feat(infra): binary mirroring and manifest generation from R2

[CalvinAllen]

Summary

Infrastructure to mirror all runtime binaries from upstream sources to builds.dtvem.io and generate manifests from our hosted binaries.

Implemented Features

Mirror Tool (scripts/mirror-binaries)

• Downloads binaries from upstream URLs in manifests
• Uploads to Cloudflare R2 with metadata files
• Tracks checksum source (upstream or dtvem)
• Parallel downloads with configurable workers
• SHA256 checksum verification
• Sync-only mode (skip existing files)
• Dry-run mode for testing

Manifest Generator (scripts/generate-manifests-from-r2)

• Scans R2 bucket for .meta.json files
• Generates manifests with builds.dtvem.io URLs
• Includes sha256_source field in output

Workflows

• mirror-all.yml: Bulk migration for initial R2 population
• mirror-sync.yml: Daily sync to catch new upstream versions
• generate-manifests-from-r2.yml: Auto-generates manifests after mirroring, deploys to R2, creates PR for embedded manifests

Schema Updates

• Added sha256_source field to manifest schema ("upstream" or "dtvem")
• Updated Go manifest.Download struct with SHA256Source field

Workflow Chain

Mirror Sync (daily) or Mirror All (manual)
         ↓
Generate Manifests from R2 (automatic)
         ↓
├─ Deploy manifests to R2 (immediate)
└─ Create PR for embedded manifests

Scope

Runtime	Versions	Downloads
Node.js	703	3,922
Python	132	482
Ruby	98	427
Total	933	4,831

Related

• Parent: feat(infra): automated build system for missing binaries #159
• PR: feat(infra): add binary mirroring workflows and tool #166

[CalvinAllen]

Completed in PR #166