Skip to content

Chrome importer should import each source with sourceMapURL if it exists #6189

Description

@canova

Currently Chrome importer generates a sourceTable entry when the sample contains some sources. But it looks like now Chrome trace profiles contains ScriptCatchup entries when the it's recorded with disabled-by-default-devtools.v8-source-rundown category (which is enabled by default now despite the name), like this:

{
  "args": {
    "data": {
      "executionContextId": 3,
      "hasSourceUrl": false,
      "isModule": true,
      "isolate": "14261501104024015583",
      "scriptId": 7,
      "sourceMapUrl": "/index-XVVABR7J.js.map",
      "url": "https://profiler.firefox.com/index-XVVABR7J.js"
    }
  },
  "cat": "disabled-by-default-devtools.v8-source-rundown",
  "dur": 2,
  "name": "ScriptCatchup",
  "ph": "X",
  "pid": 39856,
  "tdur": 1,
  "tid": 5694409,
  "ts": 194791085102,
  "tts": 56845
}

Example chrome trace:

Trace-20260715T220108.json.gz

It looks like this entry is pretty similar to our sourceTable entry! We have both url and sourceMapUrl that we need. So we should definitely use these entries to fill the source table with the correct source maps.

┆Issue is synchronized with this Jira Task

Metadata

Metadata

Assignees

Labels

profile dataIssues related to the profile format, data structure, or profile upgraderssource mapsRelated to the source maps feature

Fields

Priority

None yet

Severity

None yet

Jira

None yet

Estimate

None yet

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions