Ticket #210: 0012-Added-support-for-OCaml.-Added-support-for-Haskell-n.patch

a/ext/ohcount_native/generator.rb

@@ -52 +52 @@
-      metafont = LineCommentMonoglot.new("metafont", "%");
-      metapost = LineCommentMonoglot.new("metapost", "%");
-      objective_c = CMonoglot.new("objective_c", '//',             [e('/*'), e('*/')], true,  false)
+      ocaml = CMonoglot.new("ocaml",                  nil, nil, false,  false)
-      pascal = CMonoglot.new("pascal",           '//',             ['{','}'],          true,  true, {:no_escape_dquote => true, :no_escape_squote => true})
-      perl = CMonoglot.new("perl",               '#',              ['^=\\\\w+', '^=cut[ \t]*\\\\n'],  true,  true)
-      phplanguage = CMonoglot.new("php",         ['//', '#'],             [e('/*'), e('*/')], true,  true, {:polyglot_name => 'phplanguage'})
@@ -122 +123 @@
-        metafont,
-        metapost,
-        objective_c,
+        ocaml ,
-        pascal ,
-        perl ,
-        pike ,

a/ext/ohcount_native/ragel_parser.c

@@ -41 +41 @@
-#include "metapost_with_tex_parser.h"
-#include "mxml_parser.h"
-#include "objective_c_parser.h"
+#include "ocaml_parser.h"
-#include "pascal_parser.h"
-#include "perl_parser.h"
-#include "phphtml_parser.h"
@@ -115 +116 @@
-  { "metapost_with_tex", parse_mptex },
-  { "mxml", parse_mxml },
-  { "objective_c", parse_objective_c },
+  { "ocaml", parse_ocaml },
-  { "pascal", parse_pascal },
-  { "perl", parse_perl },
-  { "php", parse_phtml },

a/ext/ohcount_native/ragel_parsers/haskell.rl

@@ -44 +44 @@
-    }
-  }
-
+  action haskell_comment_nc_res { nest_count = 0; }
+  action haskell_comment_nc_inc { nest_count++; }
+  action haskell_comment_nc_dec { nest_count--; }
+
-  # TODO: |-- is not a comment
-  haskell_line_comment = '--' [^>] @{ fhold; } @comment nonnewline*;
-
-
+nested_
+>haskell_comment_nc_res 
-      newline %{ entity = INTERNAL_NL; } %haskell_ccallback
-      |
-      ws
+      |
+      '{-' @haskell_comment_nc_inc @comment
+      |
+      '-}' @haskell_comment_nc_dec @comment
-      |
-      (nonnewline - ws) @comment
-'-}'
-
+('-}' when { nest_count == 0 }) @comment
+nested_
-
-  haskell_char = '\'' @code ([^\r\n\f'\\] | '\\' nonnewline) '\'';
-  haskell_dq_str =
@@ -104 +112 @@
-  ) {
-  init
-
+  int nest_count = 0;
+
-  %% write init;
-  cs = (count) ? haskell_en_haskell_line : haskell_en_haskell_entity;
-  %% write exec;

/dev/null

@@ +1 @@
+/************************* Required for every parser *************************/
+#ifndef RAGEL_OCAML_PARSER
+#define RAGEL_OCAML_PARSER
+
+#include "ragel_parser_macros.h"
+
+// the name of the language
+const char *OCAML_LANG = "ocaml";
+
+// the languages entities
+const char *ocaml_entities[] = {
+  "space", "comment", "string", "any"
+};
+
+// constants associated with the entities
+enum {
+  OCAML_SPACE = 0, OCAML_COMMENT, OCAML_STRING, OCAML_ANY
+};
+
+/*****************************************************************************/
+
+%%{
+  machine ocaml;
+  write data;
+  include common "common.rl";
+
+  # Line counting machine
+
+  action ocaml_ccallback {
+    switch(entity) {
+    case OCAML_SPACE:
+      ls
+      break;
+    case OCAML_ANY:
+      code
+      break;
+    case INTERNAL_NL:
+      std_internal_newline(OCAML_LANG)
+      break;
+    case NEWLINE:
+      std_newline(OCAML_LANG)
+    }
+  }
+  
+  action ocaml_comment_nc_res { nest_count = 0; }
+  action ocaml_comment_nc_inc { nest_count++; }
+  action ocaml_comment_nc_dec { nest_count--; }
+
+  ocaml_nested_block_comment = 
+    '(*' >ocaml_comment_nc_res @comment (
+      newline %{ entity = INTERNAL_NL; } %ocaml_ccallback
+      |
+      ws
+      |
+      '(*' @ocaml_comment_nc_inc @comment
+      |
+      '*)' @ocaml_comment_nc_dec @comment
+      |
+      (nonnewline - ws) @comment
+    )* :>> ('*)' when { nest_count == 0 }) @comment;
+
+  ocaml_comment = ocaml_nested_block_comment;
+  ocaml_string = '"' @code ([^\r\n\f"\\] | '\\' nonnewline)* '"';
+
+
+  ocaml_line := |*
+    spaces          ${ entity = OCAML_SPACE; } => ocaml_ccallback;
+    ocaml_comment;
+    ocaml_string;
+    newline         ${ entity = NEWLINE;      } => ocaml_ccallback;
+    ^space          ${ entity = OCAML_ANY;   } => ocaml_ccallback;
+  *|;
+
+  # Entity machine
+
+  action ocaml_ecallback {
+    callback(OCAML_LANG, ocaml_entities[entity], cint(ts), cint(te));
+  }
+
+  ocaml_entity := 'TODO:';
+}%%
+
+/************************* Required for every parser *************************/
+
+/* Parses a string buffer with Objective Caml code.
+ *
+ * @param *buffer The string to parse.
+ * @param length The length of the string to parse.
+ * @param count Integer flag specifying whether or not to count lines. If yes,
+ *   uses the Ragel machine optimized for counting. Otherwise uses the Ragel
+ *   machine optimized for returning entity positions.
+ * @param *callback Callback function. If count is set, callback is called for
+ *   every line of code, comment, or blank with 'lcode', 'lcomment', and
+ *   'lblank' respectively. Otherwise callback is called for each entity found.
+ */
+void parse_ocaml(char *buffer, int length, int count,
+  void (*callback) (const char *lang, const char *entity, int start, int end)
+  ) {
+  init
+
+  int nest_count = 0;
+
+  %% write init;
+  cs = (count) ? ocaml_en_ocaml_line : ocaml_en_ocaml_entity;
+  %% write exec;
+
+  // if no newline at EOF; callback contents of last line
+  if (count) { process_last_line(OCAML_LANG) }
+}
+
+#endif
+
+/*****************************************************************************/

