Locale-correct structured data, and a robots.txt rule that matched nothing

Fixkamo-marketing
Shipped
August 12, 2026 at 8:36 PM UTC
Author
Kamo
Commit
8705796

Every BreadcrumbJsonLd and ItemList URL was hardcoded to /en, so the structured data on /fr/terms told crawlers its breadcrumb parent was the English page — 19 URLs across 15 files, contradicting the canonical and hreflang the same page emitted. Nine of those pages had a sync default export with no params, so they had no locale to build a URL from. robots.txt disallowed /subscribe, but localePrefix is "always" and the real paths are /<locale>/subscribe — /fr/subscribe served 200 and was never covered by the rule. The page's own noindex meant it stayed out of the index, so the dead rule was invisible; it only ever failed to save the crawl budget it was written for. contact-confirmed keeps an English OG url: its metadata is module-level and it is deliberately noindex/nofollow.

All changes

Like what you see shipping?

Every one of these updates lands in your workspace automatically. Start free and watch it grow week after week.

Start Free ForeverView Pricing