* src/lex.h: Prototype `lex.c' exported functions.

* src/reader.c: Adjust.
* src/lex.c: Formatting changes.
(safegetc): Rename as...
(xgetc): this.
This commit is contained in:
Akim Demaille
2000-10-02 07:59:08 +00:00
parent 720d742f8e
commit abadc11757
4 changed files with 288 additions and 239 deletions

View File

@@ -1,3 +1,12 @@
2000-10-02 Akim Demaille <akim@epita.fr>
* src/lex.h: Prototype `lex.c' exported functions.
* src/reader.c: Adjust.
* src/lex.c: Formatting changes.
(safegetc): Rename as...
(xgetc): this.
2000-10-02 Akim Demaille <akim@epita.fr>
* src/lalr.h: New file.

177
src/lex.c
View File

@@ -18,14 +18,6 @@
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
/*
lex is the entry point. It is called from reader.c.
It returns one of the token-type codes defined in lex.h.
When an identifier is seen, the code IDENTIFIER is returned
and the name is looked up in the symbol table using symtab.c;
symval is set to a pointer to the entry found. */
#include "system.h"
#include "getargs.h"
#include "files.h"
@@ -39,16 +31,6 @@
extern int translations;
extern void init_lex PARAMS((void));
extern char *grow_token_buffer PARAMS((char *));
extern int skip_white_space PARAMS((void));
extern void unlex PARAMS((int));
extern int lex PARAMS((void));
extern int parse_percent_token PARAMS((void));
static int safegetc PARAMS((FILE *));
static int literalchar PARAMS((char **, int *, char));
/* functions from main.c */
extern char *printable_version PARAMS ((int));
@@ -87,8 +69,8 @@ grow_token_buffer (char *p)
int
skip_white_space (void)
{
register int c;
register int inside;
int c;
int inside;
c = getc (finput);
@@ -155,29 +137,31 @@ skip_white_space (void)
/* do a getc, but give error message if EOF encountered */
static int
safegetc (FILE *f)
xgetc (FILE *f)
{
register int c = getc(f);
int c = getc (f);
if (c == EOF)
fatal (_("unexpected end of file"));
return c;
}
/* read one literal character from finput. process \ escapes.
append the normalized string version of the char to *pp.
assign the character code to *pcode
return 1 unless the character is an unescaped `term' or \n
report error for \n
*/
/*------------------------------------------------------------------.
| Read one literal character from finput. Process \ escapes. |
| Append the normalized string version of the char to *PP. Assign |
| the character code to *PCODE. Return 1 unless the character is an |
| unescaped `term' or \n report error for \n |
`------------------------------------------------------------------*/
static int
literalchar (char **pp, int *pcode, char term)
{
register int c;
register char *p;
register int code;
int c;
char *p;
int code;
int wasquote = 0;
c = safegetc(finput);
c = xgetc (finput);
if (c == '\n')
{
complain (_("unescaped newline in constant"));
@@ -193,17 +177,27 @@ literalchar (char **pp, int *pcode, char term)
}
else
{
c = safegetc(finput);
if (c == 't') code = '\t';
else if (c == 'n') code = '\n';
else if (c == 'a') code = '\007';
else if (c == 'r') code = '\r';
else if (c == 'f') code = '\f';
else if (c == 'b') code = '\b';
else if (c == 'v') code = '\013';
else if (c == '\\') code = '\\';
else if (c == '\'') code = '\'';
else if (c == '\"') code = '\"';
c = xgetc (finput);
if (c == 't')
code = '\t';
else if (c == 'n')
code = '\n';
else if (c == 'a')
code = '\007';
else if (c == 'r')
code = '\r';
else if (c == 'f')
code = '\f';
else if (c == 'b')
code = '\b';
else if (c == 'v')
code = '\013';
else if (c == '\\')
code = '\\';
else if (c == '\'')
code = '\'';
else if (c == '\"')
code = '\"';
else if (c <= '7' && c >= '0')
{
code = 0;
@@ -217,13 +211,13 @@ literalchar (char **pp, int *pcode, char term)
code &= 0xFF;
break;
}
c = safegetc(finput);
c = xgetc (finput);
}
ungetc (c, finput);
}
else if (c == 'x')
{
c = safegetc(finput);
c = xgetc (finput);
code = 0;
while (1)
{
@@ -237,12 +231,11 @@ literalchar (char **pp, int *pcode, char term)
break;
if (code >= 256 || code < 0)
{
complain (_("hexadecimal value above 255: `\\x%x'"),
code);
complain (_("hexadecimal value above 255: `\\x%x'"), code);
code &= 0xFF;
break;
}
c = safegetc(finput);
c = xgetc (finput);
}
ungetc (c, finput);
}
@@ -261,17 +254,53 @@ literalchar (char **pp, int *pcode, char term)
p = *pp;
if (code == term && wasquote)
*p++ = code;
else if (code == '\\') {*p++ = '\\'; *p++ = '\\';}
else if (code == '\'') {*p++ = '\\'; *p++ = '\'';}
else if (code == '\"') {*p++ = '\\'; *p++ = '\"';}
else if (code == '\\')
{
*p++ = '\\';
*p++ = '\\';
}
else if (code == '\'')
{
*p++ = '\\';
*p++ = '\'';
}
else if (code == '\"')
{
*p++ = '\\';
*p++ = '\"';
}
else if (code >= 040 && code < 0177)
*p++ = code;
else if (code == '\t') {*p++ = '\\'; *p++ = 't';}
else if (code == '\n') {*p++ = '\\'; *p++ = 'n';}
else if (code == '\r') {*p++ = '\\'; *p++ = 'r';}
else if (code == '\v') {*p++ = '\\'; *p++ = 'v';}
else if (code == '\b') {*p++ = '\\'; *p++ = 'b';}
else if (code == '\f') {*p++ = '\\'; *p++ = 'f';}
else if (code == '\t')
{
*p++ = '\\';
*p++ = 't';
}
else if (code == '\n')
{
*p++ = '\\';
*p++ = 'n';
}
else if (code == '\r')
{
*p++ = '\\';
*p++ = 'r';
}
else if (code == '\v')
{
*p++ = '\\';
*p++ = 'v';
}
else if (code == '\b')
{
*p++ = '\\';
*p++ = 'b';
}
else if (code == '\f')
{
*p++ = '\\';
*p++ = 'f';
}
else
{
*p++ = '\\';
@@ -296,7 +325,7 @@ unlex (int token)
int
lex (void)
{
register int c;
int c;
char *p;
if (unlexed >= 0)
@@ -330,6 +359,7 @@ lex (void)
case 'u': case 'v': case 'w': case 'x': case 'y':
case 'z':
case '.': case '_':
p = token_buffer;
while (isalnum (c) || c == '_' || c == '.')
{
@@ -366,7 +396,6 @@ lex (void)
}
case '\'':
/* parse the literal token and compute character code in code */
translations = -1;
@@ -399,7 +428,6 @@ lex (void)
}
case '\"':
/* parse the literal string token and treat as an identifier */
translations = -1;
@@ -439,7 +467,8 @@ lex (void)
do
{
c = getc (finput);
if (c == '\n') lineno++;
if (c == '\n')
lineno++;
}
while (c == ' ' || c == '\n' || c == '\t');
@@ -486,15 +515,16 @@ lex (void)
}
}
/* the following table dictates the action taken for the various
% directives. A setflag value causes the named flag to be
set. A retval action returns the code.
*/
struct percent_table_struct {
/* the following table dictates the action taken for the various %
directives. A setflag value causes the named flag to be set. A
retval action returns the code. */
struct percent_table_struct
{
const char *name;
void *setflag;
int retval;
} percent_table[] =
}
percent_table[] =
{
{ "token", NULL, TOKEN },
{ "term", NULL, TOKEN },
@@ -512,11 +542,9 @@ struct percent_table_struct {
{ "semantic_parser", NULL, SEMANTIC_PARSER },
{ "pure_parser", NULL, PURE_PARSER },
{ "prec", NULL, PREC },
{ "no_lines", &nolinesflag, NOOP}, /* -l */
{ "raw", &rawtoknumflag, NOOP }, /* -r */
{ "token_table", &toknumflag, NOOP}, /* -k */
#if 0
/* These can be utilized after main is reoganized so
open_files() is deferred 'til after read_declarations().
@@ -530,14 +558,12 @@ struct percent_table_struct {
{ "output_file", &spec_outfile, SETOPT}, /* -o */
{ "file_prefix", &spec_file_prefix, SETOPT}, /* -b */
{ "name_prefix", &spec_name_prefix, SETOPT}, /* -p */
/* These would be acceptable, but they do not affect processing */
{ "verbose", &verboseflag, NOOP}, /* -v */
{ "debug", &debugflag, NOOP}, /* -t */
/* {"help", <print usage stmt>, NOOP}, *//* -h */
/* {"version", <print version number> , NOOP}, *//* -V */
#endif
{ NULL, NULL, ILLEGAL}
};
@@ -547,9 +573,9 @@ struct percent_table_struct {
int
parse_percent_token (void)
{
register int c;
register char *p;
register struct percent_table_struct *tx;
int c;
char *p;
struct percent_table_struct *tx;
p = token_buffer;
c = getc (finput);
@@ -590,7 +616,8 @@ parse_percent_token (void)
if (p == token_buffer + maxtoken)
p = grow_token_buffer (p);
if (c == '-') c = '_';
if (c == '-')
c = '_';
*p++ = c;
c = getc (finput);
}

