From 01fa1513f4bd6f194bf6d1ca17e510003fa23312 Mon Sep 17 00:00:00 2001
From: Yorick Peterse <yorickpeterse@gmail.com>
Date: Sat, 17 Sep 2016 14:49:23 +0200
Subject: [PATCH] Lexing of processing instructions with namespaces

This adds lexing/parsing support for processing instructions that
contain namespace prefixes such as `<?foo:bar ?>`.
---
 ext/ragel/base_lexer.rl                            | 2 +-
 spec/oga/xml/lexer/processing_instruction_spec.rb  | 8 ++++++++
 spec/oga/xml/parser/processing_instruction_spec.rb | 7 +++++++
 3 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/ext/ragel/base_lexer.rl b/ext/ragel/base_lexer.rl
index d5f8070..0e552b4 100644
--- a/ext/ragel/base_lexer.rl
+++ b/ext/ragel/base_lexer.rl
@@ -161,7 +161,7 @@
     # instruction.
     #
 
-    proc_ins_start = '<?' identifier;
+    proc_ins_start = '<?' identifier (':' identifier)?;
     proc_ins_end   = '?>';
 
     # Everything except "?" OR a single "?"
diff --git a/spec/oga/xml/lexer/processing_instruction_spec.rb b/spec/oga/xml/lexer/processing_instruction_spec.rb
index 0803032..46219da 100644
--- a/spec/oga/xml/lexer/processing_instruction_spec.rb
+++ b/spec/oga/xml/lexer/processing_instruction_spec.rb
@@ -70,6 +70,14 @@ describe Oga::XML::Lexer do
       ]
     end
 
+    it 'lexes an instruction with a namespace prefix' do
+      lex('<?foo:bar?>').should == [
+        [:T_PROC_INS_START, nil, 1],
+        [:T_PROC_INS_NAME, 'foo:bar', 1],
+        [:T_PROC_INS_END, nil, 1]
+      ]
+    end
+
     describe 'using an IO as input' do
       it 'lexes an instruction with a newline after the name' do
         lex_stringio("<?foo\nbar?>").should == [
diff --git a/spec/oga/xml/parser/processing_instruction_spec.rb b/spec/oga/xml/parser/processing_instruction_spec.rb
index 32b9462..e0ad6d9 100644
--- a/spec/oga/xml/parser/processing_instruction_spec.rb
+++ b/spec/oga/xml/parser/processing_instruction_spec.rb
@@ -32,4 +32,11 @@ describe Oga::XML::Parser do
       @node.text.should == ' bar '
     end
   end
+
+  it 'parses a processing instruction with a namespace prefix' do
+    node = parse('<?foo:bar ?>').children[0]
+
+    node.should be_an_instance_of(Oga::XML::ProcessingInstruction)
+    node.name.should == 'foo:bar'
+  end
 end