| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
|
So looks like for a PTR record, the name would be the same as the service type, but would be the full instance name for a TXT. I think it's still safe to have the endswith comparison, but perhaps this difference is where the bug came from. |
Sorry, something went wrong.
|
I triggered a re-run on the tests to see whether the "this message is too long" was just stochastic. If not, I guess maybe it's because a message that was previously skipped is now being accepted, and can't be processed properly. |
Sorry, something went wrong.
|
Looks as if the tests are still failing on MacOS |
Sorry, something went wrong.
There was a problem hiding this comment.
Looks good to me! Though it would be good to get that test passing if we can.
Sorry, something went wrong.
|
Python tests on MacOS have been broken for a while |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
Observed behavior
In Kolibri, when an instance's information is updated, the instance was properly broadcasting that, but the event wasn't being handled by other instances. I confirmed that instance was properly broadcasting updated information.
Solution
The issue appeared to be that the ServiceBrowser, which accepts a specific service type to listen for, was performing an equality comparing on the updated record's name with the service type. But the record name should only end with the service type, and shouldn't match exactly, since there can be multiple devices of the same service on a network.