02111-1307 USA
*/
-%option debug nodefault noyywrap nounput never-interactive stack
+%option debug nodefault noyywrap never-interactive
%option prefix="gram_" outfile="lex.yy.c"
%{
#include "mbswidth.h"
#include "complain.h"
#include "quote.h"
+#include "struniq.h"
#include "getargs.h"
#include "gram.h"
#include "reader.h"
#define YY_USER_INIT \
do { \
LOCATION_RESET (*yylloc); \
- yylloc->file = infile; \
+ yylloc->file = current_file; \
/* This is only to avoid GCC warnings. */ \
if (yycontrol) {;}; \
} while (0)
}
-/* Report an unexpected end of file at LOC. A token or comment began
- with TOKEN_START, but an end of file was encountered and the
- expected TOKEN_END was missing. */
-
-static void
-unexpected_end_of_file (location_t loc,
- char const *token_start, char const *token_end)
-{
- complain_at (loc, _("unexpected end of file in `%s ... %s'"),
- token_start, token_end);
-}
-
-
/* STRING_OBSTACK -- Used to store all the characters that we need to
keep (to construct ID, STRINGS etc.). Use the following macros to
used, and which is used by YY_OBS_FREE to free the last string. */
static struct obstack string_obstack;
-char *last_string;
+
+/* A string representing the most recently saved token. */
+static char *last_string;
+
#define YY_OBS_GROW \
obstack_grow (&string_obstack, yytext, yyleng)
last_string = obstack_finish (&string_obstack); \
} while (0)
-#define YY_OBS_FREE \
- do { \
- obstack_free (&string_obstack, last_string); \
- } while (0)
+#define YY_OBS_FREE \
+ obstack_free (&string_obstack, last_string)
void
scanner_last_string_free (void)
YY_OBS_FREE;
}
-
-static int percent_percent_count = 0;
-
/* 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
to the top of the stack. It is not the same as the rule->length in
char *cp, location_t location);
static void handle_syncline (char *args, location_t *location);
static int convert_ucn_to_byte (char const *hex_text);
+static void unexpected_end_of_file (location_t *, char const *);
%}
%x SC_COMMENT SC_LINE_COMMENT SC_YACC_COMMENT
%%
%{
+ /* 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);
+
/* At each yylex invocation, mark the current position as the
start of the next token. */
YY_STEP;
"," return COMMA;
";" return SEMICOLON;
- [ \f\n\t\v]+ YY_STEP;
+ [ \f\n\t\v] YY_STEP;
{id} {
yylval->symbol = symbol_get (yytext, *yylloc);
}
/* Characters. We don't check there is only one. */
- "'" YY_OBS_GROW; yy_push_state (SC_ESCAPED_CHARACTER);
+ "'" YY_OBS_GROW; BEGIN SC_ESCAPED_CHARACTER;
/* Strings. */
- "\"" YY_OBS_GROW; yy_push_state (SC_ESCAPED_STRING);
+ "\"" YY_OBS_GROW; BEGIN SC_ESCAPED_STRING;
/* Comments. */
"/*" BEGIN SC_YACC_COMMENT;
"//".* YY_STEP;
/* Prologue. */
- "%{" yy_push_state (SC_PROLOGUE);
+ "%{" BEGIN SC_PROLOGUE;
/* Code in between braces. */
- "{" YY_OBS_GROW; braces_level = 0; yy_push_state (SC_BRACED_CODE);
+ "{" YY_OBS_GROW; braces_level = 0; BEGIN SC_BRACED_CODE;
/* A type. */
"<"{tag}">" {
obstack_grow (&string_obstack, yytext + 1, yyleng - 2);
YY_OBS_FINISH;
- yylval->string = last_string;
+ yylval->struniq = struniq_new (last_string);
+ YY_OBS_FREE;
return TYPE;
}
-
- "%%" {
+ "%%" {
+ static int percent_percent_count;
if (++percent_percent_count == 2)
- yy_push_state (SC_EPILOGUE);
+ BEGIN SC_EPILOGUE;
return PERCENT_PERCENT;
}
- . {
+ . {
complain_at (*yylloc, _("invalid character: %s"), quote (yytext));
YY_STEP;
}
}
- /*-------------------------------------------------------------------.
