Perf regression with node module indexing

Description

On master, I'm observing that, using Figwheel, changes to source take about 12 seconds to appear in a React Native app running in an iOS simulator, whereas with the 1.9.671 such changes take about 2 seconds.

I apologize for not having a minimal repro available. I suspect it is easily reproducible locally using a setup generated using re-natal init FutureApp (see https://github.com/drapanjanas/re-natal)

Of note: I don't have :npm-deps set. My compiler options are:

1 2 3 4 5 6 {:output-to "target/ios/not-used.js" :main "env.ios.main" :output-dir "target/ios" :optimizations :none :checked-arrays false :warnings true}

A git bisect shows that this was introduced with the change in CLJS-2212.

Environment

None

Status

Assignee

David Nolen

Reporter

Mike Fikes

Approval

Accepted

Patch

Code

Fix versions

Affects versions

1.9.854

Priority

Blocker
Configure