Skip to content
Workflow

Investigating an exception

Exceptions tell you that something failed. TraceMind helps you see which execution path led there: inside the recording, beside watches and probes, not as an isolated stack dump.

When to use this workflow

  • NullReference or InvalidOperation during a specific reproduction
  • Exception only appears after a sequence of actions
  • Stack trace points at a symptom line, not the originating bad state
  • You need to share the failure path with someone who was not in Play Mode

Prerequisites

  • Capture Exceptions enabled in Preferences (on by default in typical setups)
  • Tracked methods along the suspected path (otherwise Exception Flow may be partial)
  • A Scenario that reproduces the throw reliably

Overview

Record → Report → Exception in Tree/Timeline → Exception Flow → Method Flow → MindBoard

Step 1: Prepare the Scenario

  1. Track types involved in the failure path (owner + likely callers).
  2. Add Watches on state that might be null or invalid at throw time.
  3. Enable Probes if scene/object lifecycle could cause null state.
  4. Use Standard or Deep Profile. Minimal may omit detail you need.

Step 2: Record through the failure

  1. Start Recording.
  2. Reproduce until the exception occurs (or the invalid state immediately before it).
  3. Stop Recording.

Check the Scenario Console panel for recording messages if the throw did not appear in the Report.

Step 3: Locate the exception in the Report

Open the Report and find the exception via:

  • Tree: exception nodes in the hierarchy
  • Timeline: Events module markers
  • Flow: exception summaries on scope nodes

Note the time and scope of the throw.

Step 4: Open Exception Flow

From the exception context (Tree or related drill-down), open Exception Flow.

  1. Use Previous exception / Next exception if multiple throws exist.
  2. Run Play / Pause path toward exception to walk the reconstructed path.
  3. Read the Inspection pane for the selected node.

Partial chains are possible

Exception Flow may show limitation notes when data is incomplete, for example: - Call chain reconstructed from exception origin only - Event graph unavailable - No tracked method linked to this exception - Caller not found in recorded executions

Treat these as guidance to widen Tracking and re-record, not as product failure.

Step 5: Correlate with Watches and Probes

At the exception time:

  • Watch Studio: what values were held?
  • Probe Map: was an object destroyed or scene unloaded just before?
  • Timeline: what spiked in Methods or Execution modules?

Often the fix is state that went bad several steps before the throw line.

Step 6: Open Method Flow when structure matters

For a specific scope implicated in Exception Flow, open Method Flow to see call graph layout, filters, and class grouping options.

Export a PNG if the visual helps a ticket or review.

Step 7: Capture for the team

  1. Add to MindBoard: Exception reference + related Watch/Probe references.
  2. Note: root cause hypothesis in plain language.
  3. Export HTML or Markdown if the Report must attach to an issue tracker.

If Exception Flow is thin

Limitation message Action
No tracked method linked Track the methods on the stack path
Caller not found Track the caller's type
Origin only Add Watches on state leading into the throw; enable probes for lifecycle
No exception in Report Confirm Capture Exceptions in Preferences; confirm throw during recording

Method and Exception Flow Reports Troubleshooting