Feature initImportREPL: allow initial imports to be automatically available in repl session

[bjornregnell]

Is your feature request related to a problem? Please describe.

It is tedious and annoying to have to retype import allthethings.* everytime I start the repl with scala-cli --repl .

Describe the solution you'd like

I would like a way to state what initial imports I want to directly access in all repl sessions of the current build. It should be able to do this both with a using-directive and a command-line option.

Describe alternatives you've considered

A more general solution is to provide the corresponding feature of sbt's console / initialCommands which would be more powerful. In sbt you write for example:

console / initialCommands := "import mypack.stuff.*, otherpack.morestuff.*"

I have tried using the scala-cli option --scala-snippet but that does not work with scala-cli version 1.1.2 as it gives this warning and ignores the import:

$ scala-cli repl . --scala-snippet "import reqt.*"
Starting compilation server
Compiling project (Scala 3.3.1, JVM (17))
[warn] ./.scala-build/reqT-solver_6554fc5f28-77219591e6/src_generated/main/scala-snippet.scala:1:1
[warn] No class, trait or object is defined in the compilation unit.
[warn] The incremental compiler cannot record the dependency information in such case.
[warn] Some errors like unused import referring to a non-existent class might not be reported.
[warn] import reqt.*
[warn] ^^^^^^^^^^^^^

Additional context

This has been discussed in both #604 and #2185 - the latter has a wider scope.

[spamegg1]

Yes, a use case would be to have a file like this: for example

// this is imports.scala
//> using toolkit default
//> using dep org.apache.commons:commons-numbers-combinatorics:1.1
//> using dep org.scala-lang.modules::scala-parallel-collections:1.0.4
//> using dep org.scala-graph:graph-core_2.13:2.0.1
// more dependencies if needed

import collection.parallel.CollectionConverters.*
import scala.util.boundary, boundary.break
import org.apache.commons.numbers.combinatorics.{BinomialCoefficient, Factorial}
import scalax.collection.generic.AbstractDiEdge
// more imports if needed

then to start the repl with:

$ scala-cli repl imports.scala

and have the imports accessible in the REPL, for example:

scala> BinomialCoefficient.value(10, 2)
val res0: Long = 45

Although maybe this is not possible due to REPL's own limitations?
How does scala-cli repl --toolkit default do it? When we use that, os.??? is available immediately.

[Gedochao]

How does scala-cli repl --toolkit default do it? When we use that, os.??? is available immediately.

If you invoke any API while prefixing the full package (as os.??? is the full package path), then no imports are necessary.

Although maybe this is not possible due to REPL's own limitations?

@spamegg1 AFAIK the REPL currently does not have the API to deliver this, as well as a couple of other requested features.
However, improving this situation actually is on the 2024 Roadmap for Scala 3. So things might change soon.

[bjornregnell]

See also discussion here: https://users.scala-lang.org/t/load-a-file-on-startup-into-scala-3-repl/9943

[bjornregnell]

Also wished for here: https://users.scala-lang.org/t/how-to-make-an-import-available-on-scala-cli-repl-start-up/10463

[Gedochao]

Delivered in:

• Enable direct usage of --repl-init-script with Scala REPL >= 3.6.4-RC1 #3447
• Partial Fix #21242: Add REPL init script setting scala/scala3#22206

[bjornregnell]

Many thanks for your work on making the REPL even nicer!!!

[benhutchison]

Stretch goal: please update the docs for REPL showing how to import into a repl session.

Even the eminent programmer Claude 3.7 is giving wrong advice about this currently 😑