diff --git a/.gitignore b/.gitignore
index b76c8926..78b3fb16 100644
--- a/.gitignore
+++ b/.gitignore
@@ -8,3 +8,4 @@ Makefile.main
 build/
 vendor/
 java/lib/
+.vscode/
diff --git a/internal/tokenizer/common.go b/internal/tokenizer/common.go
new file mode 100644
index 00000000..1c47ee33
--- /dev/null
+++ b/internal/tokenizer/common.go
@@ -0,0 +1,7 @@
+// Package tokenizer implements file tokenization used by the enry content
+// classifier. This package is an implementation detail of enry and should not
+// be imported by other packages.
+package tokenizer
+
+// ByteLimit defines the maximum prefix of an input text that will be tokenized.
+const ByteLimit = 100000
diff --git a/internal/tokenizer/flex/lex.linguist_yy.c b/internal/tokenizer/flex/lex.linguist_yy.c
new file mode 100644
index 00000000..1cdb9d0e
--- /dev/null
+++ b/internal/tokenizer/flex/lex.linguist_yy.c
@@ -0,0 +1,2226 @@
+
+#line 3 "lex.linguist_yy.c"
+
+#define  YY_INT_ALIGNED short int
+
+/* A lexical scanner generated by flex */
+
+#define FLEX_SCANNER
+#define YY_FLEX_MAJOR_VERSION 2
+#define YY_FLEX_MINOR_VERSION 5
+#define YY_FLEX_SUBMINOR_VERSION 35
+#if YY_FLEX_SUBMINOR_VERSION > 0
+#define FLEX_BETA
+#endif
+
+/* First, we deal with  platform-specific or compiler-specific issues. */
+
+/* begin standard C headers. */
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+
+/* end standard C headers. */
+
+/* flex integer type definitions */
+
+#ifndef FLEXINT_H
+#define FLEXINT_H
+
+/* C99 systems have <inttypes.h>. Non-C99 systems may or may not. */
+
+#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L
+
+/* C99 says to define __STDC_LIMIT_MACROS before including stdint.h,
+ * if you want the limit (max/min) macros for int types. 
+ */
+#ifndef __STDC_LIMIT_MACROS
+#define __STDC_LIMIT_MACROS 1
+#endif
+
+#include <inttypes.h>
+typedef int8_t flex_int8_t;
+typedef uint8_t flex_uint8_t;
+typedef int16_t flex_int16_t;
+typedef uint16_t flex_uint16_t;
+typedef int32_t flex_int32_t;
+typedef uint32_t flex_uint32_t;
+typedef uint64_t flex_uint64_t;
+#else
+typedef signed char flex_int8_t;
+typedef short int flex_int16_t;
+typedef int flex_int32_t;
+typedef unsigned char flex_uint8_t; 
+typedef unsigned short int flex_uint16_t;
+typedef unsigned int flex_uint32_t;
+#endif /* ! C99 */
+
+/* Limits of integral types. */
+#ifndef INT8_MIN
+#define INT8_MIN               (-128)
+#endif
+#ifndef INT16_MIN
+#define INT16_MIN              (-32767-1)
+#endif
+#ifndef INT32_MIN
+#define INT32_MIN              (-2147483647-1)
+#endif
+#ifndef INT8_MAX
+#define INT8_MAX               (127)
+#endif
+#ifndef INT16_MAX
+#define INT16_MAX              (32767)
+#endif
+#ifndef INT32_MAX
+#define INT32_MAX              (2147483647)
+#endif
+#ifndef UINT8_MAX
+#define UINT8_MAX              (255U)
+#endif
+#ifndef UINT16_MAX
+#define UINT16_MAX             (65535U)
+#endif
+#ifndef UINT32_MAX
+#define UINT32_MAX             (4294967295U)
+#endif
+
+#endif /* ! FLEXINT_H */
+
+#ifdef __cplusplus
+
+/* The "const" storage-class-modifier is valid. */
+#define YY_USE_CONST
+
+#else	/* ! __cplusplus */
+
+/* C99 requires __STDC__ to be defined as 1. */
+#if defined (__STDC__)
+
+#define YY_USE_CONST
+
+#endif	/* defined (__STDC__) */
+#endif	/* ! __cplusplus */
+
+#ifdef YY_USE_CONST
+#define yyconst const
+#else
+#define yyconst
+#endif
+
+/* Returned upon end-of-file. */
+#define YY_NULL 0
+
+/* Promotes a possibly negative, possibly signed char to an unsigned
+ * integer for use as an array index.  If the signed char is negative,
+ * we want to instead treat it as an 8-bit unsigned char, hence the
+ * double cast.
+ */
+#define YY_SC_TO_UI(c) ((unsigned int) (unsigned char) c)
+
+/* An opaque pointer. */
+#ifndef YY_TYPEDEF_YY_SCANNER_T
+#define YY_TYPEDEF_YY_SCANNER_T
+typedef void* yyscan_t;
+#endif
+
+/* For convenience, these vars (plus the bison vars far below)
+   are macros in the reentrant scanner. */
+#define yyin yyg->yyin_r
+#define yyout yyg->yyout_r
+#define yyextra yyg->yyextra_r
+#define yyleng yyg->yyleng_r
+#define yytext yyg->yytext_r
+#define yylineno (YY_CURRENT_BUFFER_LVALUE->yy_bs_lineno)
+#define yycolumn (YY_CURRENT_BUFFER_LVALUE->yy_bs_column)
+#define yy_flex_debug yyg->yy_flex_debug_r
+
+/* Enter a start condition.  This macro really ought to take a parameter,
+ * but we do it the disgusting crufty way forced on us by the ()-less
+ * definition of BEGIN.
+ */
+#define BEGIN yyg->yy_start = 1 + 2 *
+
+/* Translate the current start state into a value that can be later handed
+ * to BEGIN to return to the state.  The YYSTATE alias is for lex
+ * compatibility.
+ */
+#define YY_START ((yyg->yy_start - 1) / 2)
+#define YYSTATE YY_START
+
+/* Action number for EOF rule of a given start state. */
+#define YY_STATE_EOF(state) (YY_END_OF_BUFFER + state + 1)
+
+/* Special action meaning "start processing a new file". */
+#define YY_NEW_FILE linguist_yyrestart(yyin ,yyscanner )
+
+#define YY_END_OF_BUFFER_CHAR 0
+
+/* Size of default input buffer. */
+#ifndef YY_BUF_SIZE
+#define YY_BUF_SIZE 16384
+#endif
+
+/* The state buf must be large enough to hold one state per character in the main buffer.
+ */
+#define YY_STATE_BUF_SIZE   ((YY_BUF_SIZE + 2) * sizeof(yy_state_type))
+
+#ifndef YY_TYPEDEF_YY_BUFFER_STATE
+#define YY_TYPEDEF_YY_BUFFER_STATE
+typedef struct yy_buffer_state *YY_BUFFER_STATE;
+#endif
+
+#ifndef YY_TYPEDEF_YY_SIZE_T
+#define YY_TYPEDEF_YY_SIZE_T
+typedef size_t yy_size_t;
+#endif
+
+#define EOB_ACT_CONTINUE_SCAN 0
+#define EOB_ACT_END_OF_FILE 1
+#define EOB_ACT_LAST_MATCH 2
+
+    #define YY_LESS_LINENO(n)
+    
+/* Return all but the first "n" matched characters back to the input stream. */
+#define yyless(n) \
+	do \
+		{ \
+		/* Undo effects of setting up yytext. */ \
+        int yyless_macro_arg = (n); \
+        YY_LESS_LINENO(yyless_macro_arg);\
+		*yy_cp = yyg->yy_hold_char; \
+		YY_RESTORE_YY_MORE_OFFSET \
+		yyg->yy_c_buf_p = yy_cp = yy_bp + yyless_macro_arg - YY_MORE_ADJ; \
+		YY_DO_BEFORE_ACTION; /* set up yytext again */ \
+		} \
+	while ( 0 )
+
+#define unput(c) yyunput( c, yyg->yytext_ptr , yyscanner )
+
+#ifndef YY_STRUCT_YY_BUFFER_STATE
+#define YY_STRUCT_YY_BUFFER_STATE
+struct yy_buffer_state
+	{
+	FILE *yy_input_file;
+
+	char *yy_ch_buf;		/* input buffer */
+	char *yy_buf_pos;		/* current position in input buffer */
+
+	/* Size of input buffer in bytes, not including room for EOB
+	 * characters.
+	 */
+	yy_size_t yy_buf_size;
+
+	/* Number of characters read into yy_ch_buf, not including EOB
+	 * characters.
+	 */
+	yy_size_t yy_n_chars;
+
+	/* Whether we "own" the buffer - i.e., we know we created it,
+	 * and can realloc() it to grow it, and should free() it to
+	 * delete it.
+	 */
+	int yy_is_our_buffer;
+
+	/* Whether this is an "interactive" input source; if so, and
+	 * if we're using stdio for input, then we want to use getc()
+	 * instead of fread(), to make sure we stop fetching input after
+	 * each newline.
+	 */
+	int yy_is_interactive;
+
+	/* Whether we're considered to be at the beginning of a line.
+	 * If so, '^' rules will be active on the next match, otherwise
+	 * not.
+	 */
+	int yy_at_bol;
+
+    int yy_bs_lineno; /**< The line count. */
+    int yy_bs_column; /**< The column count. */
+    
+	/* Whether to try to fill the input buffer when we reach the
+	 * end of it.
+	 */
+	int yy_fill_buffer;
+
+	int yy_buffer_status;
+
+#define YY_BUFFER_NEW 0
+#define YY_BUFFER_NORMAL 1
+	/* When an EOF's been seen but there's still some text to process
+	 * then we mark the buffer as YY_EOF_PENDING, to indicate that we
+	 * shouldn't try reading from the input source any more.  We might
+	 * still have a bunch of tokens to match, though, because of
+	 * possible backing-up.
+	 *
+	 * When we actually see the EOF, we change the status to "new"
+	 * (via linguist_yyrestart()), so that the user can continue scanning by
+	 * just pointing yyin at a new input file.
+	 */
+#define YY_BUFFER_EOF_PENDING 2
+
+	};
+#endif /* !YY_STRUCT_YY_BUFFER_STATE */
+
+/* We provide macros for accessing buffer states in case in the
+ * future we want to put the buffer states in a more general
+ * "scanner state".
+ *
+ * Returns the top of the stack, or NULL.
+ */
+#define YY_CURRENT_BUFFER ( yyg->yy_buffer_stack \
+                          ? yyg->yy_buffer_stack[yyg->yy_buffer_stack_top] \
+                          : NULL)
+
+/* Same as previous macro, but useful when we know that the buffer stack is not
+ * NULL or when we need an lvalue. For internal use only.
+ */
+#define YY_CURRENT_BUFFER_LVALUE yyg->yy_buffer_stack[yyg->yy_buffer_stack_top]
+
+void linguist_yyrestart (FILE *input_file ,yyscan_t yyscanner );
+void linguist_yy_switch_to_buffer (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_create_buffer (FILE *file,int size ,yyscan_t yyscanner );
+void linguist_yy_delete_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner );
+void linguist_yy_flush_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner );
+void linguist_yypush_buffer_state (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner );
+void linguist_yypop_buffer_state (yyscan_t yyscanner );
+
+static void linguist_yyensure_buffer_stack (yyscan_t yyscanner );
+static void linguist_yy_load_buffer_state (yyscan_t yyscanner );
+static void linguist_yy_init_buffer (YY_BUFFER_STATE b,FILE *file ,yyscan_t yyscanner );
+
+#define YY_FLUSH_BUFFER linguist_yy_flush_buffer(YY_CURRENT_BUFFER ,yyscanner)
+
+YY_BUFFER_STATE linguist_yy_scan_buffer (char *base,yy_size_t size ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_scan_string (yyconst char *yy_str ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_scan_bytes (yyconst char *bytes,yy_size_t len ,yyscan_t yyscanner );
+
+void *linguist_yyalloc (yy_size_t ,yyscan_t yyscanner );
+void *linguist_yyrealloc (void *,yy_size_t ,yyscan_t yyscanner );
+void linguist_yyfree (void * ,yyscan_t yyscanner );
+
+#define yy_new_buffer linguist_yy_create_buffer
+
+#define yy_set_interactive(is_interactive) \
+	{ \
+	if ( ! YY_CURRENT_BUFFER ){ \
+        linguist_yyensure_buffer_stack (yyscanner); \
+		YY_CURRENT_BUFFER_LVALUE =    \
+            linguist_yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); \
+	} \
+	YY_CURRENT_BUFFER_LVALUE->yy_is_interactive = is_interactive; \
+	}
+
+#define yy_set_bol(at_bol) \
+	{ \
+	if ( ! YY_CURRENT_BUFFER ){\
+        linguist_yyensure_buffer_stack (yyscanner); \
+		YY_CURRENT_BUFFER_LVALUE =    \
+            linguist_yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); \
+	} \
+	YY_CURRENT_BUFFER_LVALUE->yy_at_bol = at_bol; \
+	}
+
+#define YY_AT_BOL() (YY_CURRENT_BUFFER_LVALUE->yy_at_bol)
+
+/* Begin user sect3 */
+
+typedef unsigned char YY_CHAR;
+
+typedef int yy_state_type;
+
+#define yytext_ptr yytext_r
+
+static yy_state_type yy_get_previous_state (yyscan_t yyscanner );
+static yy_state_type yy_try_NUL_trans (yy_state_type current_state  ,yyscan_t yyscanner);
+static int yy_get_next_buffer (yyscan_t yyscanner );
+static void yy_fatal_error (yyconst char msg[] ,yyscan_t yyscanner );
+
+/* Done after the current pattern has been matched and before the
+ * corresponding action - sets up yytext.
