Imported From: https://issues.scala-lang.org/browse/SI-7916?orig=1
Reporter: @jedesah
Affected Versions: 2.11.0-M8
See #8422

@adriaanm said:
Since 2.11.0-RC1 is one week away, pushing all non-blockers without PR to 2.11.1-RC1. Please undo the change if I missed work in progress.

@gkossakowski said:
The 2.11.2 is out so I'm rescheduling the issue for 2.11.3.

@retronym said:
Updating fix-by version to 2.11.5.

Margaret Leber (MaggieL) said:
@moors So...I'm not familiar with the project workflow but this is still a problem in 2.11.6

@SethTisue said:
re: the status of Scala's JSR 223 support, see #8422

@SethTisue said:
I imagine it would be possible to work around this by casting the ScriptEngine returned to IMain:

val engine = new ScriptEngineManager().getEngineByName("scala")
val iMain = engine.asInstanceOf[scala.tools.nsc.interpreter.IMain]

and then just interact with it directly with there, bypassing the official JSR 223 API. (I don't know IMain well enough to give details offhand, though.)

Margaret Leber (MaggieL) said:
It's good to have a workaround for cases where one controls the source of the calls to ScriptEngineManager.
But let's not forget the whole reason for JSR-223 is to address pluggable scripting components without doing that kind of thing.

@SethTisue said:
Note that put works (as in the example in http://docs.scala-lang.org/scala/2.11/, which works in both 2.11.0 and 2.11.7), and that although get doesn't work, an easy workaround is to use eval instead. Of course, it would be better if get were supported, so I've edited the ticket name to include it.

@som-snytt said:
At some point, I pulled the ScriptEngine support out of IMain into a separate adapter. I'll try to supply a PR, but that would probably be for 2.12. It's not brain surgery, but more like pulling out quills after encountering a porcupine.

When they say, "Scala supports...," they mean, "The Scala team fully supports anyone who'd like to contribute it."

@som-snytt said:
scala/scala#4819

@lrytz said:
[~MaggieL] would you be interested in reviewing the PR scala/scala#4819? If you don't want to get the revision from github and build it from source, I packed a build and put it here: https://dl.dropboxusercontent.com/u/4881107/scala-8876c1d.zip.

Margaret Leber (MaggieL) said:
@lukas Rytz I can't promise to do it immediately; unfortunately my time is not my own. But I will do my best to try to test it for my (quite convoluted) use case as soon as possible.

Thanks for working on this, guys,

Margaret Leber (MaggieL) said:
Fiddling with this now. Dusting off a year-old branch of our app that I'm glad I kept around now. Many hacks need to be re-done...

Margaret Leber (MaggieL) said:
Well, it was a tough fight, but I finally got somewhere! Not my actual use case but a vastly simplified proof-of-concept.

Details here:
scala/scala#4819 (comment)

Margaret Leber (MaggieL) said (edited on Feb 22, 2016 1:55:30 PM UTC):
Unfortunately, I think any approach that erases type information in the objects in a ScriptContext is in for a very rough ride from folks approaching Scala as a JSR223 implemention. It's common for a JSR223 host to chain script executions, sharing their contexts between executions which may occur under different ScriptEngines. Activiti and jMeter are good examples. That said, I suppose it's better than not working at all.

Margaret Leber (MaggieL) said:
Just thought I'd mention here that discussion on the PR has schooled me on why this has to work the way it does, really hoping we can see this merged and available in 2.12

scala/scala#4819 (comment)

@SethTisue said (edited on Mar 11, 2016 10:37:39 PM UTC):
thanks Maggie! it seems nearly certain the PR will be merged once @som-snytt finds time to do a little more work on it.

@som-snytt said:
Wait, there's work?!

@SethTisue said:
isn't there?

@som-snytt said:
I guess there was. Still scala/scala#4819