View File

@@ -1,5 +1,5 @@
/* Token type definitions for bison's input reader,
Copyright (C) 1984, 1989, 1992 Free Software Foundation, Inc.
Copyright (C) 1984, 1989, 1992, 2000 Free Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
@@ -18,7 +18,10 @@ along with Bison; see the file COPYING. If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
#ifndef LEX_H_
# define LEX_H_
/* Token-type codes. */
# define ENDFILE 0
# define IDENTIFIER 1
# define COMMA 2
@@ -49,3 +52,20 @@ Boston, MA 02111-1307, USA. */
# define ILLEGAL 27
# define MAXTOKEN 1024
void init_lex PARAMS ((void));
char *grow_token_buffer PARAMS ((char *));
int skip_white_space PARAMS ((void));
void unlex PARAMS ((int));
/* Return one of the token-type codes. When an identifier is seen,
the code IDENTIFIER is returned and the name is looked up in the
symbol table using symtab.c; symval is set to a pointer to the
entry found. */
int lex PARAMS ((void));
int parse_percent_token PARAMS ((void));
#endif /* !LEX_H_ */

View File

@@ -68,17 +68,10 @@ extern int expected_conflicts;
extern char *token_buffer;
extern int maxtoken;
extern void init_lex PARAMS((void));
extern char *grow_token_buffer PARAMS((char *));
extern void tabinit PARAMS((void));
extern void free_symtab PARAMS((void));
extern void open_extra_files PARAMS((void));
extern char *printable_version PARAMS((int));
extern void unlex PARAMS((int));
extern int skip_white_space PARAMS((void));
extern int parse_percent_token PARAMS((void));
extern int lex PARAMS((void));
typedef
struct symbol_list