+ */
+#define YY_DO_BEFORE_ACTION \
+	yyg->yytext_ptr = yy_bp; \
+	yyleng = (yy_size_t) (yy_cp - yy_bp); \
+	yyg->yy_hold_char = *yy_cp; \
+	*yy_cp = '\0'; \
+	yyg->yy_c_buf_p = yy_cp;
+
+#define YY_NUM_RULES 31
+#define YY_END_OF_BUFFER 32
+/* This struct is not used in this scanner,
+   but its presence is necessary. */
+struct yy_trans_info
+	{
+	flex_int32_t yy_verify;
+	flex_int32_t yy_nxt;
+	};
+static yyconst flex_int16_t yy_accept[114] =
+    {   0,
+        0,    0,    0,    0,    0,    0,    0,    0,    0,    0,
+        0,    0,    0,    0,    0,    0,   32,   30,   17,   28,
+       29,   29,   18,   27,   27,   28,   28,   19,   19,   29,
+       27,   29,   30,   17,   28,   29,   29,   28,   26,   24,
+       25,    9,    9,    9,    9,    9,    9,    9,   16,   28,
+       29,   16,    6,    4,   19,   28,   19,   19,   28,   20,
+       29,    5,   29,    0,    0,    0,    0,    3,    0,   28,
+       24,   23,   10,    0,   12,   13,    0,    0,    7,    8,
+       19,   19,   19,    3,    0,    2,    0,    2,    2,   21,
+       22,   23,   11,   14,   15,   19,   19,   19,   19,   19,
+
+       19,    0,    2,    0,    2,    0,    0,    0,    0,    0,
+        1,    0,    0
+    } ;
+
+static yyconst flex_int32_t yy_ec[256] =
+    {   0,
+        1,    1,    1,    1,    1,    1,    1,    1,    2,    3,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    4,    5,    6,    7,    1,    8,    9,   10,   11,
+       12,   13,   14,    1,   15,   16,   17,   18,   19,   19,
+       19,   19,   19,   19,   19,   19,   19,    1,   20,   21,
+       22,   23,   24,   25,   26,   26,   26,   26,   27,   28,
+       29,   29,   29,   29,   29,   30,   29,   29,   29,   29,
+       29,   29,   29,   29,   31,   29,   29,   29,   29,   29,
+       32,    1,   33,    1,   29,    1,   26,   26,   26,   26,
+
+       34,   28,   29,   29,   29,   29,   29,   30,   29,   35,
+       29,   29,   29,   29,   29,   29,   31,   36,   29,   37,
+       29,   29,   38,   39,   40,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1
+    } ;
+
+static yyconst flex_int32_t yy_meta[41] =
+    {   0,
+        1,    2,    3,    2,    4,    5,    6,    1,    1,    5,
+        1,    1,    6,    1,    4,    7,    8,    9,    9,    1,
+        4,   10,    1,    4,    6,    9,    9,    9,    9,    9,
+        9,    1,    1,    9,    9,    9,    9,    1,    1,    1
+    } ;
+
+static yyconst flex_int16_t yy_base[125] =
+    {   0,
+        0,   39,   39,   59,  214,  213,  210,  209,  208,  207,
+      208,  207,  213,  212,  207,  206,  215,  358,  208,    0,
+      358,  204,  202,  197,  358,    0,  194,   90,  112,  185,
+      190,  165,  145,   38,   44,  199,  184,   38,  358,  174,
+      358,  358,  178,  179,  151,  176,  180,  174,  174,    0,
+      358,  168,  358,    0,    0,   38,   31,  140,  137,  140,
+      358,  358,  358,    0,  154,  139,  133,    0,  155,  141,
+      122,   54,  358,  118,  358,  358,  131,  113,  358,  358,
+       53,   84,  174,    0,  164,   96,   62,  100,   28,  358,
+      358,   74,  358,  358,  358,  157,   71,    0,    0,   77,
+
+        0,   88,   14,   94,  108,   96,  207,  131,  175,  134,
+      243,  165,  358,  280,  290,  295,  301,  302,  312,  321,
+      326,  328,  337,  347
+    } ;
+
+static yyconst flex_int16_t yy_def[125] =
+    {   0,
+      113,    1,  114,  114,  115,  115,  115,  115,  115,  115,
+      115,  115,  115,  115,  115,  115,  113,  113,  113,  116,
+      113,  113,  113,  113,  113,  116,  116,  113,  116,  117,
+      113,  113,  113,  113,  116,  113,  113,  116,  113,  118,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  116,
+      113,  113,  113,  116,   29,  116,  116,  116,  116,  117,
+      113,  113,  113,   33,  113,  113,  113,  119,  120,  116,
+      118,  121,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  116,   29,  119,  120,   85,  122,   85,   88,  113,
+      113,  121,  113,  113,  113,  113,  113,  116,   83,   83,
+
+       83,  122,   88,  122,   88,  122,  123,  122,  123,  124,
+      123,  124,    0,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113
+    } ;
+
+static yyconst flex_int16_t yy_nxt[399] =
+    {   0,
+       18,   18,   18,   18,   18,   19,   20,   21,   22,   23,
+       24,   25,   26,   21,   21,   20,   27,   28,   29,   25,
+       30,   18,   18,   18,   20,   20,   20,   20,   20,   20,
+       20,   25,   25,   20,   20,   20,   20,   31,   32,   25,
+       33,   68,   33,   49,   34,   35,   36,   68,   69,  105,
+       54,   81,   81,   37,   70,   38,   40,   40,   57,   90,
+       57,   41,  103,   91,   40,   40,   40,   40,   40,   40,
+       96,   96,   40,   40,   40,   40,   40,   40,  102,  113,
+       97,   41,   97,  113,   40,   40,   40,   40,   40,   40,
+       81,   81,   40,   40,   40,   40,   50,  113,   97,  113,
+
+       97,  113,   50,  113,  102,   55,   50,   55,   55,  107,
+      102,  107,  102,   98,   50,   50,   56,   57,   50,   57,
+       58,  104,   95,   56,   50,   50,   59,   55,  106,   55,
+       55,  108,  107,   88,  107,  107,   94,  107,   56,   57,
+       93,   57,   58,   72,   68,   56,   64,  102,   64,   65,
+       65,   65,   65,   65,   83,   83,   85,   68,   85,   66,
+      113,   67,   83,   83,   83,   85,  107,   85,  107,   82,
+       83,   86,   87,   87,   96,   96,  113,   80,  113,   79,
+       86,   87,   87,   78,   97,   77,   97,   76,   89,   99,
+       75,   99,   99,   74,   73,   72,  110,   89,   65,   99,
+
+      100,  101,   68,   63,   62,   61,   54,  100,  107,   53,
+      107,   52,   51,   49,  113,   48,   48,   47,   47,   46,
+       46,   45,   45,   44,   44,   43,   43,  113,  110,  113,
+      113,  113,  111,  111,  111,  111,  111,  111,  113,  113,
+      111,  111,  111,  111,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  110,  113,  113,  113,  111,  111,
+      111,  111,  111,  111,  113,  113,  111,  111,  111,  111,
+       39,   39,   39,   39,   39,   39,   39,   39,   39,   39,
+       42,   42,   42,   42,   42,   42,   42,   42,   42,   42,
+
+       50,   50,   50,   50,   60,  113,  113,   60,   60,   60,
+       71,   71,   84,   84,  113,   84,   84,   84,   84,   84,
+       84,   84,   88,  113,  113,  113,  113,  113,   88,   88,
+       92,  113,  113,  113,   92,   87,   87,  109,  109,  109,
+      109,  109,  109,  109,  109,  109,  109,  112,  112,  112,
+      112,  112,  112,  112,  112,  112,  112,   17,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113
+
+    } ;
+
+static yyconst flex_int16_t yy_chk[399] =
+    {   0,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        1,    1,    1,    1,    1,    1,    1,    1,    1,    1,
+        2,   34,    2,   34,    2,    2,    2,   35,   35,  103,
+       38,   56,   56,    2,   38,    2,    3,    3,   57,   72,
+       57,    3,   89,   72,    3,    3,    3,    3,    3,    3,
+       81,   81,    3,    3,    3,    3,    4,    4,   87,   92,
+       81,    4,   81,   92,    4,    4,    4,    4,    4,    4,
+      100,  100,    4,    4,    4,    4,   28,   86,   97,   86,
+
+       97,   88,   28,   88,  102,   28,   28,   28,   28,  105,
+      104,  105,  106,   82,   28,   28,   28,   28,   28,   28,
+       28,  102,   78,   28,   28,   28,   28,   29,  104,   29,
+       29,  106,  108,   88,  108,  110,   77,  110,   29,   29,
+       74,   29,   29,   71,   70,   29,   33,  108,   33,   67,
+       33,   33,   33,   66,   59,   59,   69,   65,   69,   33,
+       60,   33,   59,   59,   59,   85,  112,   85,  112,   58,
+       59,   69,   69,   69,   96,   96,  109,   52,  109,   49,
+       85,   85,   85,   48,   96,   47,   96,   46,   69,   83,
+       45,   83,   83,   44,   43,   40,  109,   85,   37,   83,
+
+       83,   83,   36,   32,   31,   30,   27,   83,  107,   24,
+      107,   23,   22,   19,   17,   16,   15,   14,   13,   12,
+       11,   10,    9,    8,    7,    6,    5,    0,  107,    0,
+        0,    0,  107,  107,  107,  107,  107,  107,    0,    0,
+      107,  107,  107,  107,  111,    0,  111,    0,    0,    0,
+        0,    0,    0,    0,    0,    0,    0,    0,    0,    0,
+        0,    0,    0,    0,  111,    0,    0,    0,  111,  111,
+      111,  111,  111,  111,    0,    0,  111,  111,  111,  111,
+      114,  114,  114,  114,  114,  114,  114,  114,  114,  114,
+      115,  115,  115,  115,  115,  115,  115,  115,  115,  115,
+
+      116,  116,  116,  116,  117,    0,    0,  117,  117,  117,
+      118,  118,  119,  119,    0,  119,  119,  119,  119,  119,
+      119,  119,  120,    0,    0,    0,    0,    0,  120,  120,
+      121,    0,    0,    0,  121,  122,  122,  123,  123,  123,
+      123,  123,  123,  123,  123,  123,  123,  124,  124,  124,
+      124,  124,  124,  124,  124,  124,  124,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113,  113,  113,
+      113,  113,  113,  113,  113,  113,  113,  113
+
+    } ;
+
+/* The intent behind this definition is that it'll catch
+ * any uses of REJECT which flex missed.
