- Generated by
1.9.8
|
Motion Master 6.0.0-alpha.86
Next-generation motion control software
|
The complete state of a procedure on one device — everything a client polling GET /api/devices/:pos/procedures/:name receives.
More...
#include <procedure.h>
Public Attributes | |
| ProcedureStatus | status = ProcedureStatus::kIdle |
| The one field a polling loop checks. | |
| uint32_t | runCount = 0 |
| How many runs this device accepted since the last rescan (a rejected start does not count). | |
| std::optional< int64_t > | startedAt |
| Epoch ms the run began; absent when never run. | |
| std::optional< int64_t > | finishedAt |
| Epoch ms the run ended; absent while running. | |
| std::vector< ProgressStep > | steps |
| The ordered step array, always the full template. | |
| std::optional< std::string > | error |
| Why the run failed, when no step captured it. | |
The complete state of a procedure on one device — everything a client polling GET /api/devices/:pos/procedures/:name receives.
This is an accumulating snapshot, not an event: every finished step keeps its terminal status and its measured value for as long as the snapshot is retained. That is what makes polling lossless — a step that both starts and finishes between two polls is still visible as kSucceeded with its value in the next one, so a client that never opens a WebSocket cannot miss a result. Only the transient kRunning blip on a fast step can be skipped, and it carries no data.
A device that never ran the procedure still yields a well-formed snapshot — kIdle, runCount 0, no timestamps, every step kIdle from the template — so a client renders one component with no empty-state special case.
| std::optional<std::string> mm::node::ProcedureSnapshot::error |
Why the run failed, when no step captured it.
A body normally fails the step that went wrong, so the reason is visible where it happened and this stays empty. It is for the failure that belongs to no step — the device turning out not to be the right kind, say — which would otherwise leave a run marked failed with nothing anywhere saying why.
| std::optional<int64_t> mm::node::ProcedureSnapshot::finishedAt |
Epoch ms the run ended; absent while running.
| uint32_t mm::node::ProcedureSnapshot::runCount = 0 |
How many runs this device accepted since the last rescan (a rejected start does not count).
Doubles as a generation counter, which is what makes polling safe without a run id: a run started elsewhere that begins and finishes between two polls would otherwise be invisible, and its result silently read as the run being watched. A changed runCount says "this is a
different run"; (devicePosition, procedureName, runCount) identifies one uniquely.
| std::optional<int64_t> mm::node::ProcedureSnapshot::startedAt |
Epoch ms the run began; absent when never run.
| ProcedureStatus mm::node::ProcedureSnapshot::status = ProcedureStatus::kIdle |
The one field a polling loop checks.
| std::vector<ProgressStep> mm::node::ProcedureSnapshot::steps |
The ordered step array, always the full template.
1.9.8