Resolve merge conflicts with main, fix pre-existing Vitest failures, and extend CI build timeouts

[Copilot]

The branch had diverged from main, producing 15 conflicts across the daily realtime-monitor article and its 14 rendered news HTML pages.

All conflicts were in generated content. The branch side carried only timestamp churn (generated_at, og:updated_time, dateModified, footer "Last updated"), while main carried real new content — the TU digital-governance cluster (HD01TU17 telecom anti-spoofing rules, HD01TU18 interoperability law), added significance-table rows, and a new cross-reference cluster node.

Resolution

• Generated content (15 files) — adopted main's newer, content-rich versions; discarded the branch's stale timestamp-only edits.
• Feature code — data/political-glossary.json, scripts/render-lib/political-context.ts, scripts/types/language.ts, and the article.ts/article-aside.ts wiring auto-merged without conflict and remain intact.
• Finalized as a true merge commit (two parents: branch HEAD + origin/main).

No source logic was authored during resolution — the choice was strictly between two existing content versions, favoring main's latest published output. The political-context feature is unaffected; its HTML output will be applied on the next render pipeline regeneration, consistent with the branch's pre-merge state.

Vitest CI fixes

After the merge, the Unit Tests (Vitest) job reported 7 failures (2 files). These were pre-existing on main, introduced by commit f41c7307 "Refine narrative aggregation for reader-friendly flow", which renamed reader labels and section titles and changed order.ts but left two test files (and one consumer) unaligned. Both root causes were fixed:

• scripts/check-headline-quality.ts (source fix, 3 tests) — Rule E (which flags H1s that merely echo the subfolder slug) was routing through the renderer's prettifyFallbackTitle, whose curated CATEGORY_FALLBACK_TITLES map turns slugs like propositions into engaging titles ("Government Bills — Swedish Legislative Agenda") and thereby defeated the bare-slug detection. Added a local plain slug→Title-Case prettifier (prettifySlugSegment) used by prettifySubfolderPath and the leaf comparison, and removed the now-unused import. Rule E once again flags lazy slug-echo H1s for curated subfolders.
• tests/article-aside.test.ts (test fix, 4 tests) — updated stale string expectations to match f41c7307's deliberate, merged renames (BLUF → "Lede and editorial decisions"; political-classification → #rm-deep-dive-political-classification; classification-results → rm-deep-dive-classification-results). f41c7307 updated render-lib.test.ts/render-lib-leaf-modules.test.ts but missed this file; all structural assertions remain intact.

Verified: both files green in isolation (36/36, 18/18), related batch (render-lib, render-lib-leaf-modules, political-context) 382/382, tsc typecheck clean, and full npm test reports 0 failures.

CI build timeout extension

The Test News / Build for News Tests job was cancelled after 15m, which then skipped its dependent News & Sitemap Tests job. Investigating the Actions logs showed the shared npm run build step now takes ~14m42s on its own (it renders ~5,400 pages — 391 subfolders × 14 languages), pushing the full build job (checkout + npm ci + build + artifact upload) to ~17min — past the 15-min cap. The passing Build Test (Vite) job (no timeout cap) confirmed the build genuinely needs ~17min.

• .github/workflows/test-news.yml — build job timeout-minutes 15 → 25.
• .github/workflows/test-homepage.yml — build job timeout-minutes 15 → 25 (identical build, same cap, same risk).

25min gives ~8min of durable headroom over the observed ~17min job as the news corpus keeps growing. On performance: the build hot spot is the CPU-bound unified/remark/rehype markdown pipeline, where bounded Promise.all concurrency gives no real wall-clock gain (single-threaded, no I/O to overlap) and would risk the render test suite; true parallelism would require a worker_threads rewrite, so the safe, deterministic timeout fix was chosen instead. CodeQL (actions) scan reports 0 alerts.

[github-actions]

🏷️ Automatic Labeling Summary

This PR has been automatically labeled based on the files changed and PR metadata.

Applied Labels: size-xs

Label Categories

• 🗳️ Content: news, dashboard, visualization, intelligence
• 💻 Technology: html-css, javascript, workflow, security
• 📊 Data: cia-data, riksdag-data, data-pipeline, schema
• 🌍 I18n: i18n, translation, rtl
• 🔒 ISMS: isms, iso-27001, nist-csf, cis-controls
• 🏗️ Infrastructure: ci-cd, deployment, performance, monitoring
• 🔄 Quality: testing, accessibility, documentation, refactor
• 🤖 AI: agent, skill, agentic-workflow

For more information, see .github/labeler.yml.