+ */
+#define REJECT reject_used_but_not_detected
+#define yymore() yymore_used_but_not_detected
+#define YY_MORE_ADJ 0
+#define YY_RESTORE_YY_MORE_OFFSET
+#line 1 "tokenizer.l"
+#line 2 "tokenizer.l"
+
+#include "linguist.h"
+
+#define feed_token(tok, typ) do { \
+    yyextra->token = (tok); \
+    yyextra->type = (typ); \
+  } while (0)
+
+#define eat_until_eol() do { \
+    int c; \
+    while ((c = input(yyscanner)) != '\n' && c != EOF && c); \
+    if (c == EOF || !c) \
+      return 0; \
+  } while (0)
+
+#define eat_until_unescaped(q) do { \
+    int c; \
+    while ((c = input(yyscanner)) != EOF && c) { \
+      if (c == '\n') \
+        break; \
+      if (c == '\\') { \
+        c = input(yyscanner); \
+        if (c == EOF || !c) \
+          return 0; \
+      } else if (c == q) \
+        break; \
+    } \
+    if (c == EOF || !c) \
+      return 0; \
+  } while (0)
+
+
+#line 589 "lex.linguist_yy.c"
+
+#define INITIAL 0
+#define sgml 1
+#define c_comment 2
+#define xml_comment 3
+#define haskell_comment 4
+#define ocaml_comment 5
+#define python_dcomment 6
+#define python_scomment 7
+
+#ifndef YY_NO_UNISTD_H
+/* Special case for "unistd.h", since it is non-ANSI. We include it way
+ * down here because we want the user's section 1 to have been scanned first.
+ * The user has a chance to override it with an option.
+ */
+#include <unistd.h>
+#endif
+
+#define YY_EXTRA_TYPE struct tokenizer_extra *
+
+/* Holds the entire state of the reentrant scanner. */
+struct yyguts_t
+    {
+
+    /* User-defined. Not touched by flex. */
+    YY_EXTRA_TYPE yyextra_r;
+
+    /* The rest are the same as the globals declared in the non-reentrant scanner. */
+    FILE *yyin_r, *yyout_r;
+    size_t yy_buffer_stack_top; /**< index of top of stack. */
+    size_t yy_buffer_stack_max; /**< capacity of stack. */
+    YY_BUFFER_STATE * yy_buffer_stack; /**< Stack as an array. */
+    char yy_hold_char;
+    yy_size_t yy_n_chars;
+    yy_size_t yyleng_r;
+    char *yy_c_buf_p;
+    int yy_init;
+    int yy_start;
+    int yy_did_buffer_switch_on_eof;
+    int yy_start_stack_ptr;
+    int yy_start_stack_depth;
+    int *yy_start_stack;
+    yy_state_type yy_last_accepting_state;
+    char* yy_last_accepting_cpos;
+
+    int yylineno_r;
+    int yy_flex_debug_r;
+
+    char *yytext_r;
+    int yy_more_flag;
+    int yy_more_len;
+
+    }; /* end struct yyguts_t */
+
+static int yy_init_globals (yyscan_t yyscanner );
+
+int linguist_yylex_init (yyscan_t* scanner);
+
+int linguist_yylex_init_extra (YY_EXTRA_TYPE user_defined,yyscan_t* scanner);
+
+/* Accessor methods to globals.
+   These are made visible to non-reentrant scanners for convenience. */
+
+int linguist_yylex_destroy (yyscan_t yyscanner );
+
+int linguist_yyget_debug (yyscan_t yyscanner );
+
+void linguist_yyset_debug (int debug_flag ,yyscan_t yyscanner );
+
+YY_EXTRA_TYPE linguist_yyget_extra (yyscan_t yyscanner );
+
+void linguist_yyset_extra (YY_EXTRA_TYPE user_defined ,yyscan_t yyscanner );
+
+FILE *linguist_yyget_in (yyscan_t yyscanner );
+
+void linguist_yyset_in  (FILE * in_str ,yyscan_t yyscanner );
+
+FILE *linguist_yyget_out (yyscan_t yyscanner );
+
+void linguist_yyset_out  (FILE * out_str ,yyscan_t yyscanner );
+
+yy_size_t linguist_yyget_leng (yyscan_t yyscanner );
+
+char *linguist_yyget_text (yyscan_t yyscanner );
+
+int linguist_yyget_lineno (yyscan_t yyscanner );
+
+void linguist_yyset_lineno (int line_number ,yyscan_t yyscanner );
+
+/* Macros after this point can all be overridden by user definitions in
+ * section 1.
+ */
+
+#ifndef YY_SKIP_YYWRAP
+#ifdef __cplusplus
+extern "C" int linguist_yywrap (yyscan_t yyscanner );
+#else
+extern int linguist_yywrap (yyscan_t yyscanner );
+#endif
+#endif
+
+#ifndef yytext_ptr
+static void yy_flex_strncpy (char *,yyconst char *,int ,yyscan_t yyscanner);
+#endif
+
+#ifdef YY_NEED_STRLEN
+static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner);
+#endif
+
+#ifndef YY_NO_INPUT
+
+#ifdef __cplusplus
+static int yyinput (yyscan_t yyscanner );
+#else
+static int input (yyscan_t yyscanner );
+#endif
+
+#endif
+
+/* Amount of stuff to slurp up with each read. */
+#ifndef YY_READ_BUF_SIZE
+#define YY_READ_BUF_SIZE 8192
+#endif
+
+/* Copy whatever the last rule matched to the standard output. */
+#ifndef ECHO
+/* This used to be an fputs(), but since the string might contain NUL's,
+ * we now use fwrite().
+ */
+#define ECHO fwrite( yytext, yyleng, 1, yyout )
+#endif
+
+/* Gets input and stuffs it into "buf".  number of characters read, or YY_NULL,
+ * is returned in "result".
+ */
+#ifndef YY_INPUT
+#define YY_INPUT(buf,result,max_size) \
+	if ( YY_CURRENT_BUFFER_LVALUE->yy_is_interactive ) \
+		{ \
+		int c = '*'; \
+		yy_size_t n; \
+		for ( n = 0; n < max_size && \
+			     (c = getc( yyin )) != EOF && c != '\n'; ++n ) \
+			buf[n] = (char) c; \
+		if ( c == '\n' ) \
+			buf[n++] = (char) c; \
+		if ( c == EOF && ferror( yyin ) ) \
+			YY_FATAL_ERROR( "input in flex scanner failed" ); \
+		result = n; \
+		} \
+	else \
+		{ \
+		errno=0; \
+		while ( (result = fread(buf, 1, max_size, yyin))==0 && ferror(yyin)) \
+			{ \
+			if( errno != EINTR) \
+				{ \
+				YY_FATAL_ERROR( "input in flex scanner failed" ); \
+				break; \
+				} \
+			errno=0; \
+			clearerr(yyin); \
+			} \
+		}\
+\
+
+#endif
+
+/* No semi-colon after return; correct usage is to write "yyterminate();" -
+ * we don't want an extra ';' after the "return" because that will cause
+ * some compilers to complain about unreachable statements.
+ */
+#ifndef yyterminate
+#define yyterminate() return YY_NULL
+#endif
+
+/* Number of entries by which start-condition stack grows. */
+#ifndef YY_START_STACK_INCR
+#define YY_START_STACK_INCR 25
+#endif
+
+/* Report a fatal error. */
+#ifndef YY_FATAL_ERROR
+#define YY_FATAL_ERROR(msg) yy_fatal_error( msg , yyscanner)
+#endif
+
+/* end tables serialization structures and prototypes */
+
+/* Default declaration of generated scanner - a define so the user can
+ * easily add parameters.
+ */
+#ifndef YY_DECL
+#define YY_DECL_IS_OURS 1
+
+extern int linguist_yylex (yyscan_t yyscanner);
+
+#define YY_DECL int linguist_yylex (yyscan_t yyscanner)
+#endif /* !YY_DECL */
+
+/* Code executed at the beginning of each rule, after yytext and yyleng
+ * have been set up.
+ */
+#ifndef YY_USER_ACTION
+#define YY_USER_ACTION
+#endif
+
+/* Code executed at the end of each rule. */
+#ifndef YY_BREAK
+#define YY_BREAK break;
+#endif
+
+#define YY_RULE_SETUP \
+	if ( yyleng > 0 ) \
+		YY_CURRENT_BUFFER_LVALUE->yy_at_bol = \
+				(yytext[yyleng - 1] == '\n'); \
+	YY_USER_ACTION
+
+/** The main scanner function which does all the work.
+ */
+YY_DECL
+{
+	register yy_state_type yy_current_state;
+	register char *yy_cp, *yy_bp;
+	register int yy_act;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+#line 38 "tokenizer.l"
+
+
+#line 819 "lex.linguist_yy.c"
+
+	if ( !yyg->yy_init )
+		{
+		yyg->yy_init = 1;
+
+#ifdef YY_USER_INIT
+		YY_USER_INIT;
+#endif
+
+		if ( ! yyg->yy_start )
+			yyg->yy_start = 1;	/* first start state */
+
+		if ( ! yyin )
+			yyin = stdin;
+
+		if ( ! yyout )
+			yyout = stdout;
+
+		if ( ! YY_CURRENT_BUFFER ) {
+			linguist_yyensure_buffer_stack (yyscanner);
+			YY_CURRENT_BUFFER_LVALUE =
+				linguist_yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner);
+		}
+
+		linguist_yy_load_buffer_state(yyscanner );
+		}
+
+	while ( 1 )		/* loops until end-of-file is reached */
+		{
+		yy_cp = yyg->yy_c_buf_p;
+
+		/* Support of yytext. */
+		*yy_cp = yyg->yy_hold_char;
+
+		/* yy_bp points to the position in yy_ch_buf of the start of
+		 * the current run.
