Repository navigation
[ZHA] Reading attributes in chunk during pairing breaks for manufacturer attrs #51
Description
This logic where we read attributes during pairing in chunks is flawed in multiple ways (regarding manufacturer code usage): https://github.com/zigpy/zha/blob/b17254948690f740a2b795d28dfad5304ad98291/zha/zigbee/cluster_handlers/__init__.py#L587-L590
- some attributes in the
attributeslist can be standard ZCL ones, other can be manufacturer specific: these need to be read separately! - the
cluster_id >= 0xFC00isn't great. Possibly use_is_manuf_specificproperty fromCustomClusters?
This needs to be reworked.
Activity
zigpy-review-bot commentedon Oct 10, 2026
Fixed: zigpy/zha#657 removed the chunked attribute read (including its
cluster_id >= 0xFC00manufacturer heuristic) from the cluster handlers. ZHA now usesCluster.read_attributes, which since zigpy/zigpy#1719 (zigpy 0.91.0) groups reads by manufacturer code before chunking them.Closing as completed.