Several structured data defects in the immigration reform article package degraded SEO and broke schema consumers.

Fixes

Breadcrumb truncation (6 article files)

itemListElement[position=3].name was truncated mid-word (50-char cutoff) in fr, de, he, nl, es, ar articles:

// Before
"name": "Schwedens Riksdag treibt umfassende Einwanderungsr"

// After
"name": "Schwedens Riksdag treibt umfassende Einwanderungsreform voran: 12 Ausschussberichte veröffentlicht"

HTML entity double-encoding in JSON-LD (4 index files)

application/ld+json scripts in en/sv/da/no indexes contained HTML entities that structured data consumers ingest as literal strings:

// Before
"headline": "Sweden's Parliament Faces Busy Week..."
"description": "Riksdagen inleder en intensiv vecka med utskottsmöten, tre kammarsessioner och EU-samråd..."

// After
"headline": "Sweden's Parliament Faces Busy Week..."
"description": "Riksdagen inleder en intensiv vecka med utskottsmöten, tre kammarsessioner och EU-samråd..."

Truncated descriptions in ItemList JSON-LD (index.html, index_sv.html)

Descriptions cut off mid-word (EU migrat, uppehålls) — completed to full sentences.

Wrong article type (index_sv.html)

Immigration reform article classified as "prospective" → "breaking".

Schema versioning (last-generation.json)

Added "schemaVersion": "2" to disambiguate the realtime-monitor schema from the generate-news-enhanced.js schema, which write different key shapes to the same file.

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.