+		 */
+		yy_bp = yy_cp;
+
+		yy_current_state = yyg->yy_start;
+		yy_current_state += YY_AT_BOL();
+yy_match:
+		do
+			{
+			register YY_CHAR yy_c = yy_ec[YY_SC_TO_UI(*yy_cp)];
+			if ( yy_accept[yy_current_state] )
+				{
+				yyg->yy_last_accepting_state = yy_current_state;
+				yyg->yy_last_accepting_cpos = yy_cp;
+				}
+			while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state )
+				{
+				yy_current_state = (int) yy_def[yy_current_state];
+				if ( yy_current_state >= 114 )
+					yy_c = yy_meta[(unsigned int) yy_c];
+				}
+			yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c];
+			++yy_cp;
+			}
+		while ( yy_current_state != 113 );
+		yy_cp = yyg->yy_last_accepting_cpos;
+		yy_current_state = yyg->yy_last_accepting_state;
+
+yy_find_action:
+		yy_act = yy_accept[yy_current_state];
+
+		YY_DO_BEFORE_ACTION;
+
+do_action:	/* This label is used only to access EOF actions. */
+
+		switch ( yy_act )
+	{ /* beginning of action switch */
+			case 0: /* must back up */
+			/* undo the effects of YY_DO_BEFORE_ACTION */
+			*yy_cp = yyg->yy_hold_char;
+			yy_cp = yyg->yy_last_accepting_cpos;
+			yy_current_state = yyg->yy_last_accepting_state;
+			goto yy_find_action;
+
+case 1:
+/* rule 1 can match eol */
+YY_RULE_SETUP
+#line 40 "tokenizer.l"
+{
+    const char *off = strrchr(yytext, ' ');
+    if (!off)
+      off = yytext;
+    else
+      ++off;
+    feed_token(strdup(off), SHEBANG_TOKEN);
+    eat_until_eol();
+    return 1;
+  }
+	YY_BREAK
+case 2:
+YY_RULE_SETUP
+#line 51 "tokenizer.l"
+{
+    const char *off = strrchr(yytext, '/');
+    if (!off)
+      off = yytext;
+    else
+      ++off;
+    if (strcmp(off, "env") == 0) {
+      eat_until_eol();
+    } else {
+      feed_token(strdup(off), SHEBANG_TOKEN);
+      eat_until_eol();
+      return 1;
+    }
+  }
+	YY_BREAK
+case 3:
+YY_RULE_SETUP
+#line 66 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 4:
+YY_RULE_SETUP
+#line 68 "tokenizer.l"
+{ BEGIN(c_comment); }
+	YY_BREAK
+/* See below for xml_comment start. */
+case 5:
+YY_RULE_SETUP
+#line 70 "tokenizer.l"
+{ BEGIN(haskell_comment); }
+	YY_BREAK
+case 6:
+YY_RULE_SETUP
+#line 71 "tokenizer.l"
+{ BEGIN(ocaml_comment); }
+	YY_BREAK
+case 7:
+YY_RULE_SETUP
+#line 72 "tokenizer.l"
+{ BEGIN(python_dcomment); }
+	YY_BREAK
+case 8:
+YY_RULE_SETUP
+#line 73 "tokenizer.l"
+{ BEGIN(python_scomment); }
+	YY_BREAK
+case 9:
+/* rule 9 can match eol */
+YY_RULE_SETUP
+#line 75 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 10:
+YY_RULE_SETUP
+#line 76 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 11:
+YY_RULE_SETUP
+#line 77 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 12:
+YY_RULE_SETUP
+#line 78 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 13:
+YY_RULE_SETUP
+#line 79 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 14:
+YY_RULE_SETUP
+#line 80 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 15:
+YY_RULE_SETUP
+#line 81 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 16:
+YY_RULE_SETUP
+#line 83 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 17:
+YY_RULE_SETUP
+#line 84 "tokenizer.l"
+{ eat_until_unescaped('"'); }
+	YY_BREAK
+case 18:
+YY_RULE_SETUP
+#line 85 "tokenizer.l"
+{ eat_until_unescaped('\''); }
+	YY_BREAK
+case 19:
+YY_RULE_SETUP
+#line 86 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 20:
+YY_RULE_SETUP
+#line 87 "tokenizer.l"
+{
+    if (strcmp(yytext, "<!--") == 0) {
+     BEGIN(xml_comment);
+    } else {
+      feed_token(strdup(yytext), SGML_TOKEN);
+      BEGIN(sgml);
+      return 1;
+    }
+  }
+	YY_BREAK
+case 21:
+YY_RULE_SETUP
+#line 96 "tokenizer.l"
+{ feed_token(strndup(yytext, strlen(yytext) - 1), REGULAR_TOKEN); eat_until_unescaped('"'); return 1; }
+	YY_BREAK
+case 22:
+YY_RULE_SETUP
+#line 97 "tokenizer.l"
+{ feed_token(strndup(yytext, strlen(yytext) - 1), REGULAR_TOKEN); eat_until_unescaped('\''); return 1; }
+	YY_BREAK
+case 23:
+YY_RULE_SETUP
+#line 98 "tokenizer.l"
+{ feed_token(strdup(yytext), REGULAR_TOKEN); *(strchr(yyextra->token, '=') + 1) = 0; return 1; }
+	YY_BREAK
+case 24:
+YY_RULE_SETUP
+#line 99 "tokenizer.l"
+{ feed_token(strdup(yytext), REGULAR_TOKEN); return 1; }
+	YY_BREAK
+case 25:
+YY_RULE_SETUP
+#line 100 "tokenizer.l"
+{ BEGIN(INITIAL); }
+	YY_BREAK
+case 26:
+/* rule 26 can match eol */
+YY_RULE_SETUP
+#line 101 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 27:
+YY_RULE_SETUP
+#line 102 "tokenizer.l"
+{ feed_token(strdup(yytext), REGULAR_TOKEN); return 1; }
+	YY_BREAK
+case 28:
+YY_RULE_SETUP
+#line 103 "tokenizer.l"
+{
+    if (strncmp(yytext, "/*", 2) == 0) {
+      if (strlen(yytext) >= 4 && strcmp(yytext + strlen(yytext) - 2, "*/") == 0) {
+        /* nothing */
+      } else {
+        BEGIN(c_comment);
+      }
+    } else {
+      feed_token(strdup(yytext), REGULAR_TOKEN);
+      return 1;
+    }
+  }
+	YY_BREAK
+case 29:
+YY_RULE_SETUP
+#line 115 "tokenizer.l"
+{ feed_token(strdup(yytext), REGULAR_TOKEN); return 1; }
+	YY_BREAK
+case 30:
+/* rule 30 can match eol */
+YY_RULE_SETUP
+#line 116 "tokenizer.l"
+{ /* nothing */ }
+	YY_BREAK
+case 31:
+YY_RULE_SETUP
+#line 118 "tokenizer.l"
+YY_FATAL_ERROR( "flex scanner jammed" );
+	YY_BREAK
+#line 1100 "lex.linguist_yy.c"
+case YY_STATE_EOF(INITIAL):
+case YY_STATE_EOF(sgml):
+case YY_STATE_EOF(c_comment):
+case YY_STATE_EOF(xml_comment):
+case YY_STATE_EOF(haskell_comment):
+case YY_STATE_EOF(ocaml_comment):
+case YY_STATE_EOF(python_dcomment):
+case YY_STATE_EOF(python_scomment):
+	yyterminate();
+
+	case YY_END_OF_BUFFER:
+		{
+		/* Amount of text matched not including the EOB char. */
+		int yy_amount_of_matched_text = (int) (yy_cp - yyg->yytext_ptr) - 1;
+
+		/* Undo the effects of YY_DO_BEFORE_ACTION. */
+		*yy_cp = yyg->yy_hold_char;
+		YY_RESTORE_YY_MORE_OFFSET
+
+		if ( YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_NEW )
+			{
+			/* We're scanning a new file or input source.  It's
+			 * possible that this happened because the user
+			 * just pointed yyin at a new source and called
+			 * linguist_yylex().  If so, then we have to assure
+			 * consistency between YY_CURRENT_BUFFER and our
+			 * globals.  Here is the right place to do so, because
+			 * this is the first action (other than possibly a
+			 * back-up) that will match for the new input source.
+			 */
+			yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars;
+			YY_CURRENT_BUFFER_LVALUE->yy_input_file = yyin;
+			YY_CURRENT_BUFFER_LVALUE->yy_buffer_status = YY_BUFFER_NORMAL;
+			}
+
+		/* Note that here we test for yy_c_buf_p "<=" to the position
+		 * of the first EOB in the buffer, since yy_c_buf_p will
+		 * already have been incremented past the NUL character
+		 * (since all states make transitions on EOB to the
+		 * end-of-buffer state).  Contrast this with the test
+		 * in input().
+		 */
+		if ( yyg->yy_c_buf_p <= &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] )
+			{ /* This was really a NUL. */
+			yy_state_type yy_next_state;
+
+			yyg->yy_c_buf_p = yyg->yytext_ptr + yy_amount_of_matched_text;
+
+			yy_current_state = yy_get_previous_state( yyscanner );
+
+			/* Okay, we're now positioned to make the NUL
+			 * transition.  We couldn't have
+			 * yy_get_previous_state() go ahead and do it
+			 * for us because it doesn't know how to deal
+			 * with the possibility of jamming (and we don't
+			 * want to build jamming into it because then it
+			 * will run more slowly).
+			 */
+
+			yy_next_state = yy_try_NUL_trans( yy_current_state , yyscanner);
+
+			yy_bp = yyg->yytext_ptr + YY_MORE_ADJ;
+
+			if ( yy_next_state )
+				{
+				/* Consume the NUL. */
+				yy_cp = ++yyg->yy_c_buf_p;
+				yy_current_state = yy_next_state;
+				goto yy_match;
+				}
+
+			else
+				{
+				yy_cp = yyg->yy_last_accepting_cpos;
+				yy_current_state = yyg->yy_last_accepting_state;
+				goto yy_find_action;
+				}
+			}
+
+		else switch ( yy_get_next_buffer( yyscanner ) )
+			{
+			case EOB_ACT_END_OF_FILE:
+				{
+				yyg->yy_did_buffer_switch_on_eof = 0;
+
+				if ( linguist_yywrap(yyscanner ) )
+					{
+					/* Note: because we've taken care in
+					 * yy_get_next_buffer() to have set up
+					 * yytext, we can now set up
+					 * yy_c_buf_p so that if some total
+					 * hoser (like flex itself) wants to
+					 * call the scanner after we return the
+					 * YY_NULL, it'll still work - another
+					 * YY_NULL will get returned.
+					 */
+					yyg->yy_c_buf_p = yyg->yytext_ptr + YY_MORE_ADJ;
+
+					yy_act = YY_STATE_EOF(YY_START);
+					goto do_action;
+					}
+
+				else
+					{
+					if ( ! yyg->yy_did_buffer_switch_on_eof )
+						YY_NEW_FILE;
+					}
+				break;
+				}
+
+			case EOB_ACT_CONTINUE_SCAN:
+				yyg->yy_c_buf_p =
+					yyg->yytext_ptr + yy_amount_of_matched_text;
+
+				yy_current_state = yy_get_previous_state( yyscanner );
+
+				yy_cp = yyg->yy_c_buf_p;
+				yy_bp = yyg->yytext_ptr + YY_MORE_ADJ;
+				goto yy_match;
+
+			case EOB_ACT_LAST_MATCH:
+				yyg->yy_c_buf_p =
+				&YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars];
+
+				yy_current_state = yy_get_previous_state( yyscanner );
+
+				yy_cp = yyg->yy_c_buf_p;
+				yy_bp = yyg->yytext_ptr + YY_MORE_ADJ;
+				goto yy_find_action;
+			}
+		break;
+		}
+
+	default:
+		YY_FATAL_ERROR(
+			"fatal flex scanner internal error--no action found" );
+	} /* end of action switch */
+		} /* end of scanning one token */
+} /* end of linguist_yylex */
+
+/* yy_get_next_buffer - try to read in a new buffer
+ *
+ * Returns a code representing an action:
+ *	EOB_ACT_LAST_MATCH -
+ *	EOB_ACT_CONTINUE_SCAN - continue scanning from current position
+ *	EOB_ACT_END_OF_FILE - end of file
+ */
+static int yy_get_next_buffer (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+	register char *dest = YY_CURRENT_BUFFER_LVALUE->yy_ch_buf;
+	register char *source = yyg->yytext_ptr;
+	register int number_to_move, i;
+	int ret_val;
+
+	if ( yyg->yy_c_buf_p > &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1] )
+		YY_FATAL_ERROR(
+		"fatal flex scanner internal error--end of buffer missed" );
+
+	if ( YY_CURRENT_BUFFER_LVALUE->yy_fill_buffer == 0 )
+		{ /* Don't try to fill the buffer, so this is an EOF. */
+		if ( yyg->yy_c_buf_p - yyg->yytext_ptr - YY_MORE_ADJ == 1 )
+			{
+			/* We matched a single character, the EOB, so
+			 * treat this as a final EOF.
+			 */
+			return EOB_ACT_END_OF_FILE;
+			}
+
+		else
+			{
+			/* We matched some text prior to the EOB, first
+			 * process it.
+			 */
+			return EOB_ACT_LAST_MATCH;
+			}
+		}
+
+	/* Try to read more data. */
+
+	/* First move last chars to start of buffer. */
+	number_to_move = (int) (yyg->yy_c_buf_p - yyg->yytext_ptr) - 1;
+
+	for ( i = 0; i < number_to_move; ++i )
+		*(dest++) = *(source++);
+
+	if ( YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_EOF_PENDING )
+		/* don't do the read, it's not guaranteed to return an EOF,
+		 * just force an EOF
+		 */
+		YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars = 0;
+
+	else
+		{
+			yy_size_t num_to_read =
+			YY_CURRENT_BUFFER_LVALUE->yy_buf_size - number_to_move - 1;
+
+		while ( num_to_read <= 0 )
+			{ /* Not enough room in the buffer - grow it. */
+
+			/* just a shorter name for the current buffer */
+			YY_BUFFER_STATE b = YY_CURRENT_BUFFER;
+
+			int yy_c_buf_p_offset =
+				(int) (yyg->yy_c_buf_p - b->yy_ch_buf);
+
+			if ( b->yy_is_our_buffer )
+				{
+				yy_size_t new_size = b->yy_buf_size * 2;
+
+				if ( new_size <= 0 )
+					b->yy_buf_size += b->yy_buf_size / 8;
+				else
+					b->yy_buf_size *= 2;
+
+				b->yy_ch_buf = (char *)
+					/* Include room in for 2 EOB chars. */
+					linguist_yyrealloc((void *) b->yy_ch_buf,b->yy_buf_size + 2 ,yyscanner );
+				}
+			else
+				/* Can't grow it, we don't own it. */
+				b->yy_ch_buf = 0;
+
+			if ( ! b->yy_ch_buf )
+				YY_FATAL_ERROR(
+				"fatal error - scanner input buffer overflow" );
+
+			yyg->yy_c_buf_p = &b->yy_ch_buf[yy_c_buf_p_offset];
+
+			num_to_read = YY_CURRENT_BUFFER_LVALUE->yy_buf_size -
+						number_to_move - 1;
+
+			}
+
+		if ( num_to_read > YY_READ_BUF_SIZE )
+			num_to_read = YY_READ_BUF_SIZE;
+
+		/* Read in more data. */
+		YY_INPUT( (&YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[number_to_move]),
+			yyg->yy_n_chars, num_to_read );
+
+		YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars;
+		}
+
+	if ( yyg->yy_n_chars == 0 )
+		{
+		if ( number_to_move == YY_MORE_ADJ )
+			{
+			ret_val = EOB_ACT_END_OF_FILE;
+			linguist_yyrestart(yyin  ,yyscanner);
+			}
+
+		else
+			{
+			ret_val = EOB_ACT_LAST_MATCH;
+			YY_CURRENT_BUFFER_LVALUE->yy_buffer_status =
+				YY_BUFFER_EOF_PENDING;
+			}
+		}
+
+	else
+		ret_val = EOB_ACT_CONTINUE_SCAN;
+
+	if ((yy_size_t) (yyg->yy_n_chars + number_to_move) > YY_CURRENT_BUFFER_LVALUE->yy_buf_size) {
+		/* Extend the array by 50%, plus the number we really need. */
+		yy_size_t new_size = yyg->yy_n_chars + number_to_move + (yyg->yy_n_chars >> 1);
+		YY_CURRENT_BUFFER_LVALUE->yy_ch_buf = (char *) linguist_yyrealloc((void *) YY_CURRENT_BUFFER_LVALUE->yy_ch_buf,new_size ,yyscanner );
+		if ( ! YY_CURRENT_BUFFER_LVALUE->yy_ch_buf )
+			YY_FATAL_ERROR( "out of dynamic memory in yy_get_next_buffer()" );
+	}
+
+	yyg->yy_n_chars += number_to_move;
+	YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] = YY_END_OF_BUFFER_CHAR;
+	YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1] = YY_END_OF_BUFFER_CHAR;
+
+	yyg->yytext_ptr = &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[0];
+
+	return ret_val;
+}
+
+/* yy_get_previous_state - get the state just before the EOB char was reached */
+
+    static yy_state_type yy_get_previous_state (yyscan_t yyscanner)
+{
+	register yy_state_type yy_current_state;
+	register char *yy_cp;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	yy_current_state = yyg->yy_start;
+	yy_current_state += YY_AT_BOL();
+
+	for ( yy_cp = yyg->yytext_ptr + YY_MORE_ADJ; yy_cp < yyg->yy_c_buf_p; ++yy_cp )
+		{
+		register YY_CHAR yy_c = (*yy_cp ? yy_ec[YY_SC_TO_UI(*yy_cp)] : 1);
+		if ( yy_accept[yy_current_state] )
+			{
+			yyg->yy_last_accepting_state = yy_current_state;
+			yyg->yy_last_accepting_cpos = yy_cp;
+			}
+		while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state )
+			{
+			yy_current_state = (int) yy_def[yy_current_state];
+			if ( yy_current_state >= 114 )
+				yy_c = yy_meta[(unsigned int) yy_c];
+			}
+		yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c];
+		}
+
+	return yy_current_state;
+}
+
+/* yy_try_NUL_trans - try to make a transition on the NUL character
+ *
+ * synopsis
+ *	next_state = yy_try_NUL_trans( current_state );
+ */
+    static yy_state_type yy_try_NUL_trans  (yy_state_type yy_current_state , yyscan_t yyscanner)
+{
+	register int yy_is_jam;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; /* This var may be unused depending upon options. */
+	register char *yy_cp = yyg->yy_c_buf_p;
+
+	register YY_CHAR yy_c = 1;
+	if ( yy_accept[yy_current_state] )
+		{
+		yyg->yy_last_accepting_state = yy_current_state;
+		yyg->yy_last_accepting_cpos = yy_cp;
+		}
+	while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state )
+		{
+		yy_current_state = (int) yy_def[yy_current_state];
+		if ( yy_current_state >= 114 )
+			yy_c = yy_meta[(unsigned int) yy_c];
+		}
+	yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c];
+	yy_is_jam = (yy_current_state == 113);
+
+	return yy_is_jam ? 0 : yy_current_state;
+}
+
+#ifndef YY_NO_INPUT
+#ifdef __cplusplus
+    static int yyinput (yyscan_t yyscanner)
+#else
+    static int input  (yyscan_t yyscanner)
+#endif
+
+{
+	int c;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	*yyg->yy_c_buf_p = yyg->yy_hold_char;
+
+	if ( *yyg->yy_c_buf_p == YY_END_OF_BUFFER_CHAR )
+		{
+		/* yy_c_buf_p now points to the character we want to return.
+		 * If this occurs *before* the EOB characters, then it's a
+		 * valid NUL; if not, then we've hit the end of the buffer.
+		 */
+		if ( yyg->yy_c_buf_p < &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] )
+			/* This was really a NUL. */
+			*yyg->yy_c_buf_p = '\0';
+
+		else
+			{ /* need more input */
+			yy_size_t offset = yyg->yy_c_buf_p - yyg->yytext_ptr;
+			++yyg->yy_c_buf_p;
+
+			switch ( yy_get_next_buffer( yyscanner ) )
+				{
+				case EOB_ACT_LAST_MATCH:
+					/* This happens because yy_g_n_b()
+					 * sees that we've accumulated a
+					 * token and flags that we need to
+					 * try matching the token before
+					 * proceeding.  But for input(),
+					 * there's no matching to consider.
+					 * So convert the EOB_ACT_LAST_MATCH
+					 * to EOB_ACT_END_OF_FILE.
+					 */
+
+					/* Reset buffer status. */
+					linguist_yyrestart(yyin ,yyscanner);
+
+					/*FALLTHROUGH*/
+
+				case EOB_ACT_END_OF_FILE:
+					{
+					if ( linguist_yywrap(yyscanner ) )
+						return 0;
+
+					if ( ! yyg->yy_did_buffer_switch_on_eof )
+						YY_NEW_FILE;
+#ifdef __cplusplus
+					return yyinput(yyscanner);
+#else
+					return input(yyscanner);
+#endif
+					}
+
+				case EOB_ACT_CONTINUE_SCAN:
+					yyg->yy_c_buf_p = yyg->yytext_ptr + offset;
+					break;
+				}
+			}
+		}
+
+	c = *(unsigned char *) yyg->yy_c_buf_p;	/* cast for 8-bit char's */
+	*yyg->yy_c_buf_p = '\0';	/* preserve yytext */
+	yyg->yy_hold_char = *++yyg->yy_c_buf_p;
+
+	YY_CURRENT_BUFFER_LVALUE->yy_at_bol = (c == '\n');
+
+	return c;
+}
+#endif	/* ifndef YY_NO_INPUT */
+
+/** Immediately switch to a different input stream.
+ * @param input_file A readable stream.
+ * @param yyscanner The scanner object.
+ * @note This function does not reset the start condition to @c INITIAL .
+ */
+    void linguist_yyrestart  (FILE * input_file , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	if ( ! YY_CURRENT_BUFFER ){
+        linguist_yyensure_buffer_stack (yyscanner);
+		YY_CURRENT_BUFFER_LVALUE =
+            linguist_yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner);
+	}
+
+	linguist_yy_init_buffer(YY_CURRENT_BUFFER,input_file ,yyscanner);
+	linguist_yy_load_buffer_state(yyscanner );
+}
+
+/** Switch to a different input buffer.
+ * @param new_buffer The new input buffer.
+ * @param yyscanner The scanner object.
+ */
+    void linguist_yy_switch_to_buffer  (YY_BUFFER_STATE  new_buffer , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	/* TODO. We should be able to replace this entire function body
+	 * with
+	 *		linguist_yypop_buffer_state();
+	 *		linguist_yypush_buffer_state(new_buffer);
+     */
+	linguist_yyensure_buffer_stack (yyscanner);
+	if ( YY_CURRENT_BUFFER == new_buffer )
+		return;
+
+	if ( YY_CURRENT_BUFFER )
+		{
+		/* Flush out information for old buffer. */
+		*yyg->yy_c_buf_p = yyg->yy_hold_char;
+		YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p;
+		YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars;
+		}
+
+	YY_CURRENT_BUFFER_LVALUE = new_buffer;
+	linguist_yy_load_buffer_state(yyscanner );
+
+	/* We don't actually know whether we did this switch during
+	 * EOF (linguist_yywrap()) processing, but the only time this flag
+	 * is looked at is after linguist_yywrap() is called, so it's safe
+	 * to go ahead and always set it.
+	 */
+	yyg->yy_did_buffer_switch_on_eof = 1;
+}
+
+static void linguist_yy_load_buffer_state  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+	yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars;
+	yyg->yytext_ptr = yyg->yy_c_buf_p = YY_CURRENT_BUFFER_LVALUE->yy_buf_pos;
+	yyin = YY_CURRENT_BUFFER_LVALUE->yy_input_file;
+	yyg->yy_hold_char = *yyg->yy_c_buf_p;
+}
+
+/** Allocate and initialize an input buffer state.
+ * @param file A readable stream.
+ * @param size The character buffer size in bytes. When in doubt, use @c YY_BUF_SIZE.
+ * @param yyscanner The scanner object.
+ * @return the allocated buffer state.
+ */
+    YY_BUFFER_STATE linguist_yy_create_buffer  (FILE * file, int  size , yyscan_t yyscanner)
+{
+	YY_BUFFER_STATE b;
+    
+	b = (YY_BUFFER_STATE) linguist_yyalloc(sizeof( struct yy_buffer_state ) ,yyscanner );
+	if ( ! b )
+		YY_FATAL_ERROR( "out of dynamic memory in linguist_yy_create_buffer()" );
+
+	b->yy_buf_size = size;
+
+	/* yy_ch_buf has to be 2 characters longer than the size given because
+	 * we need to put in 2 end-of-buffer characters.
+	 */
+	b->yy_ch_buf = (char *) linguist_yyalloc(b->yy_buf_size + 2 ,yyscanner );
+	if ( ! b->yy_ch_buf )
+		YY_FATAL_ERROR( "out of dynamic memory in linguist_yy_create_buffer()" );
+
+	b->yy_is_our_buffer = 1;
+
+	linguist_yy_init_buffer(b,file ,yyscanner);
+
+	return b;
+}
+
+/** Destroy the buffer.
+ * @param b a buffer created with linguist_yy_create_buffer()
+ * @param yyscanner The scanner object.
+ */
+    void linguist_yy_delete_buffer (YY_BUFFER_STATE  b , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	if ( ! b )
+		return;
+
+	if ( b == YY_CURRENT_BUFFER ) /* Not sure if we should pop here. */
+		YY_CURRENT_BUFFER_LVALUE = (YY_BUFFER_STATE) 0;
+
+	if ( b->yy_is_our_buffer )
+		linguist_yyfree((void *) b->yy_ch_buf ,yyscanner );
+
+	linguist_yyfree((void *) b ,yyscanner );
+}
+
+/* Initializes or reinitializes a buffer.
+ * This function is sometimes called more than once on the same buffer,
+ * such as during a linguist_yyrestart() or at EOF.
+ */
+    static void linguist_yy_init_buffer  (YY_BUFFER_STATE  b, FILE * file , yyscan_t yyscanner)
+
+{
+	int oerrno = errno;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	linguist_yy_flush_buffer(b ,yyscanner);
+
+	b->yy_input_file = file;
+	b->yy_fill_buffer = 1;
+
+    /* If b is the current buffer, then linguist_yy_init_buffer was _probably_
+     * called from linguist_yyrestart() or through yy_get_next_buffer.
+     * In that case, we don't want to reset the lineno or column.
+     */
+    if (b != YY_CURRENT_BUFFER){
+        b->yy_bs_lineno = 1;
+        b->yy_bs_column = 0;
+    }
+
+        b->yy_is_interactive = 0;
+    
+	errno = oerrno;
+}
+
+/** Discard all buffered characters. On the next scan, YY_INPUT will be called.
+ * @param b the buffer state to be flushed, usually @c YY_CURRENT_BUFFER.
+ * @param yyscanner The scanner object.
+ */
+    void linguist_yy_flush_buffer (YY_BUFFER_STATE  b , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+	if ( ! b )
+		return;
+
+	b->yy_n_chars = 0;
+
+	/* We always need two end-of-buffer characters.  The first causes
+	 * a transition to the end-of-buffer state.  The second causes
+	 * a jam in that state.
+	 */
+	b->yy_ch_buf[0] = YY_END_OF_BUFFER_CHAR;
+	b->yy_ch_buf[1] = YY_END_OF_BUFFER_CHAR;
+
+	b->yy_buf_pos = &b->yy_ch_buf[0];
+
+	b->yy_at_bol = 1;
+	b->yy_buffer_status = YY_BUFFER_NEW;
+
+	if ( b == YY_CURRENT_BUFFER )
+		linguist_yy_load_buffer_state(yyscanner );
+}
+
+/** Pushes the new state onto the stack. The new state becomes
+ *  the current state. This function will allocate the stack
+ *  if necessary.
+ *  @param new_buffer The new state.
+ *  @param yyscanner The scanner object.
+ */
+void linguist_yypush_buffer_state (YY_BUFFER_STATE new_buffer , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+	if (new_buffer == NULL)
+		return;
+
+	linguist_yyensure_buffer_stack(yyscanner);
+
+	/* This block is copied from linguist_yy_switch_to_buffer. */
+	if ( YY_CURRENT_BUFFER )
+		{
+		/* Flush out information for old buffer. */
+		*yyg->yy_c_buf_p = yyg->yy_hold_char;
+		YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p;
+		YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars;
+		}
+
+	/* Only push if top exists. Otherwise, replace top. */
+	if (YY_CURRENT_BUFFER)
+		yyg->yy_buffer_stack_top++;
+	YY_CURRENT_BUFFER_LVALUE = new_buffer;
+
+	/* copied from linguist_yy_switch_to_buffer. */
+	linguist_yy_load_buffer_state(yyscanner );
+	yyg->yy_did_buffer_switch_on_eof = 1;
+}
+
+/** Removes and deletes the top of the stack, if present.
+ *  The next element becomes the new top.
+ *  @param yyscanner The scanner object.
+ */
+void linguist_yypop_buffer_state (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+	if (!YY_CURRENT_BUFFER)
+		return;
+
+	linguist_yy_delete_buffer(YY_CURRENT_BUFFER ,yyscanner);
+	YY_CURRENT_BUFFER_LVALUE = NULL;
+	if (yyg->yy_buffer_stack_top > 0)
+		--yyg->yy_buffer_stack_top;
+
+	if (YY_CURRENT_BUFFER) {
+		linguist_yy_load_buffer_state(yyscanner );
+		yyg->yy_did_buffer_switch_on_eof = 1;
+	}
+}
+
+/* Allocates the stack if it does not exist.
+ *  Guarantees space for at least one push.
+ */
+static void linguist_yyensure_buffer_stack (yyscan_t yyscanner)
+{
+	yy_size_t num_to_alloc;
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+	if (!yyg->yy_buffer_stack) {
+
+		/* First allocation is just for 2 elements, since we don't know if this
+		 * scanner will even need a stack. We use 2 instead of 1 to avoid an
+		 * immediate realloc on the next call.
+         */
+		num_to_alloc = 1;
+		yyg->yy_buffer_stack = (struct yy_buffer_state**)linguist_yyalloc
+								(num_to_alloc * sizeof(struct yy_buffer_state*)
+								, yyscanner);
+		if ( ! yyg->yy_buffer_stack )
+			YY_FATAL_ERROR( "out of dynamic memory in linguist_yyensure_buffer_stack()" );
+								  
+		memset(yyg->yy_buffer_stack, 0, num_to_alloc * sizeof(struct yy_buffer_state*));
+				
+		yyg->yy_buffer_stack_max = num_to_alloc;
+		yyg->yy_buffer_stack_top = 0;
+		return;
+	}
+
+	if (yyg->yy_buffer_stack_top >= (yyg->yy_buffer_stack_max) - 1){
+
+		/* Increase the buffer to prepare for a possible push. */
+		int grow_size = 8 /* arbitrary grow size */;
+
+		num_to_alloc = yyg->yy_buffer_stack_max + grow_size;
+		yyg->yy_buffer_stack = (struct yy_buffer_state**)linguist_yyrealloc
+								(yyg->yy_buffer_stack,
+								num_to_alloc * sizeof(struct yy_buffer_state*)
+								, yyscanner);
+		if ( ! yyg->yy_buffer_stack )
+			YY_FATAL_ERROR( "out of dynamic memory in linguist_yyensure_buffer_stack()" );
+
+		/* zero only the new slots.*/
+		memset(yyg->yy_buffer_stack + yyg->yy_buffer_stack_max, 0, grow_size * sizeof(struct yy_buffer_state*));
+		yyg->yy_buffer_stack_max = num_to_alloc;
+	}
+}
+
+/** Setup the input buffer state to scan directly from a user-specified character buffer.
+ * @param base the character buffer
+ * @param size the size in bytes of the character buffer
+ * @param yyscanner The scanner object.
+ * @return the newly allocated buffer state object. 
+ */
+YY_BUFFER_STATE linguist_yy_scan_buffer  (char * base, yy_size_t  size , yyscan_t yyscanner)
+{
+	YY_BUFFER_STATE b;
+    
+	if ( size < 2 ||
+	     base[size-2] != YY_END_OF_BUFFER_CHAR ||
+	     base[size-1] != YY_END_OF_BUFFER_CHAR )
+		/* They forgot to leave room for the EOB's. */
+		return 0;
+
+	b = (YY_BUFFER_STATE) linguist_yyalloc(sizeof( struct yy_buffer_state ) ,yyscanner );
+	if ( ! b )
+		YY_FATAL_ERROR( "out of dynamic memory in linguist_yy_scan_buffer()" );
+
+	b->yy_buf_size = size - 2;	/* "- 2" to take care of EOB's */
+	b->yy_buf_pos = b->yy_ch_buf = base;
+	b->yy_is_our_buffer = 0;
+	b->yy_input_file = 0;
+	b->yy_n_chars = b->yy_buf_size;
+	b->yy_is_interactive = 0;
+	b->yy_at_bol = 1;
+	b->yy_fill_buffer = 0;
+	b->yy_buffer_status = YY_BUFFER_NEW;
+
+	linguist_yy_switch_to_buffer(b ,yyscanner );
+
+	return b;
+}
+
+/** Setup the input buffer state to scan a string. The next call to linguist_yylex() will
+ * scan from a @e copy of @a str.
+ * @param yystr a NUL-terminated string to scan
+ * @param yyscanner The scanner object.
+ * @return the newly allocated buffer state object.
+ * @note If you want to scan bytes that may contain NUL values, then use
+ *       linguist_yy_scan_bytes() instead.
+ */
+YY_BUFFER_STATE linguist_yy_scan_string (yyconst char * yystr , yyscan_t yyscanner)
+{
+    
+	return linguist_yy_scan_bytes(yystr,strlen(yystr) ,yyscanner);
+}
+
+/** Setup the input buffer state to scan the given bytes. The next call to linguist_yylex() will
+ * scan from a @e copy of @a bytes.
+ * @param bytes the byte buffer to scan
+ * @param len the number of bytes in the buffer pointed to by @a bytes.
+ * @param yyscanner The scanner object.
+ * @return the newly allocated buffer state object.
+ */
+YY_BUFFER_STATE linguist_yy_scan_bytes  (yyconst char * yybytes, yy_size_t  _yybytes_len , yyscan_t yyscanner)
+{
+	YY_BUFFER_STATE b;
+	char *buf;
+	yy_size_t n, i;
+    
+	/* Get memory for full buffer, including space for trailing EOB's. */
+	n = _yybytes_len + 2;
+	buf = (char *) linguist_yyalloc(n ,yyscanner );
+	if ( ! buf )
+		YY_FATAL_ERROR( "out of dynamic memory in linguist_yy_scan_bytes()" );
+
+	for ( i = 0; i < _yybytes_len; ++i )
+		buf[i] = yybytes[i];
+
+	buf[_yybytes_len] = buf[_yybytes_len+1] = YY_END_OF_BUFFER_CHAR;
+
+	b = linguist_yy_scan_buffer(buf,n ,yyscanner);
+	if ( ! b )
+		YY_FATAL_ERROR( "bad buffer in linguist_yy_scan_bytes()" );
+
+	/* It's okay to grow etc. this buffer, and we should throw it
+	 * away when we're done.
+	 */
+	b->yy_is_our_buffer = 1;
+
+	return b;
+}
+
+#ifndef YY_EXIT_FAILURE
+#define YY_EXIT_FAILURE 2
+#endif
+
+static void yy_fatal_error (yyconst char* msg , yyscan_t yyscanner)
+{
+    	(void) fprintf( stderr, "%s\n", msg );
+	exit( YY_EXIT_FAILURE );
+}
+
+/* Redefine yyless() so it works in section 3 code. */
+
+#undef yyless
+#define yyless(n) \
+	do \
+		{ \
+		/* Undo effects of setting up yytext. */ \
+        int yyless_macro_arg = (n); \
+        YY_LESS_LINENO(yyless_macro_arg);\
+		yytext[yyleng] = yyg->yy_hold_char; \
+		yyg->yy_c_buf_p = yytext + yyless_macro_arg; \
+		yyg->yy_hold_char = *yyg->yy_c_buf_p; \
+		*yyg->yy_c_buf_p = '\0'; \
+		yyleng = yyless_macro_arg; \
+		} \
+	while ( 0 )
+
+/* Accessor  methods (get/set functions) to struct members. */
+
+/** Get the user-defined data for this scanner.
+ * @param yyscanner The scanner object.
+ */
+YY_EXTRA_TYPE linguist_yyget_extra  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yyextra;
+}
+
+/** Get the current line number.
+ * @param yyscanner The scanner object.
+ */
+int linguist_yyget_lineno  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    
+        if (! YY_CURRENT_BUFFER)
+            return 0;
+    
+    return yylineno;
+}
+
+/** Get the current column number.
+ * @param yyscanner The scanner object.
+ */
+int linguist_yyget_column  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    
+        if (! YY_CURRENT_BUFFER)
+            return 0;
+    
+    return yycolumn;
+}
+
+/** Get the input stream.
+ * @param yyscanner The scanner object.
+ */
+FILE *linguist_yyget_in  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yyin;
+}
+
+/** Get the output stream.
+ * @param yyscanner The scanner object.
+ */
+FILE *linguist_yyget_out  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yyout;
+}
+
+/** Get the length of the current token.
+ * @param yyscanner The scanner object.
+ */
+yy_size_t linguist_yyget_leng  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yyleng;
+}
+
+/** Get the current token.
+ * @param yyscanner The scanner object.
+ */
+
+char *linguist_yyget_text  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yytext;
+}
+
+/** Set the user-defined data. This data is never touched by the scanner.
+ * @param user_defined The data to be associated with this scanner.
+ * @param yyscanner The scanner object.
+ */
+void linguist_yyset_extra (YY_EXTRA_TYPE  user_defined , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    yyextra = user_defined ;
+}
+
+/** Set the current line number.
+ * @param line_number
+ * @param yyscanner The scanner object.
+ */
+void linguist_yyset_lineno (int  line_number , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+        /* lineno is only valid if an input buffer exists. */
+        if (! YY_CURRENT_BUFFER )
+           yy_fatal_error( "linguist_yyset_lineno called with no buffer" , yyscanner); 
+    
+    yylineno = line_number;
+}
+
+/** Set the current column.
+ * @param line_number
+ * @param yyscanner The scanner object.
+ */
+void linguist_yyset_column (int  column_no , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+        /* column is only valid if an input buffer exists. */
+        if (! YY_CURRENT_BUFFER )
+           yy_fatal_error( "linguist_yyset_column called with no buffer" , yyscanner); 
+    
+    yycolumn = column_no;
+}
+
+/** Set the input stream. This does not discard the current
+ * input buffer.
+ * @param in_str A readable stream.
+ * @param yyscanner The scanner object.
+ * @see linguist_yy_switch_to_buffer
+ */
+void linguist_yyset_in (FILE *  in_str , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    yyin = in_str ;
+}
+
+void linguist_yyset_out (FILE *  out_str , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    yyout = out_str ;
+}
+
+int linguist_yyget_debug  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    return yy_flex_debug;
+}
+
+void linguist_yyset_debug (int  bdebug , yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    yy_flex_debug = bdebug ;
+}
+
+/* Accessor methods for yylval and yylloc */
+
+/* User-visible API */
+
+/* linguist_yylex_init is special because it creates the scanner itself, so it is
+ * the ONLY reentrant function that doesn't take the scanner as the last argument.
+ * That's why we explicitly handle the declaration, instead of using our macros.
+ */
+
+int linguist_yylex_init(yyscan_t* ptr_yy_globals)
+
+{
+    if (ptr_yy_globals == NULL){
+        errno = EINVAL;
+        return 1;
+    }
+
+    *ptr_yy_globals = (yyscan_t) linguist_yyalloc ( sizeof( struct yyguts_t ), NULL );
+
+    if (*ptr_yy_globals == NULL){
+        errno = ENOMEM;
+        return 1;
+    }
+
+    /* By setting to 0xAA, we expose bugs in yy_init_globals. Leave at 0x00 for releases. */
+    memset(*ptr_yy_globals,0x00,sizeof(struct yyguts_t));
+
+    return yy_init_globals ( *ptr_yy_globals );
+}
+
+/* linguist_yylex_init_extra has the same functionality as linguist_yylex_init, but follows the
+ * convention of taking the scanner as the last argument. Note however, that
+ * this is a *pointer* to a scanner, as it will be allocated by this call (and
+ * is the reason, too, why this function also must handle its own declaration).
+ * The user defined value in the first argument will be available to linguist_yyalloc in
+ * the yyextra field.
+ */
+
+int linguist_yylex_init_extra(YY_EXTRA_TYPE yy_user_defined,yyscan_t* ptr_yy_globals )
+
+{
+    struct yyguts_t dummy_yyguts;
+
+    linguist_yyset_extra (yy_user_defined, &dummy_yyguts);
+
+    if (ptr_yy_globals == NULL){
+        errno = EINVAL;
+        return 1;
+    }
+	
+    *ptr_yy_globals = (yyscan_t) linguist_yyalloc ( sizeof( struct yyguts_t ), &dummy_yyguts );
+	
+    if (*ptr_yy_globals == NULL){
+        errno = ENOMEM;
+        return 1;
+    }
+    
+    /* By setting to 0xAA, we expose bugs in
+    yy_init_globals. Leave at 0x00 for releases. */
+    memset(*ptr_yy_globals,0x00,sizeof(struct yyguts_t));
+    
+    linguist_yyset_extra (yy_user_defined, *ptr_yy_globals);
+    
+    return yy_init_globals ( *ptr_yy_globals );
+}
+
+static int yy_init_globals (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+    /* Initialization is the same as for the non-reentrant scanner.
+     * This function is called from linguist_yylex_destroy(), so don't allocate here.
+     */
+
+    yyg->yy_buffer_stack = 0;
+    yyg->yy_buffer_stack_top = 0;
+    yyg->yy_buffer_stack_max = 0;
+    yyg->yy_c_buf_p = (char *) 0;
+    yyg->yy_init = 0;
+    yyg->yy_start = 0;
+
+    yyg->yy_start_stack_ptr = 0;
+    yyg->yy_start_stack_depth = 0;
+    yyg->yy_start_stack =  NULL;
+
+/* Defined in main.c */
+#ifdef YY_STDINIT
+    yyin = stdin;
+    yyout = stdout;
+#else
+    yyin = (FILE *) 0;
+    yyout = (FILE *) 0;
+#endif
+
+    /* For future reference: Set errno on error, since we are called by
+     * linguist_yylex_init()
+     */
+    return 0;
+}
+
+/* linguist_yylex_destroy is for both reentrant and non-reentrant scanners. */
+int linguist_yylex_destroy  (yyscan_t yyscanner)
+{
+    struct yyguts_t * yyg = (struct yyguts_t*)yyscanner;
+
+    /* Pop the buffer stack, destroying each element. */
+	while(YY_CURRENT_BUFFER){
+		linguist_yy_delete_buffer(YY_CURRENT_BUFFER ,yyscanner );
+		YY_CURRENT_BUFFER_LVALUE = NULL;
+		linguist_yypop_buffer_state(yyscanner);
+	}
+
+	/* Destroy the stack itself. */
+	linguist_yyfree(yyg->yy_buffer_stack ,yyscanner);
+	yyg->yy_buffer_stack = NULL;
+
+    /* Destroy the start condition stack. */
+        linguist_yyfree(yyg->yy_start_stack ,yyscanner );
+        yyg->yy_start_stack = NULL;
+
+    /* Reset the globals. This is important in a non-reentrant scanner so the next time
+     * linguist_yylex() is called, initialization will occur. */
+    yy_init_globals( yyscanner);
+
+    /* Destroy the main struct (reentrant only). */
+    linguist_yyfree ( yyscanner , yyscanner );
+    yyscanner = NULL;
+    return 0;
+}
+
+/*
+ * Internal utility routines.
+ */
+
+#ifndef yytext_ptr
+static void yy_flex_strncpy (char* s1, yyconst char * s2, int n , yyscan_t yyscanner)
+{
+	register int i;
+	for ( i = 0; i < n; ++i )
+		s1[i] = s2[i];
+}
+#endif
+
+#ifdef YY_NEED_STRLEN
+static int yy_flex_strlen (yyconst char * s , yyscan_t yyscanner)
+{
+	register int n;
+	for ( n = 0; s[n]; ++n )
+		;
+
+	return n;
+}
+#endif
+
+void *linguist_yyalloc (yy_size_t  size , yyscan_t yyscanner)
+{
+	return (void *) malloc( size );
+}
+
+void *linguist_yyrealloc  (void * ptr, yy_size_t  size , yyscan_t yyscanner)
+{
+	/* The cast to (char *) in the following accommodates both
+	 * implementations that use char* generic pointers, and those
+	 * that use void* generic pointers.  It works with the latter
+	 * because both ANSI C and C++ allow castless assignment from
+	 * any pointer type to void*, and deal with argument conversions
+	 * as though doing an assignment.
+	 */
+	return (void *) realloc( (char *) ptr, size );
+}
+
+void linguist_yyfree (void * ptr , yyscan_t yyscanner)
+{
+	free( (char *) ptr );	/* see linguist_yyrealloc() for (char *) cast */
+}
+
+#define YYTABLES_NAME "yytables"
+
+#line 118 "tokenizer.l"
+
+
+
+
diff --git a/internal/tokenizer/flex/lex.linguist_yy.h b/internal/tokenizer/flex/lex.linguist_yy.h
new file mode 100644
index 00000000..dc5ba8fa
--- /dev/null
+++ b/internal/tokenizer/flex/lex.linguist_yy.h
@@ -0,0 +1,336 @@
+#ifndef linguist_yyHEADER_H
+#define linguist_yyHEADER_H 1
+#define linguist_yyIN_HEADER 1
+
+#line 6 "lex.linguist_yy.h"
+
+#define  YY_INT_ALIGNED short int
+
+/* A lexical scanner generated by flex */
+
+#define FLEX_SCANNER
+#define YY_FLEX_MAJOR_VERSION 2
+#define YY_FLEX_MINOR_VERSION 5
+#define YY_FLEX_SUBMINOR_VERSION 35
+#if YY_FLEX_SUBMINOR_VERSION > 0
+#define FLEX_BETA
+#endif
+
+/* First, we deal with  platform-specific or compiler-specific issues. */
+
+/* begin standard C headers. */
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+
+/* end standard C headers. */
+
+/* flex integer type definitions */
+
+#ifndef FLEXINT_H
+#define FLEXINT_H
+
+/* C99 systems have <inttypes.h>. Non-C99 systems may or may not. */
+
+#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L
+
+/* C99 says to define __STDC_LIMIT_MACROS before including stdint.h,
+ * if you want the limit (max/min) macros for int types. 
+ */
+#ifndef __STDC_LIMIT_MACROS
+#define __STDC_LIMIT_MACROS 1
+#endif
+
+#include <inttypes.h>
+typedef int8_t flex_int8_t;
+typedef uint8_t flex_uint8_t;
+typedef int16_t flex_int16_t;
+typedef uint16_t flex_uint16_t;
+typedef int32_t flex_int32_t;
+typedef uint32_t flex_uint32_t;
+typedef uint64_t flex_uint64_t;
+#else
+typedef signed char flex_int8_t;
+typedef short int flex_int16_t;
+typedef int flex_int32_t;
+typedef unsigned char flex_uint8_t; 
+typedef unsigned short int flex_uint16_t;
+typedef unsigned int flex_uint32_t;
+#endif /* ! C99 */
+
+/* Limits of integral types. */
+#ifndef INT8_MIN
+#define INT8_MIN               (-128)
+#endif
+#ifndef INT16_MIN
+#define INT16_MIN              (-32767-1)
+#endif
+#ifndef INT32_MIN
+#define INT32_MIN              (-2147483647-1)
+#endif
+#ifndef INT8_MAX
+#define INT8_MAX               (127)
+#endif
+#ifndef INT16_MAX
+#define INT16_MAX              (32767)
+#endif
+#ifndef INT32_MAX
+#define INT32_MAX              (2147483647)
+#endif
+#ifndef UINT8_MAX
+#define UINT8_MAX              (255U)
+#endif
+#ifndef UINT16_MAX
+#define UINT16_MAX             (65535U)
+#endif
+#ifndef UINT32_MAX
+#define UINT32_MAX             (4294967295U)
+#endif
+
+#endif /* ! FLEXINT_H */
+
+#ifdef __cplusplus
+
+/* The "const" storage-class-modifier is valid. */
+#define YY_USE_CONST
+
+#else	/* ! __cplusplus */
+
+/* C99 requires __STDC__ to be defined as 1. */
+#if defined (__STDC__)
+
+#define YY_USE_CONST
+
+#endif	/* defined (__STDC__) */
+#endif	/* ! __cplusplus */
+
+#ifdef YY_USE_CONST
+#define yyconst const
+#else
+#define yyconst
+#endif
+
+/* An opaque pointer. */
+#ifndef YY_TYPEDEF_YY_SCANNER_T
+#define YY_TYPEDEF_YY_SCANNER_T
+typedef void* yyscan_t;
+#endif
+
+/* For convenience, these vars (plus the bison vars far below)
+   are macros in the reentrant scanner. */
+#define yyin yyg->yyin_r
+#define yyout yyg->yyout_r
+#define yyextra yyg->yyextra_r
+#define yyleng yyg->yyleng_r
+#define yytext yyg->yytext_r
+#define yylineno (YY_CURRENT_BUFFER_LVALUE->yy_bs_lineno)
+#define yycolumn (YY_CURRENT_BUFFER_LVALUE->yy_bs_column)
+#define yy_flex_debug yyg->yy_flex_debug_r
+
+/* Size of default input buffer. */
+#ifndef YY_BUF_SIZE
+#define YY_BUF_SIZE 16384
+#endif
+
+#ifndef YY_TYPEDEF_YY_BUFFER_STATE
+#define YY_TYPEDEF_YY_BUFFER_STATE
+typedef struct yy_buffer_state *YY_BUFFER_STATE;
+#endif
+
+#ifndef YY_TYPEDEF_YY_SIZE_T
+#define YY_TYPEDEF_YY_SIZE_T
+typedef size_t yy_size_t;
+#endif
+
+#ifndef YY_STRUCT_YY_BUFFER_STATE
+#define YY_STRUCT_YY_BUFFER_STATE
+struct yy_buffer_state
+	{
+	FILE *yy_input_file;
+
+	char *yy_ch_buf;		/* input buffer */
+	char *yy_buf_pos;		/* current position in input buffer */
+
+	/* Size of input buffer in bytes, not including room for EOB
+	 * characters.
+	 */
+	yy_size_t yy_buf_size;
+
+	/* Number of characters read into yy_ch_buf, not including EOB
+	 * characters.
+	 */
+	yy_size_t yy_n_chars;
+
+	/* Whether we "own" the buffer - i.e., we know we created it,
+	 * and can realloc() it to grow it, and should free() it to
+	 * delete it.
+	 */
+	int yy_is_our_buffer;
+
+	/* Whether this is an "interactive" input source; if so, and
+	 * if we're using stdio for input, then we want to use getc()
+	 * instead of fread(), to make sure we stop fetching input after
+	 * each newline.
+	 */
+	int yy_is_interactive;
+
+	/* Whether we're considered to be at the beginning of a line.
+	 * If so, '^' rules will be active on the next match, otherwise
+	 * not.
+	 */
+	int yy_at_bol;
+
+    int yy_bs_lineno; /**< The line count. */
+    int yy_bs_column; /**< The column count. */
+    
+	/* Whether to try to fill the input buffer when we reach the
+	 * end of it.
+	 */
+	int yy_fill_buffer;
+
+	int yy_buffer_status;
+
+	};
+#endif /* !YY_STRUCT_YY_BUFFER_STATE */
+
+void linguist_yyrestart (FILE *input_file ,yyscan_t yyscanner );
+void linguist_yy_switch_to_buffer (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_create_buffer (FILE *file,int size ,yyscan_t yyscanner );
+void linguist_yy_delete_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner );
+void linguist_yy_flush_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner );
+void linguist_yypush_buffer_state (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner );
+void linguist_yypop_buffer_state (yyscan_t yyscanner );
+
+YY_BUFFER_STATE linguist_yy_scan_buffer (char *base,yy_size_t size ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_scan_string (yyconst char *yy_str ,yyscan_t yyscanner );
+YY_BUFFER_STATE linguist_yy_scan_bytes (yyconst char *bytes,yy_size_t len ,yyscan_t yyscanner );
+
+void *linguist_yyalloc (yy_size_t ,yyscan_t yyscanner );
+void *linguist_yyrealloc (void *,yy_size_t ,yyscan_t yyscanner );
+void linguist_yyfree (void * ,yyscan_t yyscanner );
+
+/* Begin user sect3 */
+
+#define yytext_ptr yytext_r
+
+#ifdef YY_HEADER_EXPORT_START_CONDITIONS
+#define INITIAL 0
+#define sgml 1
+#define c_comment 2
+#define xml_comment 3
+#define haskell_comment 4
+#define ocaml_comment 5
+#define python_dcomment 6
+#define python_scomment 7
+
+#endif
+
+#ifndef YY_NO_UNISTD_H
+/* Special case for "unistd.h", since it is non-ANSI. We include it way
+ * down here because we want the user's section 1 to have been scanned first.
+ * The user has a chance to override it with an option.
+ */
+#include <unistd.h>
+#endif
+
+#define YY_EXTRA_TYPE struct tokenizer_extra *
+
+int linguist_yylex_init (yyscan_t* scanner);
+
+int linguist_yylex_init_extra (YY_EXTRA_TYPE user_defined,yyscan_t* scanner);
+
+/* Accessor methods to globals.
+   These are made visible to non-reentrant scanners for convenience. */
+
+int linguist_yylex_destroy (yyscan_t yyscanner );
+
+int linguist_yyget_debug (yyscan_t yyscanner );
+
+void linguist_yyset_debug (int debug_flag ,yyscan_t yyscanner );
+
+YY_EXTRA_TYPE linguist_yyget_extra (yyscan_t yyscanner );
+
+void linguist_yyset_extra (YY_EXTRA_TYPE user_defined ,yyscan_t yyscanner );
+
+FILE *linguist_yyget_in (yyscan_t yyscanner );
+
+void linguist_yyset_in  (FILE * in_str ,yyscan_t yyscanner );
+
+FILE *linguist_yyget_out (yyscan_t yyscanner );
+
+void linguist_yyset_out  (FILE * out_str ,yyscan_t yyscanner );
+
+yy_size_t linguist_yyget_leng (yyscan_t yyscanner );
+
+char *linguist_yyget_text (yyscan_t yyscanner );
+
+int linguist_yyget_lineno (yyscan_t yyscanner );
+
+void linguist_yyset_lineno (int line_number ,yyscan_t yyscanner );
+
+/* Macros after this point can all be overridden by user definitions in
+ * section 1.
+ */
+
+#ifndef YY_SKIP_YYWRAP
+#ifdef __cplusplus
+extern "C" int linguist_yywrap (yyscan_t yyscanner );
+#else
+extern int linguist_yywrap (yyscan_t yyscanner );
+#endif
+#endif
+
+#ifndef yytext_ptr
+static void yy_flex_strncpy (char *,yyconst char *,int ,yyscan_t yyscanner);
+#endif
+
+#ifdef YY_NEED_STRLEN
+static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner);
+#endif
+
+#ifndef YY_NO_INPUT
+
+#endif
+
+/* Amount of stuff to slurp up with each read. */
+#ifndef YY_READ_BUF_SIZE
+#define YY_READ_BUF_SIZE 8192
+#endif
+
+/* Number of entries by which start-condition stack grows. */
+#ifndef YY_START_STACK_INCR
+#define YY_START_STACK_INCR 25
+#endif
+
+/* Default declaration of generated scanner - a define so the user can
+ * easily add parameters.
+ */
+#ifndef YY_DECL
+#define YY_DECL_IS_OURS 1
+
+extern int linguist_yylex (yyscan_t yyscanner);
+
+#define YY_DECL int linguist_yylex (yyscan_t yyscanner)
+#endif /* !YY_DECL */
+
+/* yy_get_previous_state - get the state just before the EOB char was reached */
+
+#undef YY_NEW_FILE
+#undef YY_FLUSH_BUFFER
+#undef yy_set_bol
+#undef yy_new_buffer
+#undef yy_set_interactive
+#undef YY_DO_BEFORE_ACTION
+
+#ifdef YY_DECL_IS_OURS
+#undef YY_DECL_IS_OURS
+#undef YY_DECL
+#endif
+
+#line 118 "tokenizer.l"
+
+
+#line 335 "lex.linguist_yy.h"
+#undef linguist_yyIN_HEADER
+#endif /* linguist_yyHEADER_H */
diff --git a/internal/tokenizer/flex/linguist.h b/internal/tokenizer/flex/linguist.h
new file mode 100644
index 00000000..eda17966
--- /dev/null
+++ b/internal/tokenizer/flex/linguist.h
@@ -0,0 +1,15 @@
+// https://github.com/github/linguist/blob/f72f2a21dfe80ebd16af3bc6216da75cd983a4f6/ext/linguist/linguist.h#L1
+enum tokenizer_type {
+  NO_ACTION,
+  REGULAR_TOKEN,
+  SHEBANG_TOKEN,
+  SGML_TOKEN,
+};
+
+struct tokenizer_extra {
+  char *token;
+  enum tokenizer_type type;
+};
+
+// TODO(bzz) port Win support from
+// https://github.com/github/linguist/commit/8e912b4d8bf2aef7948de59eba48b75cfcbc97e0
\ No newline at end of file
diff --git a/internal/tokenizer/flex/tokenize_c.go b/internal/tokenizer/flex/tokenize_c.go
new file mode 100644
index 00000000..411a9a57
--- /dev/null
+++ b/internal/tokenizer/flex/tokenize_c.go
@@ -0,0 +1,71 @@
+package flex
+
+// #include <stdlib.h>
+// #include "linguist.h"
+// #include "lex.linguist_yy.h"
+// int linguist_yywrap(yyscan_t yyscanner) {
+// 	return 1;
+// }
+import "C"
+import "unsafe"
+
+const maxTokenLen = 32 // bytes
+
+// TokenizeFlex implements tokenizer by calling Flex generated code from linguist in C
+// This is a transliteration from C https://github.com/github/linguist/blob/master/ext/linguist/linguist.c#L12
+func TokenizeFlex(content []byte) []string {
+	var buf C.YY_BUFFER_STATE
+	var scanner C.yyscan_t
+	var extra C.struct_tokenizer_extra
+	var _len C.ulong
+	var r C.int
+
+	_len = C.ulong(len(content))
+	cs := C.CBytes(content)
+	defer C.free(unsafe.Pointer(cs))
+
+	C.linguist_yylex_init_extra(&extra, &scanner)
+	buf = C.linguist_yy_scan_bytes((*C.char)(cs), _len, scanner)
+
+	ary := []string{}
+	for {
+		extra._type = C.NO_ACTION
+		extra.token = nil
+		r = C.linguist_yylex(scanner)
+		switch extra._type {
+		case C.NO_ACTION:
+			break
+		case C.REGULAR_TOKEN:
+			_len = C.strlen(extra.token)
+			if _len <= maxTokenLen {
+				ary = append(ary, C.GoStringN(extra.token, (C.int)(_len)))
+			}
+			C.free(unsafe.Pointer(extra.token))
+			break
+		case C.SHEBANG_TOKEN:
+			_len = C.strlen(extra.token)
+			if _len <= maxTokenLen {
+				s := "SHEBANG#!" + C.GoStringN(extra.token, (C.int)(_len))
+				ary = append(ary, s)
+			}
+			C.free(unsafe.Pointer(extra.token))
+			break
+		case C.SGML_TOKEN:
+			_len = C.strlen(extra.token)
+			if _len <= maxTokenLen {
+				s := C.GoStringN(extra.token, (C.int)(_len)) + ">"
+				ary = append(ary, s)
+			}
+			C.free(unsafe.Pointer(extra.token))
+			break
+		}
+		if r == 0 {
+			break
+		}
+	}
+
+	C.linguist_yy_delete_buffer(buf, scanner)
+	C.linguist_yylex_destroy(scanner)
+
+	return ary
+}
diff --git a/internal/tokenizer/tokenize.go b/internal/tokenizer/tokenize.go
index 6a721c46..e7303bc4 100644
--- a/internal/tokenizer/tokenize.go
+++ b/internal/tokenizer/tokenize.go
@@ -1,6 +1,5 @@
-// Package tokenizer implements file tokenization used by the enry content
-// classifier. This package is an implementation detail of enry and should not
-// be imported by other packages.
+// +build !flex
+
 package tokenizer
 
 import (
@@ -9,14 +8,14 @@ import (
 	"gopkg.in/src-d/enry.v1/regex"
 )
 
-const byteLimit = 100000
-
-// Tokenize returns language-agnostic lexical tokens from content. The tokens
-// returned should match what the Linguist library returns. At most the first
-// 100KB of content are tokenized.
+// Tokenize returns lexical tokens from content. The tokens returned match what
+// the Linguist library returns. At most the first ByteLimit bytes of content are tokenized.
+//
+// BUG: Until https://github.com/src-d/enry/issues/193 is resolved, there are some
+// differences between this function and the Linguist output.
 func Tokenize(content []byte) []string {
-	if len(content) > byteLimit {
-		content = content[:byteLimit]
+	if len(content) > ByteLimit {
+		content = content[:ByteLimit]
 	}
 
 	// Copy the input so that changes wrought by the tokenization steps do not
diff --git a/internal/tokenizer/tokenize_c.go b/internal/tokenizer/tokenize_c.go
new file mode 100644
index 00000000..2d640abc
--- /dev/null
+++ b/internal/tokenizer/tokenize_c.go
@@ -0,0 +1,15 @@
+// +build flex
+
+package tokenizer
+
+import "gopkg.in/src-d/enry.v1/internal/tokenizer/flex"
+
+// Tokenize returns lexical tokens from content. The tokens returned match what
+// the Linguist library returns. At most the first ByteLimit bytes of content are tokenized.
+func Tokenize(content []byte) []string {
+	if len(content) > ByteLimit {
+		content = content[:ByteLimit]
+	}
+
+	return flex.TokenizeFlex(content)
+}
diff --git a/internal/tokenizer/tokenize_test.go b/internal/tokenizer/tokenize_test.go
index 34f75db5..881df30b 100644
--- a/internal/tokenizer/tokenize_test.go
+++ b/internal/tokenizer/tokenize_test.go
@@ -5,6 +5,7 @@ import (
 	"testing"
 
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 )
 
 const (
@@ -89,27 +90,48 @@ var (
 		"XHTML", "sample", "file", "type", "#example", "background", "color", "yellow", "id", "Just", "a", "simple", "XHTML", "test", "page.",
 		"-", "|", "+", "&&", "<", "<", "-", "!", "!", "!", "=", "=", "!", ":", "=", ":", "=", ",", ",", "=", ">", ">", "=", "=", "=", "=", ">",
 		"'", ",", ">", "=", ">", "=", "=", ">", "=", ">", ":", ">", "=", ">"}
-)
 
-func TestTokenize(t *testing.T) {
-	tests := []struct {
+	tests = []struct {
 		name     string
 		content  []byte
 		expected []string
 	}{
 		{name: "content", content: []byte(testContent), expected: tokensFromTestContent},
 	}
+)
 
+func TestTokenize(t *testing.T) {
 	for _, test := range tests {
 		t.Run(test.name, func(t *testing.T) {
 			before := string(test.content)
 			tokens := Tokenize(test.content)
 			after := string(test.content)
-			assert.Equal(t, before, after, "the input slice was modified")
-			assert.Equal(t, len(test.expected), len(tokens), fmt.Sprintf("token' slice length = %v, want %v", len(test.expected), len(tokens)))
+			require.Equal(t, before, after, "the input slice was modified")
+			require.Equal(t, len(test.expected), len(tokens), fmt.Sprintf("token' slice length = %v, want %v", len(test.expected), len(tokens)))
 			for i, expectedToken := range test.expected {
 				assert.Equal(t, expectedToken, tokens[i], fmt.Sprintf("token = %v, want %v", tokens[i], expectedToken))
 			}
 		})
 	}
 }
+
+func BenchmarkTokenizer_BaselineCopy(b *testing.B) {
+	b.ReportAllocs()
+	for i := 0; i < b.N; i++ {
+		for _, test := range tests {
+			if len(test.content) > ByteLimit {
+				test.content = test.content[:ByteLimit]
+			}
+			_ = append([]byte(nil), test.content...)
+		}
+	}
+}
+
+func BenchmarkTokenizer(b *testing.B) {
+	b.ReportAllocs()
+	for i := 0; i < b.N; i++ {
+		for _, test := range tests {
+			Tokenize(test.content)
+		}
+	}
+}