v1.2.x

[Nerivec]

Important

Included in the Zigbee2MQTT July 2025 release.
For standalone serving, see the wiki.

What's Changed

• feat: support table column toggling by @Nerivec in feat: support table column toggling #78
• fix: migrate to biome v2 by @Nerivec in fix: migrate to biome v2 #79
• fix: bump zigbee2mqtt from 3cd45a7 to 0744b67 by @dependabot in fix: bump zigbee2mqtt from 3cd45a7 to 0744b67 #85
• fix: bump the development-dependencies group with 7 updates by @dependabot in fix: bump the development-dependencies group with 7 updates #86
• fix: bump vite from 6.3.5 to 7.0.0 by @dependabot in fix: bump vite from 6.3.5 to 7.0.0 #87
• fix: various small fixes by @Nerivec in fix: various small fixes #84

Full Changelog: v1.1.0...v1.2.0

Additional changelogs for Zigbee2MQTT 2.4.0 to 2.5.0 update:

• https://github.com/Nerivec/zigbee2mqtt-windfront/releases/tag/v1.1.0
• https://github.com/Nerivec/zigbee2mqtt-windfront/releases/tag/v1.0.6
• https://github.com/Nerivec/zigbee2mqtt-windfront/releases/tag/v1.0.5
• https://github.com/Nerivec/zigbee2mqtt-windfront/releases/tag/v1.0.4
• https://github.com/Nerivec/zigbee2mqtt-windfront/releases/tag/v1.0.3

---

This discussion was created from the release v1.2.0.

[khaisilk1910]

The header on the phone does not look good.
Please edit it to make it look better.
Thank you very much.

[grebenyuk51]

When using Zigbee2MQTT, the network map displays some devices as "siblings" to the Coordinator, which should not be possible according to Zigbee network topology. In a Zigbee network, the Coordinator is the root node (depth 0) and has no parent, so it cannot have siblings. Devices directly connected to the Coordinator should be its children (depth 1), and devices sharing the same parent are siblings to each other, not to the Coordinator.

This issue was observed in the Zigbee2MQTT network map, where certain devices (e.g., End Devices or Routers) are incorrectly shown as siblings to the Coordinator, often with adepth of N/A, despite functioning correctly in the network. This behavior is unexpected to me.

Morove, the network map sometimes indicates that Device A is a parent of Device B, while Device B is simultaneously shown as a parent of Device A. This is not possible in a Zigbee network, as it violates the hierarchical structure (directed acyclic graph) of the protocol. A node cannot be both a parent and a child of another node, as this would create a circular dependency in depth and addressing.

Radial, +-50 devices

[Nerivec]

The relationship handling is more complex than the original meaning of parent/child/sibling.
For example:

3.6.1.6 - After joining is complete, a joining or rejoining router device will no longer have a parent child relationship with the device it has joined or rejoined to; rather, both devices will have a ‘relationship’ of 'sibling' (0x02).

Also, this information is entirely driven by what devices report (which may include firmware assignment bugs, and/or outdated data that was not properly purged by the device).
That parent/parent your mentioned for example sounds like a firmware bug.

Just to be safe, did you confirm the bug is not a display one, and that the data has the according values (using the state.json from Settings>Tools)?

Note: the data display shows the entire dataset for deeper analysis, while the map display makes the best effort to represent what the network actually looks like based on the given data (with a margin for error of course, since different devices with different stacks may not always align their behavior, not to mention the aforementioned quirks in the data sources).