SI-8062 Fix inliner cycle with recursion, separate compilation

ICodeReaders, which decompiles JVM bytecode to ICode, was not
setting the `recursive` attribute of `IMethod`. This meant that
the inliner got into a cycle, repeatedly inlining the recursive
call.

The method name `filter` was needed to trigger this as the inliner
heuristically treats that as a more attractive inlining candidate,
based on `isMonadicMethod`.

This commit:

  - refactors the checking / setting of `virtual`
  - adds this to ICodeReaders
  - tests the case involving `invokevirtual`

I'm not sure how to setup a test that fails without the other changes
to `ICodeReader` (for invokestatic and invokespecial).

3 files changed, 9 insertions, 0 deletions

diff --git a/test/files/pos/t8062.flags b/test/files/pos/t8062.flags
new file mode 100644
index 0000000000..49d036a887
--- /dev/null
+++ b/test/files/pos/t8062.flags
@@ -0,0 +1 @@
+-optimize
diff --git a/test/files/pos/t8062/A_1.scala b/test/files/pos/t8062/A_1.scala
new file mode 100644
index 0000000000..ca0411dae8
--- /dev/null
+++ b/test/files/pos/t8062/A_1.scala
@@ -0,0 +1,5 @@
+package warmup
+
+object Warmup {
+  def filter[A](p: Any => Boolean): Any = filter[Any](p)
+}
diff --git a/test/files/pos/t8062/B_2.scala b/test/files/pos/t8062/B_2.scala
new file mode 100644
index 0000000000..f0a6761488
--- /dev/null
+++ b/test/files/pos/t8062/B_2.scala
@@ -0,0 +1,3 @@
+object Test {
+  warmup.Warmup.filter[Any](x => false)
+}