Files

Claude Code d436dc3598 P0-1 partial + AppImage fallback: clean placeholder for every broken image

P0-1 (featured-image fallback path):
- AppImage.tsx: default fallbackSrc changed from /assets/images/no_image.png
  to /assets/images/article-placeholder.svg, so any article image that
  loads but 404s gets the new clean dark gradient placeholder instead of
  the old grey BB_Gray-style box. The smoke-test article
  (matrox-video-and-amagi-collaborate-...) had a featured_image URL that
  404s on broadcastbeat.com; it now renders the new placeholder.

P0-1 batch (Wayback recovery) is NOT done — see PENDING_ads.md. The
attempted backfill via Wayback found no snapshots for the Matrox smoke
article (it is a 2026 NAB story Wayback never crawled), and the batch
loop tripped a 403 from one of its outbound calls. Proper backfill
needs per-vendor press-kit scraping and a new bb-media storage bucket.

PENDING_ads.md updated with: featured-image backfill TODOs, the banners
that still need local creatives, the hardcoded slug refactor for
NewsTicker / FeaturedBento / ArticleFeed, and the SEO re-enable gate.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-14 17:30:57 +00:00

3.1 KiB

Raw Blame History

Pending — broadcastbeat.com tonight's deferred work

Featured-image backfill (P0-1, deferred)

The clean placeholder at /assets/images/article-placeholder.svg is live for every article whose featured_image is NULL, contains no_image, BB_Gray, placeholder, or 404s on load (see src/components/ui/AppImage.tsx). That removes the bad UX immediately for ~518 articles.

The actual content backfill (Wayback → manufacturer press kit → real image in Supabase storage) is NOT done. The smoke-test attempt /tmp/bb-image-backfill.py ran on coolify01:

Matrox+Amagi smoke article (matrox-video-and-amagi-collaborate-...): Wayback returned no archived snapshot of https://www.broadcastbeat.com/<slug>/, so the og:image extraction path didn't apply. Many of the broken-image articles are recent NAB 2026 posts that have never been crawled by Wayback.
Batch loop aborted on an HTTP 403 from one of its outbound calls (Wayback API rate limit or transient Supabase RLS — needs follow-up).

What needs to happen next session:

For each broken article, try multiple sources in order:
- Wayback Machine snapshot of broadcastbeat.com/<slug>/ (skip if 404).
- Manufacturer press-kit page based on title keywords (Matrox, Amagi, Telestream, Calrec, AJA, etc. — keep a per-vendor mapping of press kit URLs).
- Generic image search via the vendor's news page.
Upload to a NEW Supabase storage bucket bb-media (currently doesn't exist; only cms-images, banner-creatives, avatars, group-images, city-images are present). Need to create with public-read.
Update bb.wp_imported_posts.featured_image with the public storage URL.
Smoke-test on the Matrox+Amagi article first per Ryan's instruction.

Banner creatives (P0-2 follow-up)

Adsanity iframe path removed. Banners pruned because they had no local image. Each one needs a local creative dropped into public/legacy/ads/ and an entry added to src/lib/ads.ts:

AJA Video 300x250
Zixi 300x250
Lectrosonics 300x250
Opengear 300x250
Studio Hero 300x250
Magewell 300x250
Blackmagic 300x600 (was originally Adsanity ad-210571)

Hardcoded article slugs (P0-4 follow-up)

/articles/[slug] and /news/[slug] now redirect to /news for unknown slugs, so the dead links from NewsTicker.tsx, FeaturedBento.tsx, and ArticleFeed.tsx no longer 404 — they redirect to the news index. But those components still hardcode fake slugs:

NewsTicker headlines array (7 hardcoded entries)
FeaturedBento hero + rotatingStories (1 + 15 hardcoded entries)
ArticleFeed seed entries (~6 hardcoded)

These should be refactored to pull real recent slugs from bb.wp_imported_posts so the headlines link to actual articles, not the redirect catch-all. Out of scope for tonight.

SEO re-enable

src/app/layout.tsx currently sets robots: { index: false, follow: false } sitewide. Flip both to true once:

Phase B (PR rewrite system) lands so the index doesn't get filled with raw press releases competing with the source.
Phase C (news sitemap at /news/sitemap-news.xml) lands.
Google Rich Results Test passes on a sample /news/<slug> URL.