Two tests for Ruby code generator:

- A golden-file test that ensures protoc produces known-valid output.
- A Ruby test that loads that golden file and ensures it actually works
  with the extension.

This split strategy allows us to test end-to-end without needing to
integrate the Ruby gem build system and the protoc build system. This is
desirable because we do not want a gem build/install to depend on
building protoc, and we do not want building protoc to depend on
building and testing the gem.

1 files changed, 17 insertions, 0 deletions

diff --git a/ruby/tests/generated_code_test.rb b/ruby/tests/generated_code_test.rb
new file mode 100644
index 00000000..daef357a
--- /dev/null
+++ b/ruby/tests/generated_code_test.rb
@@ -0,0 +1,17 @@
+#!/usr/bin/ruby
+
+# generated_code.rb is in the same directory as this test.
+$LOAD_PATH.unshift(File.expand_path(File.dirname(__FILE__)))
+
+require 'generated_code'
+require 'test/unit'
+
+class GeneratedCodeTest < Test::Unit::TestCase
+  def test_generated_msg
+    # just test that we can instantiate the message. The purpose of this test
+    # is to ensure that the output of the code generator is valid Ruby and
+    # successfully creates message definitions and classes, not to test every
+    # aspect of the extension (basic.rb is for that).
+    m = A::B::C::TestMessage.new()
+  end
+end