Import normalizer: offline tool to normalize the raw archive spreadsheets #663

Merged
marcel merged 172 commits from docs/import-migration into main 2026-05-28 15:05:51 +02:00
Showing only changes of commit 4bc96c3772 - Show all commits

View File

@@ -72,10 +72,15 @@ jobs:
# Svelte default escaping, never {@html}. This guard flags any {@html ...}
# whose expression references a raw-date variable. A comment mentioning
# "{@html}" without a raw token inside the braces does NOT match.
pattern='\{@html[^}]*(metaDateRaw|documentDateRaw|rawDate)'
# Self-test: the regex must catch the dangerous form and ignore the comment form.
# The token list MUST cover every variable that carries the raw value:
# DocumentDate.svelte exposes it via the `raw` prop, so `\braw\b` is included.
# Grow this list whenever a new raw-bearing variable name is introduced.
pattern='\{@html[^}]*(metaDateRaw|documentDateRaw|rawDate|\braw\b)'
# Self-test: the regex must catch the dangerous forms and ignore the comment form.
printf '{@html doc.metaDateRaw}\n' | grep -qP "$pattern" \
|| { echo "FAIL: guard self-test — regex missed the unsafe {@html metaDateRaw} form"; exit 1; }
printf '{@html raw}\n' | grep -qP "$pattern" \
|| { echo "FAIL: guard self-test — regex missed the unsafe {@html raw} form (DocumentDate prop)"; exit 1; }
printf 'never use {@html} for this\n' | grep -qvP "$pattern" \
|| { echo "FAIL: guard self-test — regex wrongly flagged a {@html} comment"; exit 1; }
if grep -rPln "$pattern" --include='*.svelte' frontend/src/; then