[ruby/irb] Refactor eval_path and `SourceFinder::Source`

(https://github.com/ruby/irb/pull/870)

* Assign `@eval_path` through `irb_path=` method

This simplifies the original caching logic for the `eval_path` method
and makes it easier to understand.

* Refactor SourceFinder::Source

https://github.com/ruby/irb/commit/c63e4c4035

1 files changed, 26 insertions, 16 deletions

diff --git a/lib/irb/context.rb b/lib/irb/context.rb
index 814a8bd4ad..e50958978f 100644
--- a/lib/irb/context.rb
+++ b/lib/irb/context.rb
@@ -77,7 +77,7 @@ module IRB
       else
         @irb_name = IRB.conf[:IRB_NAME]+"#"+IRB.JobManager.n_jobs.to_s
       end
-      @irb_path = "(" + @irb_name + ")"
+      self.irb_path = "(" + @irb_name + ")"
 
       case input_method
       when nil
@@ -121,11 +121,11 @@ module IRB
       when '-'
         @io = FileInputMethod.new($stdin)
         @irb_name = '-'
-        @irb_path = '-'
+        self.irb_path = '-'
       when String
         @io = FileInputMethod.new(input_method)
         @irb_name = File.basename(input_method)
-        @irb_path = input_method
+        self.irb_path = input_method
       else
         @io = input_method
       end
@@ -246,9 +246,27 @@ module IRB
     # Can be either name from <code>IRB.conf[:IRB_NAME]</code>, or the number of
     # the current job set by JobManager, such as <code>irb#2</code>
     attr_accessor :irb_name
-    # Can be either the #irb_name surrounded by parenthesis, or the
-    # +input_method+ passed to Context.new
-    attr_accessor :irb_path
+
+    # Can be one of the following:
+    # - the #irb_name surrounded by parenthesis
+    # - the +input_method+ passed to Context.new
+    # - the file path of the current IRB context in a binding.irb session
+    attr_reader :irb_path
+
+    # Sets @irb_path to the given +path+ as well as @eval_path
+    # @eval_path is used for evaluating code in the context of IRB session
+    # It's the same as irb_path, but with the IRB name postfix
+    # This makes sure users can distinguish the methods defined in the IRB session
+    # from the methods defined in the current file's context, especially with binding.irb
+    def irb_path=(path)
+      @irb_path = path
+
+      if File.exist?(path)
+        @eval_path = "#{path}(#{IRB.conf[:IRB_NAME]})"
+      else
+        @eval_path = path
+      end
+    end
 
     # Whether multiline editor mode is enabled or not.
     #
@@ -557,7 +575,7 @@ module IRB
 
       if IRB.conf[:MEASURE] && !IRB.conf[:MEASURE_CALLBACKS].empty?
         last_proc = proc do
-          result = @workspace.evaluate(line, eval_path, line_no)
+          result = @workspace.evaluate(line, @eval_path, line_no)
         end
         IRB.conf[:MEASURE_CALLBACKS].inject(last_proc) do |chain, item|
           _name, callback, arg = item
@@ -568,20 +586,12 @@ module IRB
           end
         end.call
       else
-        result = @workspace.evaluate(line, eval_path, line_no)
+        result = @workspace.evaluate(line, @eval_path, line_no)
       end
 
       set_last_value(result)
     end
 
-    private def eval_path
-      # We need to use differente path to distinguish source_location of method defined in the actual file and method defined in irb session.
-      if !defined?(@irb_path_existence) || @irb_path_existence[0] != irb_path
-        @irb_path_existence = [irb_path, File.exist?(irb_path)]
-      end
-      @irb_path_existence[1] ? "#{irb_path}(#{IRB.conf[:IRB_NAME]})" : irb_path
-    end
-
     def inspect_last_value # :nodoc:
       @inspect_method.inspect_value(@last_value)
     end