Writing
The complete chronological archive.
All writing
Page 2 of 10
I Spent an Hour Proving the TLS Was Fine
Aug 2026·Debug a KOReader Tailscale proxy failure step by step: check DNS and routing, rule out TLS, and add HTTP CONNECT tunnelling to LuaSocket and LuaSec.
The Plugin Only Wrote, It Never Read
Aug 2026·How I extended a KOReader bookmark plugin without rebuilding existing work: check remote branches, add search, fix device bugs, and report test scope.
Jailbreaking a Paperwhite in 2026
Aug 2026·A practical Kindle Paperwhite jailbreak guide for firmware 5.18.1.1.1: identify the model, block updates, run Sanctuary, and verify it worked.
Replacing Hand-Rolled Acquia Deployments With a CDK-Defined, Verification-Gated Pipeline
Aug 2026·Replace fragile Git deployment scripts with an Acquia deployment pipeline using AWS CDK, supported CLI operations, and verification gates.
Replacing Split CI/CD With a Verified CodePipeline-to-Acquia Deployment Path
Aug 2026·Split CI/CD consolidation removes duplicate checks by moving verified Acquia deployments and PR gates into AWS CodePipeline before retiring GitHub Actions.
Building a KPDCL Electricity Dashboard Without an API
Aug 2026·A KPDCL electricity dashboard built from Smart BS Plus data, SQLite, and Glance shows how silent SVG and CSS failures can lie while returning 200.
Why Configured Telemetry Does Not Prove Production Tracing
Aug 2026·Production tracing records state transitions but cannot make queue commits atomic or prove that retries and side effects are correct.
Measure Django Cache Performance Before Adding More Caching
Aug 2026·Measure Django cache performance by matching requests, tracing get, loader, and set stages, and testing whether hits remove limiting work.
Privacy-Safe Caddy Access Logs for Production Debugging
Aug 2026·Build access logging with privacy-safe Caddy records, one-event retrieval, observed retention, and deliberate loss detection.
How to Harden Cloudflare WAF Without Bot Management
Aug 2026·Keep Cloudflare WAF enforcement scope within reviewed evidence using narrow rules, conservative sampling, phase checks, and rollback.