2 Commits

Author	SHA1	Message	Date
svemagie	a166af2306	chore: sync upstream — performance, webmentions v2, OG v3 ... - _data: switch to cachedFetch wrapper (10s timeout + 4h watch cache) - js/webmentions.js: owner reply threading, platform provenance badges, DOM dedup, Micropub reply support - js/comments.js: owner detection, reply system, Alpine.store integration - _includes/components/webmentions.njk: data-wm-* attrs, provenance badge slots, reply buttons - _includes/components/comments.njk: owner-aware comment form, threaded replies - widgets/toc.njk: Alpine.js tocScanner upgrade (replaces is-land/inline-JS) - lib/og.js + og-cli.js: OG card v3 (light theme, avatar, batched spawn, DESIGN_VERSION=3) - eleventy.config.js: hasOgImage cache, memoized date filters, batched OG/unfurl, post-build GC, YouTube check opt - base.njk: Inter font preloads + toc-scanner.js script - critical.css: font-face declarations (font-display:optional) - tailwind.css: font-display swap→optional - tailwind.config.js: prose link colors -700→-600 - Color design system: accent-700/300 → accent-600/400 across components Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-03-15 23:56:56 +01:00
Ricardo	86cbc1ee5d	fix: run OG image generation in subprocess to prevent OOM kill ... The OG generation in the eleventy.before hook consumed too much memory alongside Eleventy's data cascade, causing the Eleventy process to be OOM-killed on Cloudron. Fix by running OG generation in a separate child process with its own 768MB heap limit. Also write the manifest incrementally (every 10 images) to preserve progress if interrupted.	2026-02-18 08:49:23 +01:00