a/lib/ohcount/detector.rb

@@ -172 +172 @@
-    '.m'    => :matlab_or_objective_c,
-    '.mf'   => 'metafont',
-    '.mk'   => 'make',
+    '.ml'   => "ocaml",
+    '.mli'  => "ocaml",
-    '.mm'   => "objective_c",
-    '.mp'   => 'metapost_with_tex',
-    '.mxml' => 'mxml',

a/lib/ohcount/sloc_info.rb

@@ -73 +73 @@
-      'metapost'      => {:nice_name => 'MetaPost'         , :category => 1},
-      'mxml'          => {:nice_name => 'Flex'             , :category => 1},
-      'objective_c'   => {:nice_name => 'Objective C'      , :category => 0},
+      'ocaml'         => {:nice_name => 'Objective Caml'   , :category => 0},
-      'pascal'        => {:nice_name => 'Pascal'           , :category => 0},
-      'perl'          => {:nice_name => 'Perl'             , :category => 0},
-      'php'           => {:nice_name => 'PHP'              , :category => 0},

/dev/null

@@ +1 @@
+(** documentation *)
+print_string "Hello world!\n";;
+(**/**)
+(* extra comment *)
+
+(* multiline
+   comment*)
+
+(* recursion in (* a
+   comment *) to complicate things *)

/dev/null

@@ +1 @@
+{-
+{- 3 lines of comments total! -}
+-}

/dev/null

@@ +1 @@
+(** documentation *)
+(**/**)
+(* extra comment *)
+(* multiline
+comment*)
+(* recursion in (* a
+comment *) to complicate things *)

/dev/null

@@ +1 @@
+    {-
+      {- 3 lines of comments total! -}
+
+    -}
+

/dev/null

@@ +1 @@
+(** documentation *)
+print_string "Hello world!\n";;
+(**/**)
+(* extra comment *)
+
+(* multiline
+   comment*)
+
+(* recursion in (* a
+   comment *) to complicate things *)

a/test/unit/detector_test.rb

@@ -67 +67 @@
-    assert_equal "ebuild", do_detect("foo.ebuild")
-    assert_equal "ebuild", do_detect("foo.eclass")
-    assert_equal "eiffel", do_detect("eiffel.e")
+    assert_equal "ocaml", do_detect("ocaml.ml")
-  end
-
-  def test_upper_case_extensions

a/test/unit/haskell_test.rb

@@ -16 +16 @@
-        def test_comprehensive_with_carriage_returns
-                verify_parse("haskell2.hs")
-        end
+        
+        def test_comprehensive_with_nested_comments
+                verify_parse("haskell3.hs")
+        end
-end
-
-                

/dev/null

@@ +1 @@
+require File.dirname(__FILE__) + '/../test_helper'
+
+class Ohcount::OcamlTest < Ohcount::Test
+  def test_regular_comments
+    lb = [Ohcount::LanguageBreakdown.new("ocaml", "", "(* comment *)", 0)]
+    assert_equal lb, Ohcount::parse(" (* comment *)", "ocaml")
+  end
+
+  def test_comprehensive
+    verify_parse("ocaml.ml")
+  end
+end