v1.2.9

[1.2.9] - 2026-06-02

Fixed

• Missing product images that never recovered. image_url was
  only written at create-time, so a first-scrape miss became permanent
  even when later scrapes would have succeeded. JS-heavy SPAs
  (digitec.ch via Puppeteer) hit this when render timing varied.
  Scheduler now backfills image_url whenever the current value is
  null and the scrape produced one — existing broken products
  self-heal on their next scheduled check.
• Lazy-load attribute coverage. extractGenericImage now also
  reads data-original, data-lazy, data-lazy-src, data-srcset
  and srcset (picking the first URL out of srcset descriptor lists).
  Reduces first-scrape misses on lazy-loaded image elements.