Why KNX Thermostat Setpoints Revert in Home Assistant and How to Fix It

·1 views

A Home Assistant user discovered that manually adjusting a KNX thermostat's target temperature appeared to work but immediately snapped back to its previous value. The cause lies in how KNX thermostats handle their five built-in preset modes — Standby, Building Protection, Economy, Comfort, and Auto — each of which stores its own fixed setpoint directly on the physical device. When the thermostat is in Comfort or Economy mode, it treats its stored setpoint as authoritative and overrides any external write from Home Assistant within seconds. Only Auto mode allows Home Assistant to freely set and retain a target temperature. The fix is straightforward: switch the climate entity's preset to Auto before adjusting the temperature, and the new value will stick.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Codename One Replaces Wrapped Javadoc with Hugo-Integrated API Docs

Codename One, an open-source framework for building cross-platform apps from Java or Kotlin, has overhauled how its API documentation integrates with its developer website. Previously, Javadoc pages were loaded through a JavaScript wrapper, causing mismatches in navigation, styling, and dark mode between the API docs and the rest of the site. The team built a new Maven doclet that feeds the API model directly into their Hugo static site generator, allowing both tools to handle what they do best. Each Java type now generates a Hugo content file rendered with the same theme and typography as the guide and blog, while a standard Javadoc archive is still produced for offline use. A redesigned API search index, compressed to 341 KB gzipped, lets users find API members alongside other site content within a unified interface.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Build vs Buy: Why Manufacturers Should Rethink CRM Strategy for External Users

Most manufacturers default to buying off-the-shelf CRM software, but the real cost and complexity lie in the external-facing layer used by dealers, distributors, and service technicians. Standard CRM platforms are built around contacts and pipelines, while manufacturing sales depend on ERP-held data like contract pricing, part codes, and production slots. Enterprise options such as Salesforce Manufacturing Cloud and HubSpot Sales Hub carry steep per-seat fees, and external portal users — often outnumbering internal staff — can nearly double the annual licence bill. A hypothetical mid-size manufacturer with 15 internal users and 300 dealer accounts could face over $139,000 per year in list-price Salesforce licences alone. This cost structure, combined with quoting complexity and the recent end-of-sale of Salesforce CPQ, is prompting manufacturers to evaluate which CRM components are worth building in-house versus purchasing.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Mapping Legacy Position Codes to Oracle HCM: Why a Translation Layer Matters

Data migration to Oracle HCM often hits a critical snag when source systems store worker assignments using a single legacy position code, while Oracle requires multiple distinct values such as job code, department, and business unit. Organizations frequently redesign their job and position structures during implementation, meaning legacy codes rarely map directly to Oracle values. A mapping-table approach — rather than hardcoded IF-THEN logic — serves as a translation layer, making conversions easier to maintain, test, and update as Oracle codes evolve. This strategy also improves governance by centralizing all mappings for review by functional and business teams. Crucially, worker assignments can only be loaded after departments, jobs, and positions have already been created in Oracle HCM, a dependency chain that project teams often underestimate during planning.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How Tekton, Cosign, and Kyverno Create a Fully Signed Kubernetes CI/CD Pipeline

A homelab engineer has documented a four-component CI/CD security stack built on RKE2 Kubernetes, ensuring every container image is built, scanned, signed, and verified before deployment. The pipeline uses Tekton for in-cluster image builds, SonarQube and Trivy for code quality and vulnerability scanning, and Tekton Chains to attach cryptographic signatures and SLSA provenance to each artifact. Images are stored in a self-hosted Harbor registry alongside their signatures, provenance data, and software bill of materials. Kyverno acts as the admission gatekeeper, rejecting any image at the cluster door that does not carry a valid cryptographic signature. The setup replaces Jenkins and Nexus with cloud-native alternatives, reducing complexity while enforcing an end-to-end verifiable software supply chain.

0 comments Read more at DEV Community