Open Apollon77 opened 5 months ago
Under "7.19.1.9. Struct", in the "Fabric-Scoped Struct" section, the text says:
The global FabricIndex field of a fabric-scoped struct SHOULD NOT be indicated in a write interaction.
The global FabricIndex field of a fabric-scoped struct SHALL be ignored in a write interaction.
It looks like this has been inconsistent all along, actually. The intent of the spec is what 7.19.1.9 says, so 7.13.7 needs to be fixed. I filed https://github.com/CHIP-Specifications/connectedhomeip-spec/issues/9213 to track this, but this is not an SDK issue so closing this.
That said, ideally the Python bits would follow the SHOULD NOT and not send the FabricIndex.... chip-tool does not do that when executing the relevant YAML, for example.
Thank you @bzbarsky-apple ... ok in fact I add logic then in matter.,js to always overwrite the fabricIndex, irrelevant from what comes in
Test issue(s)
While executing python test ACE_1_2.py I found that the acl entries written by Python contained an fabricIndex field with value 0. see https://github.com/project-chip/matter.js/actions/runs/9092348098/job/24988983271#step:4:1080
The value is really encoded in the Tlv data.
According to specifications 7.13.7. FabricIndex Field:
I do not know if this is a python test issue or a more general issue.
Platform
all
Anything else?
No response