): href="#…" anchors all resolve to id="…" targets in #14

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 18 additions & 3 deletions scripts/dream-link-check.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,14 +3,19 @@
# the built static site. Emits a surface-dependent count; non-zero missing
# refs prints LINK-INTEGRITY-FAIL. Used as a dream-cycle evaluator entrypoint.
#
# Two ref families are resolved:
# Three ref families are resolved:
# 1. href/src attributes — ordinary markup links and asset references.
# 2. meta[content] media URLs (og:image, twitter:image, …). These were a
# shared blind spot: dream-meta-tags-scan.sh COUNTS the tags but never
# looks at the URL, and this script only ever parsed href/src, so an
# og:image pointing at a file absent from dist/ passed both gates. The
# social card is a required asset in website/assets.manifest.json, and
# this makes that requirement observable from the link surface too.
# 3. in-page anchors — every href="#id" must resolve to an id in the same
# shipped document. Family 1 skips refs containing '#' by design, so a
# nav link to a renamed or removed section passed this gate silently
# (2026-09-28 finding; that PR never landed). Baseline 0 dangling,
# operator-verified 2026-10-02, so dangling > 0 now FAILS.
#
# Social URLs are absolute on the canonical domain (they must be, for the
# crawlers), so "site-relative" here includes same-origin absolute URLs: the
Expand Down Expand Up @@ -38,7 +43,7 @@ resolve_ref() {
esac
}

miss=0; checked=0
miss=0; checked=0; anchors=0; dangling=0
while IFS= read -r f; do
d=$(dirname "$f")

Expand All @@ -58,6 +63,16 @@ while IFS= read -r f; do
checked=$((checked+1))
[ -e "$d/$t" ] || [ -e "./$t" ] || { echo "MISSING: $f -> $ref"; miss=$((miss+1)); }
done

# 3. in-page anchors — same-document only: a fragment never names a file, so
# resolution is an id lookup within the referring page itself.
ids=$(grep -oE 'id="[^"]+"' "$f" 2>/dev/null | sed -E 's/^id="([^"]*)"$/\1/' | sort -u)
for ref in $(grep -oE 'href="#[^"]+"' "$f" 2>/dev/null | sed -E 's/^href="#([^"]*)"$/\1/'); do
anchors=$((anchors+1))
printf '%s\n' "$ids" | grep -qxF -- "$ref" \
|| { echo "DANGLING-ANCHOR: $f -> #$ref"; dangling=$((dangling+1)); }
done
done < <(find . -name '*.html')
echo "internal-refs-checked: $checked missing: $miss"
[ "$miss" -eq 0 ] && echo LINK-INTEGRITY-OK || echo LINK-INTEGRITY-FAIL
echo "anchors-checked: $anchors dangling: $dangling"
[ "$miss" -eq 0 ] && [ "$dangling" -eq 0 ] && echo LINK-INTEGRITY-OK || echo LINK-INTEGRITY-FAIL
Loading