Merge pull request #3858 from eregon/various-fixes

Add Prism::Source#line_to_byte_offset, optimize Prism::Node#tunnel and fix docs of BlockNode

Author: Earlopain

config.yml

@@ -1269,17 +1269,17 @@ nodes:
       - name: opening_loc
         type: location
         comment: |
-          Represents the location of the opening `|`.
+          Represents the location of the opening `{` or `do`.
 
               [1, 2, 3].each { |i| puts x }
-                               ^
+                             ^
       - name: closing_loc
         type: location
         comment: |
-          Represents the location of the closing `|`.
+          Represents the location of the closing `}` or `end`.
 
               [1, 2, 3].each { |i| puts x }
-                                 ^
+                                          ^
     comment: |
       Represents a block of ruby code.
 

lib/prism/parse_result.rb

@@ -76,6 +76,15 @@ def slice(byte_offset, length)
       source.byteslice(byte_offset, length) or raise
     end
 
+    # Converts the line number to a byte offset corresponding to the start of that line
+    def line_to_byte_offset(line)
+      l = line - @start_line
+      if l < 0 || l >= offsets.size
+        raise ArgumentError, "line #{line} is out of range"
+      end
+      offsets[l]
+    end
+
     # Binary search through the offsets to find the line number for the given
     # byte offset.
     def line(byte_offset)

sig/prism/parse_result.rbs

@@ -14,6 +14,7 @@ module Prism
     def encoding: () -> Encoding
     def lines: () -> Array[String]
     def slice: (Integer byte_offset, Integer length) -> String
+    def line_to_byte_offset: (Integer line) -> Integer
     def line: (Integer byte_offset) -> Integer
     def line_start: (Integer byte_offset) -> Integer
     def line_end: (Integer byte_offset) -> Integer

templates/lib/prism/node.rb.erb

@@ -184,24 +184,13 @@ module Prism
       queue = [self] #: Array[Prism::node]
       result = [] #: Array[Prism::node]
 
+      search_offset = source.line_to_byte_offset(line) + column
+
       while (node = queue.shift)
         result << node
 
         node.each_child_node do |child_node|
-          child_location = child_node.location
-
-          start_line = child_location.start_line
-          end_line = child_location.end_line
-
-          if start_line == end_line
-            if line == start_line && column >= child_location.start_column && column < child_location.end_column
-              queue << child_node
-              break
-            end
-          elsif (line == start_line && column >= child_location.start_column) || (line == end_line && column < child_location.end_column)
-            queue << child_node
-            break
-          elsif line > start_line && line < end_line
+          if child_node.start_offset <= search_offset && search_offset < child_node.end_offset
             queue << child_node
             break
           end

test/prism/ruby/source_test.rb

@@ -0,0 +1,47 @@
+# frozen_string_literal: true
+
+require_relative "../test_helper"
+
+module Prism
+  class SourceTest < TestCase
+    def test_line_to_byte_offset
+      parse_result = Prism.parse(<<~SRC)
+      abcd
+      efgh
+      ijkl
+      SRC
+      source = parse_result.source
+
+      assert_equal 0, source.line_to_byte_offset(1)
+      assert_equal 5, source.line_to_byte_offset(2)
+      assert_equal 10, source.line_to_byte_offset(3)
+      assert_equal 15, source.line_to_byte_offset(4)
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(5) }
+      assert_equal "line 5 is out of range", e.message
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(0) }
+      assert_equal "line 0 is out of range", e.message
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(-1) }
+      assert_equal "line -1 is out of range", e.message
+    end
+
+    def test_line_to_byte_offset_with_start_line
+      parse_result = Prism.parse(<<~SRC, line: 11)
+      abcd
+      efgh
+      ijkl
+      SRC
+      source = parse_result.source
+
+      assert_equal 0, source.line_to_byte_offset(11)
+      assert_equal 5, source.line_to_byte_offset(12)
+      assert_equal 10, source.line_to_byte_offset(13)
+      assert_equal 15, source.line_to_byte_offset(14)
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(15) }
+      assert_equal "line 15 is out of range", e.message
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(10) }
+      assert_equal "line 10 is out of range", e.message
+      e = assert_raise(ArgumentError) { source.line_to_byte_offset(9) }
+      assert_equal "line 9 is out of range", e.message
+    end
+  end
+end