chrome: tolerate missing args data#527
Conversation
src/import/chrome.ts
Outdated
| const id = event.id || pidTid | ||
|
|
||
| if (event.args.data == null) { | ||
| console.log('Ignoring CpuProfile event with no data', event) |
There was a problem hiding this comment.
These logs may be noisy, or not descriptive enough. Let me know!
|
@jefflee-figma Okay, this seems pretty harmless to me Can you add a minimal test to this (file should be as small as possible) to prevent this from regressing in the future? |
Addressed in e180d4e. This is a real-world trace with lots of irrelevant content trimmed out, and some URLs changed to protect the innocent. 🙂 I also added c03112f which removes the log spam. It didn't seem that helpful. |
|
This is almost ready to go -- can you move the test file into the appropriate directory in Thanks! |
1398bd8 to
ac3b0c7
Compare
Yep, thanks for clarifying that. This is addressed in ac3b0c7. I assumed that the directory name should be the major version listed under "About Chrome", but let me know if that is supposed to be something else. |
|
@jefflee-figma This is now live on https://speedscope.app and published to npm as speedscope@1.25.0 |
|
Much appreciated!
…On Wed, Dec 3, 2025 at 10:51 AM Jamie Wong ***@***.***> wrote:
*jlfwong* left a comment (jlfwong/speedscope#527)
<#527 (comment)>
@jefflee-figma <https://github.com/jefflee-figma> This is now live on
https://speedscope.app and published to npm as ***@***.***
—
Reply to this email directly, view it on GitHub
<#527 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BA3RNTFG3VM4DLK6BW4ZWR3374WMVAVCNFSM6AAAAACJXLMOXGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMMBYGMZDEMRRGU>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
This change allows Speedscope to tolerate Chrome profile chunks where
args.datais undefined. It's unclear whether this is a Chrome bug, but we are noticing recent profiles that crash Speedscope without this tweak. FWIW, it also appears that Firefox seems to die on a similar issue.Happy to provide an example trace that repros, but out of an abundance of caution I'd prefer to email rather than post publicly. Feel free to shoot me an email: jefflee@figma.com