Prompt the canonical site URL during bootstrap script to rewrite website crawlers.

Scripts/bootstrap

@@ -24,6 +24,11 @@ fi
 
 XCODEPROJ="./Site.xcodeproj"
 
+# The placeholder origin the crawler files ship with. It is an RFC 2606 reserved
+# domain, so an un-bootstrapped copy can never point a crawler at a real site.
+PLACEHOLDER_URL="https://site.example.com"
+PLACEHOLDER_URL_PATTERN='https://site\.example\.com'
+
 # --- Prompts ------------------------------------------------------------------
 
 printf 'Site display name (PascalCase, e.g. Berlin) [Site]: '
@@ -35,9 +40,31 @@ printf 'Project slug (lowercase, used for owner/db/compose, e.g. loud-berlin) [%
 read -r SLUG
 SLUG="${SLUG:-$DEFAULT_SLUG}"
 
+printf 'Canonical site URL (scheme + host, e.g. https://berlin.example.com) [%s]: ' "$PLACEHOLDER_URL"
+read -r URL
+URL="${URL:-$PLACEHOLDER_URL}"
+URL="${URL%/}" # the rewrites append their own path, so drop a trailing slash
+
+# The URL is interpolated into a sed replacement (with `|` as the delimiter) and
+# into the crawler files, so reject anything that is not a plain scheme + host.
+case "$URL" in
+	http://* | https://*) ;;
+	*)
+		echo "error: the canonical site URL must start with http:// or https:// (got \"$URL\")." >&2
+		exit 1
+		;;
+esac
+case "$URL" in
+	*[!-A-Za-z0-9:/._~]*)
+		echo "error: the canonical site URL contains unexpected characters (got \"$URL\")." >&2
+		exit 1
+		;;
+esac
+
 echo
 echo "  display name : $NAME       (server name \"${NAME}Website\", ${NAME}.xcodeproj)"
 echo "  slug         : $SLUG       (container owner, database name/user, compose project)"
+echo "  canonical URL: $URL       (robots.txt sitemap reference, sitemap.xml entry)"
 echo
 printf 'Apply these values? [y/N]: '
 read -r CONFIRM
@@ -95,6 +122,14 @@ rewrite "$W/docker-compose.override.yml" \
 rewrite "$W/Makefile" \
 	-e "s/),site)/),${SLUG})/g"
 
+# Crawler files: the absolute origin they must carry. `|` is the sed delimiter,
+# since both the placeholder and the replacement contain slashes.
+rewrite "$W/Resources/Static/robots.txt" \
+	-e "s|${PLACEHOLDER_URL_PATTERN}|${URL}|g"
+
+rewrite "$W/Resources/Static/sitemap.xml" \
+	-e "s|${PLACEHOLDER_URL_PATTERN}|${URL}|g"
+
 # Service README.
 rewrite "$W/README.md" \
 	-e "s/^# Site Website\$/# ${NAME} Website/" \
@@ -147,6 +182,7 @@ Next steps:
        - $W/Sources/Library/Catalogs/Localizable.xcstrings   (copy)
        - $W/Sources/Library                                   (landing / 404 pages)
        - $W/Resources/Static                                  (css, js, favicon, icons)
+       - $W/Resources/Static/site.webmanifest                 (name / short_name)
   2. Set a real database password in a git-ignored $W/.env
      (the committed .env.local defaults the password to the slug — do NOT ship that).
   3. Point the git remote at your new repository:
@@ -157,3 +193,12 @@ Next steps:
 The Persistence package still ships an ExampleRecord / ExampleRepository sample
 model — replace it with your real domain models when you add persistence.
 EOF
+
+if [ "$URL" = "$PLACEHOLDER_URL" ]; then
+	cat <<EOF
+
+WARNING: the canonical site URL was left as the placeholder "${PLACEHOLDER_URL}".
+Set the real origin in $W/Resources/Static/robots.txt and sitemap.xml before
+going live, or crawlers will be pointed at a reserved example domain.
+EOF
+fi