Cannot load.ivy where a classifier is required

[laurencer]

In sbt you can normally specify the classifier - e.g.

libraryDependencies ++= "net.sf.json-lib" % "json-lib" % "2.4" % "jdk15"

This doesn't work with load.ivy and its not clear how to load a JAR which requires a classifier.

[lihaoyi]

This is fixed as of 662859f, which allows you to use coursier to add a classifier:

Welcome to the Ammonite Repl version-SNAPSHOT
(Scala 2.12.2 Java 1.8.0_112)
If you like Ammonite, please support our development at www.patreon.com/lihaoyi
@ interp.load.ivy(
    // Directly using coursier API. Can pass in exclusions,
    // attributes, configurations, classifiers, etc.
    coursier.Dependency(
      coursier.Module("net.sf.json-lib", "json-lib"),
      "2.4",
      attributes = coursier.Attributes(classifier = "jdk15")
    ),
    "xom" % "xom" % "1.1"
  )
https://repo1.maven.org/maven2/xom/xom/1.1/xom-1.1.pom
  100.0% [##########] 1.7 KiB (1.6 KiB / s)
https://repo1.maven.org/maven2/net/sf/json-lib/json-lib/2.4/json-lib-2.4.pom
  100.0% [##########] 12.8 KiB (11.2 KiB / s)
https://repo1.maven.org/maven2/xom/xom/1.1/xom-1.1.jar
  100.0% [##########] 421.5 KiB (336.9 KiB / s)
https://repo1.maven.org/maven2/net/sf/json-lib/json-lib/2.4/json-lib-2.4-jdk15.jar
  100.0% [##########] 155.4 KiB (120.5 KiB / s)

@ val serializer = new net.sf.json.xml.XMLSerializer
serializer: net.sf.json.xml.XMLSerializer = net.sf.json.xml.XMLSerializer@6d1c0769
@

[lihaoyi]

http://www.lihaoyi.com/Ammonite/#0.9.0 is out

[borice]

@lihaoyi How is this supposed to work with the new version:
Welcome to the Ammonite Repl 2.2.0 (Scala 2.13.3 Java 1.8.0_252)

I see there's now a coursierapi, where Dependency and Module live, but not Attributes.

[borice]

I found something that works.

interp.load.ivy(
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0",
      Dependency.of("edu.stanford.nlp", "stanford-corenlp", "4.2.0").withClassifier("models"),
      Dependency.of("edu.stanford.nlp", "stanford-corenlp", "4.2.0").withClassifier("models-english")
  ) 

would be nice if the syntax would more closely match what's used in SBT for simple copy-paste testing, like:

SBT:
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0",
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0" classifier "models" classifier "models-english"

to translate to:

interp.load.ivy(
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0",
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0" classifier "models" classifier "models-english"
)

or, even if it needs to be duplicated, something like:

interp.load.ivy(
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0",
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0" classifier "models",
      "edu.stanford.nlp"              %  "stanford-corenlp"     % "4.2.0" classifier "models-english"
)

would be nice, if possible.

[wsargent]

Should add import coursierapi._ into that Dependency.of solution as well.