There has been a regression on most inspector open tests.

• cold.open: +4%
• simple.open: +10%
• custom.open: +4%
• complicated.open: +7%

Alert is https://treeherder.mozilla.org/perf.html#/alerts?id=23087
According to this alert, we should look at https://bugzilla.mozilla.org/show_bug.cgi?id=1500142 (remove inspector toolbox bindings)

But there is also https://bugzilla.mozilla.org/show_bug.cgi?id=1572980 (use reps.js intead of old-reps) which landing very close to the alert, and also made the inspector load 10% more data. Could also have an impact.

I think the regression may be Bug 1494680 (which is the one Julian wanted to put here I think, but linked to another, test-only related)

The need to non-lazy load is because we need the object inspector reducer (which is small, but bundled in the reps.js file)
We should check if we can only import the reducer itself (and lazy load reps.js)
We may have to remove flow usage from the objectInspector for that (because we can't transpile in other places than in the debugger for now)
Might be a good opportunity to move to the typescript annotation syntax greg was talking about.

(In reply to Nicolas Chevobbe [:nchevobbe] from comment #1)

I think the regression may be Bug 1494680 (which is the one Julian wanted to put here I think, but linked to another, test-only related)

Thanks! How did I even get to this bug number, I have no idea!
The good news is that it doesn't have anything to do with your bug :)

The regression is definitely from Bug 1500142.
Kicked off a few try runs:

• baseline: https://treeherder.mozilla.org/#/jobs?repo=try&revision=7cd2eded42e1cb236cf46a855456eb66b975dd48
• fix for reps loading: https://treeherder.mozilla.org/#/jobs?repo=try&revision=d938bf12bd0f4700448005514323b0822179d0d7
• fix for reps + backout of 1500142: https://treeherder.mozilla.org/#/jobs?repo=try&revision=3497956f4e0a258c8a344968ee0140a639f2ece1

Comparing "fix for reps + backout of 1500142" as new vs "fix for reps loading" as old:
https://treeherder.mozilla.org/perf.html#/comparesubtest?originalProject=try&newProject=try&newRevision=3497956f4e0a258c8a344968ee0140a639f2ece1&originalSignature=1759151&newSignature=1759151&framework=12&originalRevision=d938bf12bd0f4700448005514323b0822179d0d7

That means that backing out Bug 1500142 fixed the performance issue.