avoid java.net.URL

[jtjeferreira]

I was using sbt without internet connection and noticed that update task taking a lot of time (around 30s). I did a thread dump with a SigQuit ( Ctrl + \) and noticed that the thread was blocked because java.net.URL#equals blocks:

"pool-514-thread-3" #918 prio=5 os_prio=31 cpu=292.62ms elapsed=128.20s tid=0x000000012d484a00 nid=0x1071b waiting for monitor entry  [0x0000000175550000]
   java.lang.Thread.State: BLOCKED (on object monitor)
	at java.net.InetAddress$NameServiceAddresses.get([email protected]/InetAddress.java:838)
	- waiting to lock <0x0000000737d5e758> (a java.net.InetAddress$NameServiceAddresses)
	at java.net.InetAddress.getAllByName0([email protected]/InetAddress.java:1533)
	at java.net.InetAddress.getAllByName([email protected]/InetAddress.java:1385)
	at java.net.InetAddress.getAllByName([email protected]/InetAddress.java:1306)
	at java.net.InetAddress.getByName([email protected]/InetAddress.java:1256)
	at java.net.URL.getHostAddress([email protected]/URL.java:831)
	- locked <0x00000007073fd680> (a java.net.URL)
	at java.net.URLStreamHandler.getHostAddress([email protected]/URLStreamHandler.java:447)
	at java.net.URLStreamHandler.hostsEqual([email protected]/URLStreamHandler.java:459)
	at java.net.URLStreamHandler.sameFile([email protected]/URLStreamHandler.java:431)
	at java.net.URLStreamHandler.equals([email protected]/URLStreamHandler.java:352)
	at java.net.URL.equals([email protected]/URL.java:980)
	at scala.runtime.BoxesRunTime.equals2(BoxesRunTime.java:137)
	at scala.runtime.BoxesRunTime.equals(BoxesRunTime.java:123)
	at scala.Some.equals(Option.scala:513)
	at sbt.librarymanagement.Artifact.equals(Artifact.scala:22)
	at scala.runtime.BoxesRunTime.equals2(BoxesRunTime.java:137)
	at scala.runtime.BoxesRunTime.equals(BoxesRunTime.java:123)
	at scala.Tuple2.equals(Tuple2.scala:24)
	at scala.runtime.BoxesRunTime.equals2(BoxesRunTime.java:137)
	at scala.runtime.BoxesRunTime.equals(BoxesRunTime.java:123)
	at scala.collection.IterableLike.sameElements(IterableLike.scala:304)
	at scala.collection.IterableLike.sameElements$(IterableLike.scala:294)
	at scala.collection.AbstractIterable.sameElements(Iterable.scala:56)
	at scala.collection.GenSeqLike.equals(GenSeqLike.scala:481)
	at scala.collection.GenSeqLike.equals$(GenSeqLike.scala:480)
	at scala.collection.AbstractSeq.equals(Seq.scala:45)
	at sbt.librarymanagement.ModuleReport.equals(ModuleReport.scala:34)
	at scala.runtime.BoxesRunTime.equals2(BoxesRunTime.java:137)
	at scala.runtime.BoxesRunTime.equals(BoxesRunTime.java:123)
	at scala.collection.IterableLike.sameElements(IterableLike.scala:304)
	at scala.collection.IterableLike.sameElements$(IterableLike.scala:294)
	at scala.collection.AbstractIterable.sameElements(Iterable.scala:56)
	at scala.collection.GenSeqLike.equals(GenSeqLike.scala:481)
	at scala.collection.GenSeqLike.equals$(GenSeqLike.scala:480)
	at scala.collection.AbstractSeq.equals(Seq.scala:45)
	at sbt.librarymanagement.ConfigurationReport.equals(ConfigurationReport.scala:21)

In this PR I tried to replace java.net.URL with java.net.URI in a backwards compatible manner in Artifact, Developer, and ScmInfo.

Although I managed to make mima happy, I think there is a problem in the Artifact#apply method because it changed from having a parameter url: Option[java.net.URL] to uri: Option[java.net.URI] (which erases to Option and makes mima happy). I also cant add a new backwards compatible apply method because it would erase to the same signature.

Any ideas how to move forward?

[jtjeferreira]

@adpi2 I see you are working on #442. Do you think I can drop binary compatibility here?

[adpi2]

@adpi2 I see you are working on #442. Do you think I can drop binary compatibility here?

Yes, if you open a new PR to wip/sbt-2.x.

In sbt 1.x you should try not to break it and if you cannot, we should make sure it won't break any plugin out there.

[jtjeferreira]

@adpi2 I see you are working on #442. Do you think I can drop binary compatibility here?

Yes, if you open a new PR to wip/sbt-2.x.

I will do that

if you cannot, we should make sure it won't break any plugin out there.

how can I check if the change will break other plugins? There is some kind of community build?

[jtjeferreira]

@adpi2 I opened #443