I see claims in KEEPs or YouTrack issues from the ...
# random
y
I see claims in KEEPs or YouTrack issues from the Kotlin team that xyz code feature is barely used based on checking millions of lines of open-source code on GitHub. How are such statistics made? I'd really like to run such analysis on my own! Usually it's a frontend feature like
catch(_: Exception)
for instance. Is this done using the Kotlin ANTLR grammar maybe? I'd love to know
m
Typically, it's a combination of a GitHub search for simple regex requests and our users' project base for more complex analysis that requires a compiler frontend. The user project base includes many of our internal projects, such as IntelliJ IDEA, YouTrack, our libraries, and a bunch of external ones to cover different frameworks. We can build these projects against a custom compiler branch, so it's easy to validate a change or write a diagnostic to find some patterns
y
Is the users' project base public somewhere?
m
No, it includes also private repositories