Markus Stange

I don't think this is ready to be worked on just yet. For now it's just my opinion; I would like to get input from others about how this problem...

> I ran into this problem and **it makes me question whether I know how to interpret the profiler or not.** In particular, I noticed the flame graph visualization appears...

Here's another proposal for showing sample gaps in the stack chart:

You can also give this a try: https://github.com/mstange/fxprof-perf-convert It's not finished yet but it should work for most use cases.

I think it would be good to handle source maps like we handle symbolication: Give the raw profile to profiler.firefox.com first, and then resolve source maps asynchronously, with help from...

I see, interesting. Well, the browser has to have fetched this JS before, because it's executing it. Maybe there's some efficient way to have that existing JS parsing detect source...

Indeed there is! The JS engine already detects source map URLs. They can be queried using [`ScriptSource::sourceMapURL`](https://searchfox.org/mozilla-central/rev/86c98c486f03b598d0f80356b69163fd400ec8aa/js/src/vm/JSScript.h#1021-1022). So we can put those source map URLs into the profile JSON.

Here's one proposal: We could have **both** a `weight` column **and** a `duration` column on the samples table. For on-CPU samples, the `weight` would be 1 and the `duration` would...

> When you select a time range, the numbers in the call tree and flame graph "jump" as you drag past the weighted samples. The "blocked range" is included in...

Here's what this would look like for an example profile. This example profile declares an intended sampling interval of 1ms but ends up sampling every 1.4ms. ## All samples ```...