- | Whatever the start condition (but those which correspond to |
- | entities `swallowed' by Bison: SC_YACC_COMMENT, SC_ESCAPED_STRING, |
- | and SC_ESCAPED_CHARACTER), no M4 character must escape as is. |
- `-------------------------------------------------------------------*/
-
-<SC_COMMENT,SC_LINE_COMMENT,SC_STRING,SC_CHARACTER,SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
-{
- \[ obstack_sgrow (&string_obstack, "@<:@");
- \] obstack_sgrow (&string_obstack, "@:>@");
-}
-
-
/*---------------------------------------------------------------.
| Scanning a Yacc comment. The initial `/ *' is already eaten. |
`---------------------------------------------------------------*/
BEGIN INITIAL;
}
- [^*]+|"*" ;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "/*", "*/");
- BEGIN INITIAL;
- }
+ .|\n ;
+ <<EOF>> unexpected_end_of_file (yylloc, "*/");
}
<SC_COMMENT>
{
- "*"{splice}"/" YY_OBS_GROW; yy_pop_state ();
- [^*\[\]]+|"*" YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "/*", "*/");
- yy_pop_state ();
- }
+ "*"{splice}"/" YY_OBS_GROW; BEGIN c_context;
+ <<EOF>> unexpected_end_of_file (yylloc, "*/");
}
<SC_LINE_COMMENT>
{
- "\n" YY_OBS_GROW; yy_pop_state ();
- ([^\n\[\]]|{splice})+ YY_OBS_GROW;
- <<EOF>> yy_pop_state ();
+ "\n" YY_OBS_GROW; BEGIN c_context;
+ {splice} YY_OBS_GROW;
+ <<EOF>> BEGIN c_context;
}
<SC_ESCAPED_STRING>
{
"\"" {
- assert (yy_top_state () == INITIAL);
YY_OBS_GROW;
YY_OBS_FINISH;
yylval->string = last_string;
- yy_pop_state ();
rule_length++;
+ BEGIN INITIAL;
return STRING;
}
- [^\"\\]+ YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "\"", "\"");
- assert (yy_top_state () == INITIAL);
- YY_OBS_FINISH;
- yylval->string = last_string;
- yy_pop_state ();
- return STRING;
- }
+ .|\n YY_OBS_GROW;
+ <<EOF>> unexpected_end_of_file (yylloc, "\"");
}
/*---------------------------------------------------------------.
{
"'" {
YY_OBS_GROW;
- assert (yy_top_state () == INITIAL);
- {
- 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;
- yy_pop_state ();
- rule_length++;
- return ID;
- }
- }
-
- [^\'\\]+ YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "'", "'");
- assert (yy_top_state () == INITIAL);
YY_OBS_FINISH;
- yylval->string = last_string;
- yy_pop_state ();
- return CHARACTER;
+ 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;
+ rule_length++;
+ BEGIN INITIAL;
+ return ID;
}
+
+ .|\n YY_OBS_GROW;
+ <<EOF>> unexpected_end_of_file (yylloc, "'");
}
quote (yytext));
YY_OBS_GROW;
}
- /* FLex wants this rule, in case of a `\<<EOF>>'. */
- \\ YY_OBS_GROW;
}
<SC_CHARACTER>
{
- "'" {
- YY_OBS_GROW;
- assert (yy_top_state () != INITIAL);
- yy_pop_state ();
- }
-
- [^'\[\]\\]+ YY_OBS_GROW;
- \\{splice}[^\[\]] YY_OBS_GROW;
- {splice} YY_OBS_GROW;
- /* Needed for `\<<EOF>>', `\\<<newline>>[', and `\\<<newline>>]'. */
- \\ YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "'", "'");
- assert (yy_top_state () != INITIAL);
- yy_pop_state ();
- }
+ "'" YY_OBS_GROW; BEGIN c_context;
+ \\{splice}[\'\\] YY_OBS_GROW;
+ <<EOF>> unexpected_end_of_file (yylloc, "'");
}
<SC_STRING>
{
- "\"" {
- assert (yy_top_state () != INITIAL);
- YY_OBS_GROW;
- yy_pop_state ();
- }
-
- [^\"\[\]\\]+ YY_OBS_GROW;
