| label | name | web | github | origin | contributors | status | type | active | priority | category | hide-from-homepage | description | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
scalacProfiling |
Exposing Compilation Performance |
|
Completed |
project |
false |
-1 |
tooling |
true |
Enabling statistics in the compiler and creating the infrastructure around it. |
Read on how to use scalac-profiling to speed up your compile times in this
scala-lang blog post.
The goal of this proposal is to expose the cost that Scala code has in Scalac's compile times, and display it to users via a console and an interactive environment.
From the proposal:
"As a user of the Scala compiler investigating slow build times is difficult. We propose enhancing the current flags and tooling within the Scala compiler to identify hotspots in or caused by users' code. For example a macro may be significant or called more often than expected (e.g. in implicit resolution). Providing tooling that generates user-understandable reports e.g. per macro times called and total times or identifying poor implicit resolution allows users to tune their builds for optimal performance."
The proposal focuses on three areas:
- Data collection (which requires changes to the Scala compiler.)
- Data visualisation and comparison (which requires a compiler plugin + an IDE integration)
- Reproducibility (which requires the development of an sbt plugin)
All the changes to the Scala compiler have been merged upstream as of 2.12.5.
This project is now completed. More information is available in the scalac-profiling repository.