Support backtrace after method calls

GitHub: fix #2902, #2917

The current implementation traverses stack to retrieve backtrace. But
stack will be changed when some operations are occurred. It means that
backtrace may be broken after some operations.

This change (1) saves the minimum information to retrieve backtrace when
exception is raised and (2) restores backtrace from the minimum
information when backtrace is needed. It reduces overhead for creating
backtrace Ruby objects.

The space for the minimum information is reused by multiple
exceptions. So memory allocation isn't occurred for each exception.

1 files changed, 29 insertions, 6 deletions

diff --git a/test/t/exception.rb b/test/t/exception.rb
index d27813028..b54d01c46 100644
--- a/test/t/exception.rb
+++ b/test/t/exception.rb
@@ -373,12 +373,35 @@ assert('Raise in ensure') do
   end
 end
 
-assert('Raise in rescue') do
-  assert_raise(ArgumentError) do
-    begin
-      raise "" # RuntimeError
-    rescue
-      raise ArgumentError
+assert('GC in rescue') do
+  line = nil
+  begin
+    [1].each do
+      [2].each do
+        [3].each do
+          line = __LINE__; raise "XXX"
+        end
+      end
+    end
+  rescue => exception
+    GC.start
+    assert_equal("#{__FILE__}:#{line}:in Object.call",
+                 exception.backtrace.first)
+  end
+end
+
+assert('Method call in rescue') do
+  line = nil
+  begin
+    [1].each do
+      [2].each do
+        line = __LINE__; raise "XXX"
+      end
+    end
+  rescue => exception
+    [3].each do
     end
+    assert_equal("#{__FILE__}:#{line}:in Object.call",
+                 exception.backtrace.first)
   end
 end