Correlating Trace Data with Code
What you'll learn: How to map slow type instantiations in trace files back to the exact source code responsible, enabling targeted performance fixes.
The Detective Work
You've generated trace data using --generateTrace and identified expensive type operations. Now comes the crucial step: connecting those abstract performance hotspots to actual lines of code you can optimize.
Think of trace data as a crime scene report—it tells you what happened (type instantiation took 500ms), but you need to find where in your source files the culprit resides.
Reading Type Names in Traces
In Chrome DevTools or @typescript/analyze-trace, you'll see entries like:
checkTypeRelatedTo: 450ms
Type: ConditionalType_123
Related to: MappedType_456
These cryptic type IDs are your breadcrumbs. The challenge is mapping them to source locations.
The Correlation Process
Look for source file references in the trace events—many entries include file paths and positions. DevTools flame charts let you click into events to see which .ts file triggered the check.
Search for distinctive type names. If your trace shows DeepPartial<UserConfig> is expensive, grep your codebase for where DeepPartial is used with UserConfig.
Cross-reference instantiation counts. If a type is instantiated 10,000 times, look for places in tight loops or heavily-called functions where that type appears.
Narrowing Down the Location
Start with the most expensive single operations from your trace analysis. Open the referenced file and examine:
- Complex generic constraints near that line
- Recursive mapped types
- Large union operations in type parameters
- Conditional types with deep nesting
Key Takeaway: Effective trace analysis requires mapping performance metrics back to specific source code patterns—find the expensive type, locate its usage site, then refactor that exact code for gains.