Jamie Wong

RE: small chunks, they are more cache-optimal, but they also result in more total network bandwidth needed because compression is better for larger chunks because of de-duplication of content between...

Not too sure offhand what would cause that, but I'm open to accepting PRs that fix the problem!

Does changing it to "line-profiler" fix the problem for you?

If you'd like to submit a pull request to make the fix, I can test it out locally and merge if it works

Hi @KAJAL-GOSALIYA! Do you have a sample trace you can share? Does this happen for all profiles imported in this way? Import from Instrument 10 is pretty shaky right now...

Ahh, yeah -- this is pretty broken by Instruments 10 in ways I haven't figured out yet. See https://github.com/jlfwong/speedscope/issues/196#issuecomment-471421141 for more context. I'm not personally invested in fixing this bug...

I'm trying to build the `perf` output for the test file provided in #393. To do so, in the `/sample/programs/cpp` directory, I run `make linux-perf`, which spins up a docker...

@javierhonduco Ahhh thanks for the catch! I now have it generating a valid profile. Thank you!

Interesting! Perhaps it would work to retry requests that failed with `mode: no-cors` (see https://developer.mozilla.org/en-US/docs/Web/API/WindowOrWorkerGlobalScope/fetch). Here's the relevant code: https://github.com/jlfwong/speedscope/blob/b7806a1c5f079efb4e2d5c9cb0b7d3f1346b0e82/src/views/application.tsx#L520-L527

I would accept a PR to fix this assuming it doesn't always force fetching without cookies. Fetching with cookies is an important use case for people who want to embed...