Where does workflow-trace-viewer data come from? #1440
-
|
Sorry if this is a silly question. I see that code in workflow-trace-viewer reads from a port over ADB, but I don't see where anything in the Workflow library writes to that port. Am I missing something obvious or are library consumers expected to implement the writes themselves from the workflow-tracing APIs? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Hi @Kritarie. This started as an internal project which we have only partially upstreamed into OSS. So the full picture is not yet in this repo, you are right! The 'writing' side is currently still internal, and likely will be until earlyish next year. The summary of that though is it uses a Debug Only WorkflowRuntimeTracer to capture JSON about the workflow nodes being rendered in the form of Node data class. Then this is forwarded both to a .JSON file and to a domain socket. The file is saved so that it can be opened later by the viewer, the domain socket (over ADB) is what can connect live to the viewer. |
Beta Was this translation helpful? Give feedback.
Hi @Kritarie.
This started as an internal project which we have only partially upstreamed into OSS. So the full picture is not yet in this repo, you are right!
The 'writing' side is currently still internal, and likely will be until earlyish next year.
The summary of that though is it uses a Debug Only WorkflowRuntimeTracer to capture JSON about the workflow nodes being rendered in the form of Node data class.
Then this is forwarded both to a .JSON file and to a domain socket. The file is saved so that it can be opened later by the viewer, the domain socket (over ADB) is what can connect live to the viewer.