- \\{splice}[^\[\]] YY_OBS_GROW;
- {splice} YY_OBS_GROW;
- /* Needed for `\<<EOF>>', `\\<<newline>>[', and `\\<<newline>>]'. */
- \\ YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "\"", "\"");
- assert (yy_top_state () != INITIAL);
- yy_pop_state ();
- }
+ "\"" YY_OBS_GROW; BEGIN c_context;
+ \\{splice}[\"\\] YY_OBS_GROW;
+ <<EOF>> unexpected_end_of_file (yylloc, "\"");
}
<SC_BRACED_CODE,SC_PROLOGUE,SC_EPILOGUE>
{
- /* Characters. We don't check there is only one. */
- "'" YY_OBS_GROW; yy_push_state (SC_CHARACTER);
-
- /* Strings. */
- "\"" YY_OBS_GROW; yy_push_state (SC_STRING);
-
- /* Comments. */
- "/"{splice}"*" YY_OBS_GROW; yy_push_state (SC_COMMENT);
- "/"{splice}"/" YY_OBS_GROW; yy_push_state (SC_LINE_COMMENT);
-
- /* Not comments. */
- "/" YY_OBS_GROW;
+ "'" 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;
}
braces_level--;
if (braces_level < 0)
{
- yy_pop_state ();
YY_OBS_FINISH;
yylval->string = last_string;
rule_length++;
+ BEGIN INITIAL;
return BRACED_CODE;
}
}
+ /* Tokenize `<<%' correctly (as `<<' `%') rather than incorrrectly
+ (as `<' `<%'). */
+ "<"{splice}"<" YY_OBS_GROW;
+
"$"("<"{tag}">")?(-?[0-9]+|"$") { handle_dollar (current_braced_code,
yytext, *yylloc); }
"@"(-?[0-9]+|"$") { handle_at (current_braced_code,
yytext, *yylloc); }
- /* `"<"{splice}"<"' tokenizes `<<%' correctly (as `<<' `%') rather
- than incorrrectly (as `<' `<%'). */
- [^\"$%\'/<@\[\]\{\}]+|[$%/<@]|"<"{splice}"<" YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "{", "}");
- yy_pop_state ();
- YY_OBS_FINISH;
- yylval->string = last_string;
- return BRACED_CODE;
- }
-
+ <<EOF>> unexpected_end_of_file (yylloc, "}");
}
<SC_PROLOGUE>
{
"%}" {
- yy_pop_state ();
YY_OBS_FINISH;
yylval->string = last_string;
+ BEGIN INITIAL;
return PROLOGUE;
}
- [^%\[\]/\'\"]+ YY_OBS_GROW;
- "%" YY_OBS_GROW;
-
- <<EOF>> {
- unexpected_end_of_file (*yylloc, "%{", "%}");
- yy_pop_state ();
- YY_OBS_FINISH;
- yylval->string = last_string;
- return PROLOGUE;
- }
+ <<EOF>> unexpected_end_of_file (yylloc, "%}");
}
<SC_EPILOGUE>
{
- [^\[\]]+ YY_OBS_GROW;
-
<<EOF>> {
- yy_pop_state ();
YY_OBS_FINISH;
yylval->string = last_string;
+ BEGIN INITIAL;
return EPILOGUE;
}
}
+ /*----------------------------------------------------------------.
+ | By default, grow the string obstack with the input, escaping M4 |
+ | quoting characters. |
+ `----------------------------------------------------------------*/
+
+<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;
+}
+
+
%%
/*------------------------------------------------------------------.
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;
+ current_file = xstrdup (file);
+ location->file = current_file;
location->last_line = lineno;
}
+
+/*-------------------------------------------------------------.
+| 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. |
+`-------------------------------------------------------------*/
+
+static void
+unexpected_end_of_file (location_t *loc, char const *token_end)
+{
+ size_t i = strlen (token_end);
+
+ 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;
+
+ while (i != 0)
+ unput (token_end[--i]);
+}
+
+
/*-------------------------.
| Initialize the scanner. |
`-------------------------*/