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[pethers]

@copilot analyse and improve code and test quality and fix root cause of
TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request)
TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request)Failing after 4m

and resolve merge conflicts

[Copilot]

Copilot wasn't able to review this pull request because it exceeds the maximum number of files (300). Try reducing the number of changed files and requesting a review from Copilot again.

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[pethers]

@copilot analyse and improve all changes. MUST fix so all PR checks pass and all code is correct

TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request)
TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request)Failing after 4m

[Copilot]

@copilot analyse and improve all changes. MUST fix so all PR checks pass and all code is correct

[TypeScript & JavaScript Testing / Unit Tests (Vitest) (pull_request)](https://githu...

Fixed in 76f56fc. The 7 Vitest failures were pre-existing on main (introduced by f41c7307 "Refine narrative aggregation for reader-friendly flow"), surfaced here after the merge:

• 3 check-headline-quality (Rule E) — root cause in source: prettifySubfolderPath/matchesPrettifiedSubfolder routed through the renderer's curated prettifyFallbackTitle, which maps propositions → "Government Bills — …" and stopped Rule E from flagging lazy bare-slug H1s. Fixed by adding a local plain slug→Title-Case prettifier so Rule E compares against the bare label.
• 4 article-aside — those expectations still asserted the pre-rename strings (BLUF, #rm-political-classification, rm-classification-results). f41c7307 deliberately renamed the labels (BLUF→"Lede", section titles→"Deep Dive: …") and updated render-lib.test.ts/render-lib-leaf-modules.test.ts but missed article-aside.test.ts. Updated the stale strings; all structural assertions kept intact.

Verified: both files green in isolation (36/36, 18/18), related batch 382/382, tsc typecheck clean, full npm test reports 0 failures.

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json

[pethers]

@copilot and fix issues in

failing checks
Test News / Build for News Tests (pull_request)
Test News / Build for News Tests (pull_request)Cancelled after 15m
in progress checks
Loading
TypeScript & JavaScript Testing / Multi-Language Sanity Tests (pull_request)
TypeScript & JavaScript Testing / Multi-Language Sanity Tests (pull_request)Started 1 minute ago — This check has started...
skipped checks
Test News / News & Sitemap Tests (pull_request)
Test News / News & Sitemap Tests (pull_request)Skipped 2 minutes ago

extend timeouts so all test have time to complete. improve performance if possible

[Copilot]

@copilot and fix issues in

failing checks
...

Fixed in the latest commit. Root cause of the cancellations: the shared npm run build step now takes ~14m42s on its own (renders ~5,400 pages — 391 subfolders × 14 languages), pushing the full build job (checkout + npm ci + build + artifact upload) to ~17min. The Test News / Build for News Tests job was capped at timeout-minutes: 15, so it was cancelled mid-render, which in turn skipped the dependent News & Sitemap Tests job. (The Multi-Language Sanity Tests check completed once its build artifact was available.)

• Extended test-news.yml build timeout-minutes 15 → 25.
• Extended test-homepage.yml build timeout-minutes 15 → 25 (identical build, same cap, same risk).
• 25min gives durable headroom (~8min over the observed ~17min job) as the news corpus keeps growing.

On performance: the build hot spot is the CPU-bound unified/remark/rehype markdown pipeline. Bounded Promise.all concurrency wouldn't meaningfully cut wall-clock (single-threaded, no real I/O to overlap) and would risk the render test suite; true parallelism would require a worker_threads rewrite. I opted for the safe, deterministic timeout fix rather than risk regressions.

[pethers]

@copilot resolve the merge conflicts in this pull request

[Copilot]

@copilot resolve the merge conflicts in this pull request

Resolved in 029fcf8. Two conflicts from main's recent merges:

• scripts/render-lib/article.ts — kept both import lines (branch's enrichArticleMarkdownWithPoliticalContext + main's scannability imports from PR feat: add visual scannability — confidence chips, in-article TOC, progressive disclosure, methodology footer #2807).
• .github/workflows/test-homepage.yml — main bumped timeout 15→20; our branch already has 25 (supersedes). Combined comment history.

tsc typecheck clean, 159/159 relevant tests passing.

[github-actions]

🔍 Lighthouse Performance Audit

Category	Score	Status
Performance	85/100	🟡
Accessibility	95/100	🟢
Best Practices	90/100	🟢
SEO	95/100	🟢

📥 Download full Lighthouse report

Budget Compliance: Performance budgets enforced via budget.json