[ruby/prism] Fix parser translator with trailing backslash in %W /%I array

https://docs.ruby-lang.org/en/master/syntax/literals_rdoc.html#label-25w+and+-25W-3A+String-Array+Literals
> %W allow escape sequences described in Escape Sequences. However the continuation line <newline> is not usable because it is interpreted as the escaped newline described above.

https://github.com/ruby/prism/commit/f5c7460ad5

lib/prism/translation/parser/lexer.rb

@@ -425,7 +425,12 @@ module Prism
                   end
 
                   current_string << unescape_string(value, quote_stack.last)
-                  if (backslash_count = token.value[/(\\{1,})\n/, 1]&.length).nil? || backslash_count.even? || !interpolation?(quote_stack.last)
+                  relevant_backslash_count = if quote_stack.last.start_with?("%W", "%I")
+                                               0 # the last backslash escapes the newline
+                                             else
+                                               token.value[/(\\{1,})\n/, 1]&.length || 0
+                                             end
+                  if relevant_backslash_count.even? || !interpolation?(quote_stack.last)
                     tokens << [:tSTRING_CONTENT, [current_string, range(start_offset, start_offset + current_length)]]
                     break
                   end

test/prism/fixtures/strings.txt

@@ -99,6 +99,34 @@ bar)
  d   
 ]
 
+%w[
+  foo\nbar baz\n\n\
+  bat\n\\\n\foo
+]
+
+%W[
+  foo\nbar baz\n\n\
+  bat\n\\\n\foo
+]
+
+%w[foo\
+  bar
+  baz\\
+  bat
+  1\n
+  2
+  3\\n
+]
+
+%W[foo\
+  bar
+  baz\\
+  bat
+  1\n
+  2
+  3\\n
+]
+
 %W[f\u{006f 006f}]
 
 %W[a b#{c}d e]