Skip to content

Extract Devantler Tech into business-site and retain GitHub Pages #3086

Closed
1 of 1 issue completed
Closed

Description

@devantler

🤖 Generated by the Agentic Engineer

Outcome

Move the public Devantler Tech website into devantler-tech/business-site, as explicitly requested by the maintainer on 2026-10-07. Aggregate it at applications/business-site in the monorepo. This supersedes the older proposed devantler.tech repository name. Hosting stays on GitHub Pages; no cluster tenant, paid hosting or DNS change is required for this static site.

Why

The website is a first-class application product and needs its own source, tests, review and issue queue, rather than being a special case in the portfolio aggregator. The maintainer has now prioritized this migration. The separately requested client portal belongs in devantler-tech/client-portal, with its own application product and dedicated platform tenant; it is not part of the static website extraction.

Delivery sequence

  1. Register business-site in the organization configuration, with team ownership, labels, pinned-action permissions and Observe-first adoption. The provider cannot send the five required organization properties at creation; the documented one-time bootstrap must set them, including AutomaticallyRequestCopilotCodeReview=false. No protected branch or review bypass.
  2. Extract the authored website, assets, CV generator, production checks and pinned workflows into the new repository. Preserve source history/provenance and the current public paths. Keep monorepo ADRs and agent deployment helpers in the monorepo.
  3. Add the website as an isolated submodule and first-class portfolio application. Repoint aggregator checks without dropping their contracts.
  4. Put the complete reusable publishing workflow in business-site. Retain the existing monorepo Pages resource and custom domain: a thin reviewed caller resolves the website gitlink and invokes an immutable publishing-workflow pin with that exact source revision. Prove publication before retiring the duplicated monorepo implementation. Source adoption and publishing-workflow adoption are separate reviewed pins; no domain transfer is needed.

Acceptance criteria

  • Public devantler-tech/business-site exists, classified as Application and declared in .github GitHub-as-code, with team-owned CODEOWNERS, canonical AGENTS.md and reviewed CI.
  • The website source lives in the new repository; applications/business-site is a pinned submodule and portfolio/product-card references identify the correct owner.
  • All existing production behavior and affected contracts pass with Node 24 and LC_ALL=C. English/Danish navigation, journal, artwork, project bookmarks, CV, theme selection, mobile layout and the introductory prices remain intact.
  • A reviewable PR build artifact provides the preview path. Production deployment is restricted to reviewed main, with no new paid service.
  • A successful existing-resource Pages deployment using the new repository's exact reviewed source and callable publisher, plus actual live https://devantler.tech browser paths, proves delivery. The monorepo retains Pages resource ownership and the thin dispatch adapter; website source, CI and the full publishing implementation belong to business-site. No DNS change, domain transfer, competing publisher or paid service, and a documented pin-based rollback path.
  • Company/contact follow-up Publish verified company and contact details from business-site #3917 remains open. No legal/contact details, customer claims, purchases or subscriptions are invented.

The requested removal of the duplicate technical catalogue is a separate, independently tested website polish slice; it must not be hidden inside a behavior-preserving extraction.

Activity

  1. coderabbitai commented on Aug 27, 2026

    @coderabbitai
    ⚠️ Possible Duplicate Issue(s)

    📝 Issue Planner

    Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.

    • Create Plan

    🧪 Issue enrichment is currently in open beta.

    You can configure auto-planning by selecting labels in the issue_enrichment configuration.

    To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

    issue_enrichment:
      auto_enrich:
        enabled: false

    💬 Have feedback or questions? Drop into our discord!

  2. moved this to 📥 Backlog in 🌊 Project Boardon Aug 27, 2026
  3. self-assigned this
    on Oct 7, 2026
  4. changed the title [-]Extract devantler.tech into its own repo (hosting stays on GitHub Pages)[/-] [+]Extract Devantler Tech into business-site and retain GitHub Pages[/+] on Oct 7, 2026
  5. changed the issue type fromtoon Oct 7, 2026
  6. added a commit that references this issue on Oct 7, 2026
  7. added 2 commits that reference this issue on Oct 7, 2026
  8. added 2 commits that reference this issue on Oct 9, 2026
  9. moved this from 👀 In Review to ✅ Done in 🌊 Project Boardon Oct 9, 2026
  10. devantler commented on Oct 9, 2026

    @devantler
    ContributorAuthor

    🤖 Generated by the Agentic Engineer

    Migration completion readback

    Reviewed cleanup #4058 was merged through the normal protected exact-head squash path at 11:25:25 UTC on 9 October 2026. Actual squash c998f0ce7cbc3cd523b97529e8f06950bee50317 has a GitHub-valid signature, parent f54592a44ea8869a91f59ab6e0d086eacf532f5e, and exactly the reviewed head's tree. Issue #3086 actually closed as completed at 11:25:27 UTC; its canonical board card is readback-verified Done.

    The independent business site was published before legacy retirement through reviewed #4056 and successful Pages run 37916186339. The actual public receipt binds source ae3a41b912246c078d32a51191ccd4583d490ede to caller f54592a44ea8869a91f59ab6e0d086eacf532f5e. Real EN/DA Home, all thirteen portfolio headings, keyboard/theme/mobile paths and fifteen distinct loaded Journal covers were exercised; screenshot-based visual sign-off remains unavailable and is separately tracked, not invented.

    The final cleanup #4058 removes 128 tracked frozen-legacy application files while retaining all seven architecture decisions and architecture instructions. The independent application gitlink and immutable publisher are unchanged. Removed source/assets remain recoverable in Git history; rollback is pinned-source based, not a second writable website.

    Current-head substantive CodeRabbit review is GREEN, with no actionable findings, inline comments or unresolved threads. The retirement guard was RED before removal and GREEN afterward, preserving restoration/missing-architecture negative controls. Both draft and promotion CI 37921730860 passed, with all 127 jobs read over two complete pages in each run. Final pre-merge native evidence covered all 267 contexts over three complete pages, all SUCCESS/SKIPPED, and all eight head-matching workflows were successful. Full five-comment/zero-review/zero-thread and REST inline reads were complete, the current head/base were unchanged, and the host holder check showed only this invocation's own processes. No admin, force, auto-merge, cancellation, check waiver or new provider request was used.

    Configured native code-quality analysis remains UNVERIFIED/UNKNOWN, not waived or relabelled complete. The local 109-script batch ended with exit 1, not blanket green; focused host reruns and hosted evidence are separate. The fresh screenshot capture limitation is not called visual proof.

    The original extraction sequence was independently read back: monorepo #3943 merged / #3941 closed, business-site #2 merged / #1 closed, then monorepo #3951 merged. The source is a public, GitHub-as-code managed application with maintainer ownership; the actual PR preview artifact exists and the publication caller uses the immutable reviewed source. Fresh canonical main is the cleanup squash, and its source/publisher pins remain unchanged. The actual public receipt was freshly read after the cleanup merge and still reports source ae3a41b912246c078d32a51191ccd4583d490ede / caller f54592a44ea8869a91f59ab6e0d086eacf532f5e, as expected: cleanup does not request another Pages deployment or claim its new squash as the live caller.

    Remaining work is explicit: business-site #5's screenshot/design sign-off, default-off Journal presentation and revised offer/maintenance/identity-copy activation, legal obligations, and the real client portal. #3917 remains OPEN. The loopback client portal is a fictional in-memory prototype, not authentication, persistence, customer isolation, payments or production delivery. No public portal routing or paid service was enabled. The session schedule remains disabled.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

Fields

Priority

None yet

Projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions