XCALLOC -> CALLOC, XMALLOC -> MALLOC.

[bison.git] / src / scan-gram.l

diff --git a/src/scan-gram.l b/src/scan-gram.l
index da92baafe6ebe1c2262178dd338b43c125d0afc6..41aefd620440633787aba050e6f15c91d12a5439 100644 (file)
--- a/src/scan-gram.l
+++ b/src/scan-gram.l
@@ -24,136 +24,69 @@
 
 %{
 #include "system.h"
 
 %{
 #include "system.h"

-#include "mbswidth.h"
+
+#include <mbswidth.h>
+#include <get-errno.h>
+#include <quote.h>
+

 #include "complain.h"
 #include "complain.h"

-#include "quote.h"
-#include "struniq.h"
+#include "files.h"

 #include "getargs.h"
 #include "gram.h"
 #include "reader.h"
 #include "getargs.h"
 #include "gram.h"
 #include "reader.h"

+#include "uniqstr.h"

 
 

-/* Each time we match a string, move the end cursor to its end. */
-#define YY_USER_INIT                           \
-do {                                           \
-  LOCATION_RESET (*yylloc);                    \
-  yylloc->file = infile;                       \
-   /* This is only to avoid GCC warnings. */   \
-  if (yycontrol) {;};                          \
-} while (0)
-
-#define YY_USER_ACTION  extend_location (yylloc, yytext, yyleng);
-#define YY_STEP         LOCATION_STEP (*yylloc)
-
-#define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size))
-
-
-/* Read bytes from FP into buffer BUF of size SIZE.  Return the
-   number of bytes read.  Remove '\r' from input, treating \r\n
-   and isolated \r as \n.  */
-
-static size_t
-no_cr_read (FILE *fp, char *buf, size_t size)
-{
-  size_t s = fread (buf, 1, size, fp);
-  if (s)
-    {
-      char *w = memchr (buf, '\r', s);
-      if (w)
-       {
-         char const *r = ++w;
-         char const *lim = buf + s;
-
-         for (;;)
-           {
-             /* Found an '\r'.  Treat it like '\n', but ignore any
-                '\n' that immediately follows.  */
-             w[-1] = '\n';
-             if (r == lim)
-               {
-                 int ch = getc (fp);
-                 if (ch != '\n' && ungetc (ch, fp) != ch)
-                   break;
-               }
-             else if (*r == '\n')
-               r++;
-
-             /* Copy until the next '\r'.  */
-             do
-               {
-                 if (r == lim)
-                   return w - buf;
-               }
-             while ((*w++ = *r++) != '\r');
-           }
-
-         return w - buf;
-       }
-    }
-
-  return s;
-}
-
-
-/* Extend *LOC to account for token TOKEN of size SIZE.  */
-
-static void
-extend_location (location_t *loc, char const *token, int size)
-{
-  int line = loc->last_line;
-  int column = loc->last_column;
-  char const *p0 = token;
-  char const *p = token;
-  char const *lim = token + size;
-
-  for (p = token; p < lim; p++)
-    switch (*p)
-      {
-      case '\r':
-       /* \r shouldn't survive no_cr_read.  */
-       abort ();
-
-      case '\n':
-       line++;
-       column = 1;
-       p0 = p + 1;
-       break;
+#define YY_USER_INIT                                   \
+  do                                                   \
+    {                                                  \
+      scanner_cursor.file = current_file;              \
+      scanner_cursor.line = 1;                         \
+      scanner_cursor.column = 1;                       \
+    }                                                  \
+  while (0)

 
 

-      case '\t':
-       column += mbsnwidth (p0, p - p0, 0);
-       column += 8 - ((column - 1) & 7);
-       p0 = p + 1;
-       break;
-      }
+/* Location of scanner cursor.  */
+boundary scanner_cursor;

 
 

-  loc->last_line = line;
-  loc->last_column = column + mbsnwidth (p0, p - p0, 0);
-}
+static void adjust_location (location *, char const *, size_t);
+#define YY_USER_ACTION  adjust_location (loc, yytext, yyleng);

 
 

+static size_t no_cr_read (FILE *, char *, size_t);
+#define YY_INPUT(buf, result, size) ((result) = no_cr_read (yyin, buf, size))

 
 
 
 

-/* STRING_OBSTACK -- Used to store all the characters that we need to
+/* OBSTACK_FOR_STRING -- Used to store all the characters that we need to

    keep (to construct ID, STRINGS etc.).  Use the following macros to
    use it.
 
    keep (to construct ID, STRINGS etc.).  Use the following macros to
    use it.
 

-   Use YY_OBS_GROW to append what has just been matched, and
-   YY_OBS_FINISH to end the string (it puts the ending 0).
-   YY_OBS_FINISH also stores this string in LAST_STRING, which can be
-   used, and which is used by YY_OBS_FREE to free the last string.  */
+   Use STRING_GROW to append what has just been matched, and
+   STRING_FINISH to end the string (it puts the ending 0).
+   STRING_FINISH also stores this string in LAST_STRING, which can be
+   used, and which is used by STRING_FREE to free the last string.  */

 
 

-static struct obstack string_obstack;
+static struct obstack obstack_for_string;

 
 

-#define YY_OBS_GROW   \
-  obstack_grow (&string_obstack, yytext, yyleng)
+/* A string representing the most recently saved token.  */
+static char *last_string;

 
 

-#define YY_OBS_FINISH                                  \
+
+#define STRING_GROW   \
+  obstack_grow (&obstack_for_string, yytext, yyleng)
+
+#define STRING_FINISH                                  \

   do {                                                 \
   do {                                                 \

-    obstack_1grow (&string_obstack, '\0');             \
-    last_string = obstack_finish (&string_obstack);    \
+    obstack_1grow (&obstack_for_string, '\0');         \
+    last_string = obstack_finish (&obstack_for_string);        \

   } while (0)
 
   } while (0)
 

-#define YY_OBS_FREE \
-  obstack_free (&string_obstack, last_string)
+#define STRING_FREE \
+  obstack_free (&obstack_for_string, last_string)

 
 

+void
+scanner_last_string_free (void)
+{
+  STRING_FREE;
+}

 
 /* Within well-formed rules, RULE_LENGTH is the number of values in
    the current rule so far, which says where to find `$0' with respect
 
 /* Within well-formed rules, RULE_LENGTH is the number of values in
    the current rule so far, which says where to find `$0' with respect


@@ -163,17 +96,16 @@ static struct obstack string_obstack;
    Outside of well-formed rules, RULE_LENGTH has an undefined value.  */
 static int rule_length;
 
    Outside of well-formed rules, RULE_LENGTH has an undefined value.  */
 static int rule_length;
 

-static void handle_dollar (braced_code_t code_kind,
-                          char *cp, location_t location);
-static void handle_at (braced_code_t code_kind,
-                      char *cp, location_t location);
-static void handle_syncline (char *args, location_t *location);
+static void handle_dollar (braced_code code_kind, char *cp, location loc);
+static void handle_at (braced_code code_kind, char *cp, location loc);
+static void handle_syncline (char *args);

 static int convert_ucn_to_byte (char const *hex_text);
 static int convert_ucn_to_byte (char const *hex_text);

-static void unexpected_end_of_file (location_t *, char const *);
+static void unexpected_end_of_file (boundary, char const *);

 
 %}
 %x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT
 %x SC_STRING SC_CHARACTER
 
 %}
 %x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT
 %x SC_STRING SC_CHARACTER

+%x SC_AFTER_IDENTIFIER

 %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
 %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE
 
 %x SC_ESCAPED_STRING SC_ESCAPED_CHARACTER
 %x SC_BRACED_CODE SC_PROLOGUE SC_EPILOGUE
 


@@ -196,18 +128,41 @@ splice     (\\[ \f\t\v]*\n)*
   /* Nesting level of the current code in braces.  */
   int braces_level IF_LINT (= 0);
 
   /* Nesting level of the current code in braces.  */
   int braces_level IF_LINT (= 0);
 

-  /* Scanner context when scanning C code.  */
-  int c_context IF_LINT (= 0);
+  /* Parent context state, when applicable.  */
+  int context_state IF_LINT (= 0);
+
+  /* Location of most recent identifier, when applicable.  */
+  location id_loc IF_LINT (= *loc);

 
 

-  /* A string representing the most recently saved token.  */
-  char *last_string;
+  /* Where containing code started, when applicable.  */
+  boundary code_start IF_LINT (= loc->start);

 
 

-  /* At each yylex invocation, mark the current position as the
-     start of the next token.  */
-  YY_STEP;
+  /* Where containing comment or string or character literal started,
+     when applicable.  */
+  boundary token_start IF_LINT (= loc->start);

 %}
 
 
 %}
 
 

+  /*-----------------------.
+  | Scanning white space.  |
+  `-----------------------*/
+
+<INITIAL,SC_AFTER_IDENTIFIER>
+{
+  [ \f\n\t\v]  ;
+
+  /* Comments. */
+  "/*"         token_start = loc->start; context_state = YY_START; BEGIN SC_YACC_COMMENT;
+  "//".*       ;
+
+  /* #line directives are not documented, and may be withdrawn or
+     modified in future versions of Bison.  */
+  ^"#line "{int}" \"".*"\"\n" {
+    handle_syncline (yytext + sizeof "#line " - 1);
+  }
+}
+
+

   /*----------------------------.
   | Scanning Bison directives.  |
   `----------------------------*/
   /*----------------------------.
   | Scanning Bison directives.  |
   `----------------------------*/


@@ -248,75 +203,103 @@ splice    (\\[ \f\t\v]*\n)*
   "%verbose"              return PERCENT_VERBOSE;
   "%yacc"                 return PERCENT_YACC;
 
   "%verbose"              return PERCENT_VERBOSE;
   "%yacc"                 return PERCENT_YACC;
 

-  {directive}             {
-    complain_at (*yylloc, _("invalid directive: %s"), quote (yytext));
-    YY_STEP;
+  {directive} {
+    complain_at (*loc, _("invalid directive: %s"), quote (yytext));

   }
 
   }
 

-  ^"#line "{int}" \""[^\"]*"\"\n"   handle_syncline (yytext + strlen ("#line "), yylloc); YY_STEP;
-

   "="                     return EQUAL;
   "="                     return EQUAL;

-  ":"                     rule_length = 0; return COLON;

   "|"                     rule_length = 0; return PIPE;
   "|"                     rule_length = 0; return PIPE;

-  ","                     return COMMA;

   ";"                     return SEMICOLON;
 
   ";"                     return SEMICOLON;
 

-  [ \f\n\t\v]  YY_STEP;
+  "," {
+    warn_at (*loc, _("stray `,' treated as white space"));
+  }

 
 

-  {id}        {
-    yylval->symbol = symbol_get (yytext, *yylloc);
+  {id} {
+    val->symbol = symbol_get (yytext, *loc);
+    id_loc = *loc;

     rule_length++;
     rule_length++;

-    return ID;
+    BEGIN SC_AFTER_IDENTIFIER;

   }
 
   {int} {
     unsigned long num;
   }
 
   {int} {
     unsigned long num;

-    errno = 0;
+    set_errno (0);

     num = strtoul (yytext, 0, 10);
     num = strtoul (yytext, 0, 10);

-    if (INT_MAX < num || errno)
+    if (INT_MAX < num || get_errno ())

       {
       {

-       complain_at (*yylloc, _("integer out of range: %s"), quote (yytext));
+       complain_at (*loc, _("integer out of range: %s"), quote (yytext));

        num = INT_MAX;
       }
        num = INT_MAX;
       }

-    yylval->integer = num;
+    val->integer = num;

     return INT;
   }
 
   /* Characters.  We don't check there is only one.  */
     return INT;
   }
 
   /* Characters.  We don't check there is only one.  */

-  "'"         YY_OBS_GROW; BEGIN SC_ESCAPED_CHARACTER;
+  "'"        STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER;

 
   /* Strings. */
 
   /* Strings. */

-  "\""        YY_OBS_GROW; BEGIN SC_ESCAPED_STRING;
-
-  /* Comments. */
-  "/*"        BEGIN SC_YACC_COMMENT;
-  "//".*      YY_STEP;
+  "\""       STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_STRING;

 
   /* Prologue. */
 
   /* Prologue. */

-  "%{"        BEGIN SC_PROLOGUE;
+  "%{"        code_start = loc->start; BEGIN SC_PROLOGUE;

 
   /* Code in between braces.  */
 
   /* Code in between braces.  */

-  "{"         YY_OBS_GROW; braces_level = 0; BEGIN SC_BRACED_CODE;
+  "{" {
+    STRING_GROW;
+    braces_level = 0;
+    code_start = loc->start;
+    BEGIN SC_BRACED_CODE;
+  }

 
   /* A type. */
   "<"{tag}">" {
 
   /* A type. */
   "<"{tag}">" {

-    obstack_grow (&string_obstack, yytext + 1, yyleng - 2);
-    YY_OBS_FINISH;
-    yylval->struniq = struniq_new (last_string);
-    YY_OBS_FREE;
+    obstack_grow (&obstack_for_string, yytext + 1, yyleng - 2);
+    STRING_FINISH;
+    val->uniqstr = uniqstr_new (last_string);
+    STRING_FREE;

     return TYPE;
   }
 
   "%%" {
     static int percent_percent_count;
     if (++percent_percent_count == 2)
     return TYPE;
   }
 
   "%%" {
     static int percent_percent_count;
     if (++percent_percent_count == 2)

-      BEGIN SC_EPILOGUE;
+      {
+       code_start = loc->start;
+       BEGIN SC_EPILOGUE;
+      }

     return PERCENT_PERCENT;
   }
 
   . {
     return PERCENT_PERCENT;
   }
 
   . {

-    complain_at (*yylloc, _("invalid character: %s"), quote (yytext));
-    YY_STEP;
+    complain_at (*loc, _("invalid character: %s"), quote (yytext));
+  }
+}
+
+
+  /*-----------------------------------------------------------------.
+  | Scanning after an identifier, checking whether a colon is next.  |
+  `-----------------------------------------------------------------*/
+
+<SC_AFTER_IDENTIFIER>
+{
+  ":" {
+    rule_length = 0;
+    *loc = id_loc;
+    BEGIN INITIAL;
+    return ID_COLON;
+  }
+  . {
+    scanner_cursor.column -= mbsnwidth (yytext, yyleng, 0);
+    yyless (0);
+    *loc = id_loc;
+    BEGIN INITIAL;
+    return ID;
+  }
+  <<EOF>> {
+    *loc = id_loc;
+    BEGIN INITIAL;
+    return ID;

   }
 }
 
   }
 }
 


@@ -327,13 +310,9 @@ splice      (\\[ \f\t\v]*\n)*
 
 <SC_YACC_COMMENT>
 {
 
 <SC_YACC_COMMENT>
 {

-  "*/" {
-    YY_STEP;
-    BEGIN INITIAL;
-  }
-
+  "*/"     BEGIN context_state;

   .|\n    ;
   .|\n    ;

-  <<EOF>>  unexpected_end_of_file (yylloc, "*/");
+  <<EOF>>  unexpected_end_of_file (token_start, "*/");

 }
 
 
 }
 
 


@@ -343,8 +322,8 @@ splice       (\\[ \f\t\v]*\n)*
 
 <SC_COMMENT>
 {
 
 <SC_COMMENT>
 {

-  "*"{splice}"/"  YY_OBS_GROW; BEGIN c_context;
-  <<EOF>>        unexpected_end_of_file (yylloc, "*/");
+  "*"{splice}"/"  STRING_GROW; BEGIN context_state;
+  <<EOF>>        unexpected_end_of_file (token_start, "*/");

 }
 
 
 }
 
 


@@ -354,9 +333,9 @@ splice       (\\[ \f\t\v]*\n)*
 
 <SC_LINE_COMMENT>
 {
 
 <SC_LINE_COMMENT>
 {

-  "\n"          YY_OBS_GROW; BEGIN c_context;
-  {splice}      YY_OBS_GROW;
-  <<EOF>>       BEGIN c_context;
+  "\n"          STRING_GROW; BEGIN context_state;
+  {splice}      STRING_GROW;
+  <<EOF>>       BEGIN context_state;

 }
 
 
 }
 
 


@@ -368,16 +347,17 @@ splice     (\\[ \f\t\v]*\n)*
 <SC_ESCAPED_STRING>
 {
   "\"" {
 <SC_ESCAPED_STRING>
 {
   "\"" {

-    YY_OBS_GROW;
-    YY_OBS_FINISH;
-    yylval->string = last_string;
+    STRING_GROW;
+    STRING_FINISH;
+    loc->start = token_start;
+    val->chars = last_string;

     rule_length++;
     BEGIN INITIAL;
     return STRING;
   }
 
     rule_length++;
     BEGIN INITIAL;
     return STRING;
   }
 

-  .|\n     YY_OBS_GROW;
-  <<EOF>>   unexpected_end_of_file (yylloc, "\"");
+  .|\n     STRING_GROW;
+  <<EOF>>   unexpected_end_of_file (token_start, "\"");

 }
 
   /*---------------------------------------------------------------.
 }
 
   /*---------------------------------------------------------------.


@@ -388,20 +368,21 @@ splice     (\\[ \f\t\v]*\n)*
 <SC_ESCAPED_CHARACTER>
 {
   "'" {
 <SC_ESCAPED_CHARACTER>
 {
   "'" {

-    YY_OBS_GROW;
-    YY_OBS_FINISH;
-    yylval->symbol = symbol_get (last_string, *yylloc);
-    symbol_class_set (yylval->symbol, token_sym, *yylloc);
-    symbol_user_token_number_set (yylval->symbol,
-                                 (unsigned char) last_string[1], *yylloc);
-    YY_OBS_FREE;
+    STRING_GROW;
+    STRING_FINISH;
+    loc->start = token_start;
+    val->symbol = symbol_get (last_string, *loc);
+    symbol_class_set (val->symbol, token_sym, *loc);
+    symbol_user_token_number_set (val->symbol,
+                                 (unsigned char) last_string[1], *loc);
+    STRING_FREE;

     rule_length++;
     BEGIN INITIAL;
     return ID;
   }
 
     rule_length++;
     BEGIN INITIAL;
     return ID;
   }
 

-  .|\n     YY_OBS_GROW;
-  <<EOF>>   unexpected_end_of_file (yylloc, "'");
+  .|\n     STRING_GROW;
+  <<EOF>>   unexpected_end_of_file (token_start, "'");

 }
 
 
 }
 
 


@@ -414,52 +395,42 @@ splice     (\\[ \f\t\v]*\n)*
   \\[0-7]{1,3} {
     unsigned long c = strtoul (yytext + 1, 0, 8);
     if (UCHAR_MAX < c)
   \\[0-7]{1,3} {
     unsigned long c = strtoul (yytext + 1, 0, 8);
     if (UCHAR_MAX < c)

-      {
-       complain_at (*yylloc, _("invalid escape sequence: %s"),
-                    quote (yytext));
-       YY_STEP;
-      }
+      complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));

     else
     else

-      obstack_1grow (&string_obstack, c);
+      obstack_1grow (&obstack_for_string, c);

   }
 
   }
 

-  \\x[0-9a-fA-F]+ {
+  \\x[0-9abcdefABCDEF]+ {

     unsigned long c;
     unsigned long c;

-    errno = 0;
+    set_errno (0);

     c = strtoul (yytext + 2, 0, 16);
     c = strtoul (yytext + 2, 0, 16);

-    if (UCHAR_MAX < c || errno)
-      {
-       complain_at (*yylloc, _("invalid escape sequence: %s"),
-                    quote (yytext));
-       YY_STEP;
-      }
+    if (UCHAR_MAX < c || get_errno ())
+      complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));

     else
     else

-      obstack_1grow (&string_obstack, c);
+      obstack_1grow (&obstack_for_string, c);

   }
 
   }
 

-  \\a  obstack_1grow (&string_obstack, '\a');
-  \\b  obstack_1grow (&string_obstack, '\b');
-  \\f  obstack_1grow (&string_obstack, '\f');
-  \\n  obstack_1grow (&string_obstack, '\n');
-  \\r  obstack_1grow (&string_obstack, '\r');
-  \\t  obstack_1grow (&string_obstack, '\t');
-  \\v  obstack_1grow (&string_obstack, '\v');
-  \\[\"\'?\\]  obstack_1grow (&string_obstack, yytext[1]);
-  \\(u|U[0-9a-fA-F]{4})[0-9a-fA-F]{4} {
+  \\a  obstack_1grow (&obstack_for_string, '\a');
+  \\b  obstack_1grow (&obstack_for_string, '\b');
+  \\f  obstack_1grow (&obstack_for_string, '\f');
+  \\n  obstack_1grow (&obstack_for_string, '\n');
+  \\r  obstack_1grow (&obstack_for_string, '\r');
+  \\t  obstack_1grow (&obstack_for_string, '\t');
+  \\v  obstack_1grow (&obstack_for_string, '\v');
+
+  /* \\[\"\'?\\] would be shorter, but it confuses xgettext.  */
+  \\("\""|"'"|"?"|"\\")  obstack_1grow (&obstack_for_string, yytext[1]);
+
+  \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} {

     int c = convert_ucn_to_byte (yytext);
     if (c < 0)
     int c = convert_ucn_to_byte (yytext);
     if (c < 0)

-      {
-       complain_at (*yylloc, _("invalid escape sequence: %s"),
-                    quote (yytext));
-       YY_STEP;
-      }
+      complain_at (*loc, _("invalid escape sequence: %s"), quote (yytext));

     else
     else

-      obstack_1grow (&string_obstack, c);
+      obstack_1grow (&obstack_for_string, c);

   }
   \\(.|\n)     {
   }
   \\(.|\n)     {

-    complain_at (*yylloc, _("unrecognized escape sequence: %s"),
-                quote (yytext));
-    YY_OBS_GROW;
+    complain_at (*loc, _("unrecognized escape sequence: %s"), quote (yytext));
+    STRING_GROW;

   }
 }
 
   }
 }
 


@@ -471,9 +442,9 @@ splice       (\\[ \f\t\v]*\n)*
 
 <SC_CHARACTER>
 {
 
 <SC_CHARACTER>
 {

-  "'"              YY_OBS_GROW; BEGIN c_context;
-  \\{splice}[\'\\]  YY_OBS_GROW;
-  <<EOF>>          unexpected_end_of_file (yylloc, "'");
+  "'"                  STRING_GROW; BEGIN context_state;
+  \\{splice}[^$@\[\]]  STRING_GROW;
+  <<EOF>>              unexpected_end_of_file (token_start, "'");

 }
 
 
 }
 
 


@@ -484,9 +455,9 @@ splice       (\\[ \f\t\v]*\n)*
 
 <SC_STRING>
 {
 
 <SC_STRING>
 {

-  "\""             YY_OBS_GROW; BEGIN c_context;
-  \\{splice}[\"\\]  YY_OBS_GROW;
-  <<EOF>>          unexpected_end_of_file (yylloc, "\"");
+  "\""                 STRING_GROW; BEGIN context_state;
+  \\{splice}[^$@\[\]]  STRING_GROW;
+  <<EOF>>              unexpected_end_of_file (token_start, "\"");

 }
 
 
 }
 
 


@@ -496,10 +467,29 @@ splice     (\\[ \f\t\v]*\n)*
 
 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
 {
 
 <SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
 {

-  "'"            YY_OBS_GROW; c_context = YY_START; BEGIN SC_CHARACTER;
-  "\""           YY_OBS_GROW; c_context = YY_START; BEGIN SC_STRING;
-  "/"{splice}"*"  YY_OBS_GROW; c_context = YY_START; BEGIN SC_COMMENT;
-  "/"{splice}"/"  YY_OBS_GROW; c_context = YY_START; BEGIN SC_LINE_COMMENT;
+  "'" {
+    STRING_GROW;
+    context_state = YY_START;
+    token_start = loc->start;
+    BEGIN SC_CHARACTER;
+  }
+  "\"" {
+    STRING_GROW;
+    context_state = YY_START;
+    token_start = loc->start;
+    BEGIN SC_STRING;
+  }
+  "/"{splice}"*" {
+    STRING_GROW;
+    context_state = YY_START;
+    token_start = loc->start;
+    BEGIN SC_COMMENT;
+  }
+  "/"{splice}"/" {
+    STRING_GROW;
+    context_state = YY_START;
+    BEGIN SC_LINE_COMMENT;
+  }

 }
 
 
 }
 
 


@@ -510,15 +500,16 @@ splice     (\\[ \f\t\v]*\n)*
 
 <SC_BRACED_CODE>
 {
 
 <SC_BRACED_CODE>
 {

-  "{"|"<"{splice}"%"  YY_OBS_GROW; braces_level++;
-  "%"{splice}">"      YY_OBS_GROW; braces_level--;
+  "{"|"<"{splice}"%"  STRING_GROW; braces_level++;
+  "%"{splice}">"      STRING_GROW; braces_level--;

   "}" {
   "}" {

-    YY_OBS_GROW;
+    STRING_GROW;

     braces_level--;
     if (braces_level < 0)
       {
     braces_level--;
     if (braces_level < 0)
       {

-       YY_OBS_FINISH;
-       yylval->string = last_string;
+       STRING_FINISH;
+       loc->start = code_start;
+       val->chars = last_string;

        rule_length++;
        BEGIN INITIAL;
        return BRACED_CODE;
        rule_length++;
        BEGIN INITIAL;
        return BRACED_CODE;


@@ -527,14 +518,14 @@ splice     (\\[ \f\t\v]*\n)*
 
   /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly
      (as `<' `<%').  */
 
   /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly
      (as `<' `<%').  */

-  "<"{splice}"<"  YY_OBS_GROW;
+  "<"{splice}"<"  STRING_GROW;

 
   "$"("<"{tag}">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,
 
   "$"("<"{tag}">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,

-                                                  yytext, *yylloc); }
+                                                  yytext, *loc); }

   "@"(-?[0-9]+|"$")               { handle_at (current_braced_code,
   "@"(-?[0-9]+|"$")               { handle_at (current_braced_code,

-                                              yytext, *yylloc); }
+                                              yytext, *loc); }

 
 

-  <<EOF>>  unexpected_end_of_file (yylloc, "}");
+  <<EOF>>  unexpected_end_of_file (code_start, "}");

 }
 
 
 }
 
 


@@ -545,13 +536,14 @@ splice     (\\[ \f\t\v]*\n)*
 <SC_PROLOGUE>
 {
   "%}" {
 <SC_PROLOGUE>
 {
   "%}" {

-    YY_OBS_FINISH;
-    yylval->string = last_string;
+    STRING_FINISH;
+    loc->start = code_start;
+    val->chars = last_string;

     BEGIN INITIAL;
     return PROLOGUE;
   }
 
     BEGIN INITIAL;
     return PROLOGUE;
   }
 

-  <<EOF>>  unexpected_end_of_file (yylloc, "%}");
+  <<EOF>>  unexpected_end_of_file (code_start, "%}");

 }
 
 
 }
 
 


@@ -563,8 +555,9 @@ splice       (\\[ \f\t\v]*\n)*
 <SC_EPILOGUE>
 {
   <<EOF>> {
 <SC_EPILOGUE>
 {
   <<EOF>> {

-    YY_OBS_FINISH;
-    yylval->string = last_string;
+    STRING_FINISH;
+    loc->start = code_start;
+    val->chars = last_string;

     BEGIN INITIAL;
     return EPILOGUE;
   }
     BEGIN INITIAL;
     return EPILOGUE;
   }


@@ -578,26 +571,110 @@ splice    (\\[ \f\t\v]*\n)*
 
 <SC_COMMENT,SC_LINE_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
 {
 
 <SC_COMMENT,SC_LINE_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
 {

-  \$   obstack_sgrow (&string_obstack, "$][");
-  \@   obstack_sgrow (&string_obstack, "@@");
-  \[   obstack_sgrow (&string_obstack, "@{");
-  \]   obstack_sgrow (&string_obstack, "@}");
-  .|\n  YY_OBS_GROW;
+  \$   obstack_sgrow (&obstack_for_string, "$][");
+  \@   obstack_sgrow (&obstack_for_string, "@@");
+  \[   obstack_sgrow (&obstack_for_string, "@{");
+  \]   obstack_sgrow (&obstack_for_string, "@}");
+  .|\n  STRING_GROW;

 }
 
 
 %%
 
 }
 
 
 %%
 

+/* Set *LOC and adjust scanner cursor to account for token TOKEN of
+   size SIZE.  */
+
+static void
+adjust_location (location *loc, char const *token, size_t size)
+{
+  int line = scanner_cursor.line;
+  int column = scanner_cursor.column;
+  char const *p0 = token;
+  char const *p = token;
+  char const *lim = token + size;
+
+  loc->start = scanner_cursor;
+
+  for (p = token; p < lim; p++)
+    switch (*p)
+      {
+      case '\n':
+       line++;
+       column = 1;
+       p0 = p + 1;
+       break;
+
+      case '\t':
+       column += mbsnwidth (p0, p - p0, 0);
+       column += 8 - ((column - 1) & 7);
+       p0 = p + 1;
+       break;
+      }
+
+  scanner_cursor.line = line;
+  scanner_cursor.column = column + mbsnwidth (p0, p - p0, 0);
+
+  loc->end = scanner_cursor;
+}
+
+
+/* Read bytes from FP into buffer BUF of size SIZE.  Return the
+   number of bytes read.  Remove '\r' from input, treating \r\n
+   and isolated \r as \n.  */
+
+static size_t
+no_cr_read (FILE *fp, char *buf, size_t size)
+{
+  size_t s = fread (buf, 1, size, fp);
+  if (s)
+    {
+      char *w = memchr (buf, '\r', s);
+      if (w)
+       {
+         char const *r = ++w;
+         char const *lim = buf + s;
+
+         for (;;)
+           {
+             /* Found an '\r'.  Treat it like '\n', but ignore any
+                '\n' that immediately follows.  */
+             w[-1] = '\n';
+             if (r == lim)
+               {
+                 int ch = getc (fp);
+                 if (ch != '\n' && ungetc (ch, fp) != ch)
+                   break;
+               }
+             else if (*r == '\n')
+               r++;
+
+             /* Copy until the next '\r'.  */
+             do
+               {
+                 if (r == lim)
+                   return w - buf;
+               }
+             while ((*w++ = *r++) != '\r');
+           }
+
+         return w - buf;
+       }
+    }
+
+  return s;
+}
+
+

 /*------------------------------------------------------------------.
 | TEXT is pointing to a wannabee semantic value (i.e., a `$').      |
 |                                                                   |
 | Possible inputs: $[<TYPENAME>]($|integer)                         |
 |                                                                   |
 /*------------------------------------------------------------------.
 | TEXT is pointing to a wannabee semantic value (i.e., a `$').      |
 |                                                                   |
 | Possible inputs: $[<TYPENAME>]($|integer)                         |
 |                                                                   |

-| Output to the STRING_OBSTACK a reference to this semantic value.  |
+| Output to OBSTACK_FOR_STRING a reference to this semantic value.  |

 `------------------------------------------------------------------*/
 
 static inline void
 `------------------------------------------------------------------*/
 
 static inline void

-handle_action_dollar (char *text, location_t location)
+handle_action_dollar (char *text, location loc)

 {
   const char *type_name = NULL;
   char *cp = text + 1;
 {
   const char *type_name = NULL;
   char *cp = text + 1;


@@ -615,38 +692,37 @@ handle_action_dollar (char *text, location_t location)
   if (*cp == '$')
     {
       if (!type_name)
   if (*cp == '$')
     {
       if (!type_name)

-       type_name = symbol_list_n_type_name_get (current_rule, location, 0);
+       type_name = symbol_list_n_type_name_get (current_rule, loc, 0);

       if (!type_name && typed)
       if (!type_name && typed)

-       complain_at (location, _("$$ of `%s' has no declared type"),
+       complain_at (loc, _("$$ of `%s' has no declared type"),

                     current_rule->sym->tag);
       if (!type_name)
        type_name = "";
                     current_rule->sym->tag);
       if (!type_name)
        type_name = "";

-      obstack_fgrow1 (&string_obstack,
+      obstack_fgrow1 (&obstack_for_string,

                      "]b4_lhs_value([%s])[", type_name);
     }
   else
     {
       long num;
                      "]b4_lhs_value([%s])[", type_name);
     }
   else
     {
       long num;

-      errno = 0;
+      set_errno (0);

       num = strtol (cp, 0, 10);
 
       num = strtol (cp, 0, 10);
 

-      if (INT_MIN <= num && num <= rule_length && ! errno)
+      if (INT_MIN <= num && num <= rule_length && ! get_errno ())

        {
          int n = num;
          if (!type_name && n > 0)
        {
          int n = num;
          if (!type_name && n > 0)

-           type_name = symbol_list_n_type_name_get (current_rule, location,
-                                                    n);
+           type_name = symbol_list_n_type_name_get (current_rule, loc, n);

          if (!type_name && typed)
          if (!type_name && typed)

-           complain_at (location, _("$%d of `%s' has no declared type"),
-                     n, current_rule->sym->tag);
+           complain_at (loc, _("$%d of `%s' has no declared type"),
+                        n, current_rule->sym->tag);

          if (!type_name)
            type_name = "";
          if (!type_name)
            type_name = "";

-         obstack_fgrow3 (&string_obstack,
+         obstack_fgrow3 (&obstack_for_string,

                          "]b4_rhs_value([%d], [%d], [%s])[",
                          rule_length, n, type_name);
        }
       else
                          "]b4_rhs_value([%d], [%d], [%s])[",
                          rule_length, n, type_name);
        }
       else

-       complain_at (location, _("integer out of range: %s"), quote (text));
+       complain_at (loc, _("integer out of range: %s"), quote (text));

     }
 }
 
     }
 }
 


@@ -657,13 +733,13 @@ handle_action_dollar (char *text, location_t location)
 `---------------------------------------------------------------*/
 
 static inline void
 `---------------------------------------------------------------*/
 
 static inline void

-handle_symbol_code_dollar (char *text, location_t location)
+handle_symbol_code_dollar (char *text, location loc)

 {
   char *cp = text + 1;
   if (*cp == '$')
 {
   char *cp = text + 1;
   if (*cp == '$')

-    obstack_sgrow (&string_obstack, "]b4_dollar_dollar[");
+    obstack_sgrow (&obstack_for_string, "]b4_dollar_dollar[");

   else
   else

-    complain_at (location, _("invalid value: %s"), quote (text));
+    complain_at (loc, _("invalid value: %s"), quote (text));

 }
 
 
 }
 
 


@@ -673,18 +749,17 @@ handle_symbol_code_dollar (char *text, location_t location)
 `-----------------------------------------------------------------*/
 
 static void
 `-----------------------------------------------------------------*/
 
 static void

-handle_dollar (braced_code_t braced_code_kind,
-              char *text, location_t location)
+handle_dollar (braced_code braced_code_kind, char *text, location loc)

 {
   switch (braced_code_kind)
     {
     case action_braced_code:
 {
   switch (braced_code_kind)
     {
     case action_braced_code:

-      handle_action_dollar (text, location);
+      handle_action_dollar (text, loc);

       break;
 
     case destructor_braced_code:
     case printer_braced_code:
       break;
 
     case destructor_braced_code:
     case printer_braced_code:

-      handle_symbol_code_dollar (text, location);
+      handle_symbol_code_dollar (text, loc);

       break;
     }
 }
       break;
     }
 }


@@ -692,33 +767,33 @@ handle_dollar (braced_code_t braced_code_kind,
 
 /*------------------------------------------------------.
 | TEXT is a location token (i.e., a `@...').  Output to |
 
 /*------------------------------------------------------.
 | TEXT is a location token (i.e., a `@...').  Output to |

-| STRING_OBSTACK a reference to this location.          |
+| OBSTACK_FOR_STRING a reference to this location.      |

 `------------------------------------------------------*/
 
 static inline void
 `------------------------------------------------------*/
 
 static inline void

-handle_action_at (char *text, location_t location)
+handle_action_at (char *text, location loc)

 {
   char *cp = text + 1;
   locations_flag = 1;
 
   if (*cp == '$')
     {
 {
   char *cp = text + 1;
   locations_flag = 1;
 
   if (*cp == '$')
     {

-      obstack_sgrow (&string_obstack, "]b4_lhs_location[");
+      obstack_sgrow (&obstack_for_string, "]b4_lhs_location[");

     }
   else
     {
       long num;
     }
   else
     {
       long num;

-      errno = 0;
+      set_errno (0);

       num = strtol (cp, 0, 10);
 
       num = strtol (cp, 0, 10);
 

-      if (INT_MIN <= num && num <= rule_length && ! errno)
+      if (INT_MIN <= num && num <= rule_length && ! get_errno ())

        {
          int n = num;
        {
          int n = num;

-         obstack_fgrow2 (&string_obstack, "]b4_rhs_location([%d], [%d])[",
+         obstack_fgrow2 (&obstack_for_string, "]b4_rhs_location([%d], [%d])[",

                          rule_length, n);
        }
       else
                          rule_length, n);
        }
       else

-       complain_at (location, _("integer out of range: %s"), quote (text));
+       complain_at (loc, _("integer out of range: %s"), quote (text));

     }
 }
 
     }
 }
 


@@ -729,13 +804,13 @@ handle_action_at (char *text, location_t location)
 `---------------------------------------------------------------*/
 
 static inline void
 `---------------------------------------------------------------*/
 
 static inline void

-handle_symbol_code_at (char *text, location_t location)
+handle_symbol_code_at (char *text, location loc)

 {
   char *cp = text + 1;
   if (*cp == '$')
 {
   char *cp = text + 1;
   if (*cp == '$')

-    obstack_sgrow (&string_obstack, "]b4_at_dollar[");
+    obstack_sgrow (&obstack_for_string, "]b4_at_dollar[");

   else
   else

-    complain_at (location, _("invalid value: %s"), quote (text));
+    complain_at (loc, _("invalid value: %s"), quote (text));

 }
 
 
 }
 
 


@@ -745,18 +820,17 @@ handle_symbol_code_at (char *text, location_t location)
 `-------------------------------------------------------------------*/
 
 static void
 `-------------------------------------------------------------------*/
 
 static void

-handle_at (braced_code_t braced_code_kind,
-          char *text, location_t location)
+handle_at (braced_code braced_code_kind, char *text, location loc)

 {
   switch (braced_code_kind)
     {
     case action_braced_code:
 {
   switch (braced_code_kind)
     {
     case action_braced_code:

-      handle_action_at (text, location);
+      handle_action_at (text, loc);

       break;
 
     case destructor_braced_code:
     case printer_braced_code:
       break;
 
     case destructor_braced_code:
     case printer_braced_code:

-      handle_symbol_code_at (text, location);
+      handle_symbol_code_at (text, loc);

       break;
     }
 }
       break;
     }
 }


@@ -823,36 +897,37 @@ convert_ucn_to_byte (char const *ucn)
 `----------------------------------------------------------------*/
 
 static void
 `----------------------------------------------------------------*/
 
 static void

-handle_syncline (char *args, location_t *location)
+handle_syncline (char *args)

 {
   int lineno = strtol (args, &args, 10);
   const char *file = NULL;
   file = strchr (args, '"') + 1;
   *strchr (file, '"') = 0;
 {
   int lineno = strtol (args, &args, 10);
   const char *file = NULL;
   file = strchr (args, '"') + 1;
   *strchr (file, '"') = 0;

-  /* FIXME: Leaking...  Can't free, as some locations are still
-     pointing to the old file name.  */
-  infile = xstrdup (file);
-  location->file = infile;
-  location->last_line = lineno;
+  scanner_cursor.file = current_file = xstrdup (file);
+  scanner_cursor.line = lineno;
+  scanner_cursor.column = 1;

 }
 
 
 }
 
 

-/*-------------------------------------------------------------.
-| Report an unexpected end of file at LOC.  An end of file was |
-| encountered and the expected TOKEN_END was missing.  After   |
-| reporting the problem, pretend that TOKEN_END was found.     |
-`-------------------------------------------------------------*/
+/*------------------------------------------------------------------------.
+| Report an unexpected EOF in a token or comment starting at START.       |
+| An end of file was encountered and the expected TOKEN_END was missing.  |
+| After reporting the problem, pretend that TOKEN_END was found.          |
+`------------------------------------------------------------------------*/

 
 static void
 
 static void

-unexpected_end_of_file (location_t *loc, char const *token_end)
+unexpected_end_of_file (boundary start, char const *token_end)

 {
   size_t i = strlen (token_end);
 
 {
   size_t i = strlen (token_end);
 

-  complain_at (*loc, _("missing `%s' at end of file"), token_end);
+  location loc;
+  loc.start = start;
+  loc.end = scanner_cursor;
+  complain_at (loc, _("missing `%s' at end of file"), token_end);

 
 

-  /* Adjust location's last column so that any later message does not
-     mention the characters just inserted.  */
-  loc->last_column -= i;
+  /* Adjust scanner cursor so that any later message does not count
+     the characters about to be inserted.  */
+  scanner_cursor.column -= i;

 
   while (i != 0)
     unput (token_end[--i]);
 
   while (i != 0)
     unput (token_end[--i]);


@@ -866,7 +941,7 @@ unexpected_end_of_file (location_t *loc, char const *token_end)
 void
 scanner_initialize (void)
 {
 void
 scanner_initialize (void)
 {

-  obstack_init (&string_obstack);
+  obstack_init (&obstack_for_string);

 }
 
 
 }
 
 


@@ -877,7 +952,7 @@ scanner_initialize (void)
 void
 scanner_free (void)
 {
 void
 scanner_free (void)
 {

-  obstack_free (&string_obstack, 0);
+  obstack_free (&obstack_for_string, 0);

   /* Reclaim Flex's buffers.  */
   yy_delete_buffer (YY_CURRENT_BUFFER);
 }
   /* Reclaim Flex's buffers.  */
   yy_delete_buffer (YY_CURRENT_BUFFER);
 }