/* PSPP - computes sample statistics.
- Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
- Written by Ben Pfaff <blp@gnu.org>.
+ Copyright (C) 1997-9, 2000, 2006 Free Software Foundation, Inc.
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License as
#include <stdarg.h>
#include <stdlib.h>
#include <libpspp/alloc.h>
+#include <libpspp/assertion.h>
#include <language/command.h>
#include <libpspp/message.h>
-#include <language/line-buffer.h>
#include <libpspp/magic.h>
#include <data/settings.h>
+#include <libpspp/getl.h>
#include <libpspp/str.h>
#include "size_max.h"
#define _(msgid) gettext (msgid)
#define N_(msgid) msgid
-/*
-#define DUMP_TOKENS 1
-*/
-\f
-/* Global variables. */
-
-extern const char *keywords[T_N_KEYWORDS + 1];
+#define DUMP_TOKENS 0
-/* Current token. */
-int token;
+struct lexer
+{
+ struct string line_buffer;
-/* T_POS_NUM, T_NEG_NUM: the token's value. */
-double tokval;
+ struct source_stream *ss;
-/* T_ID: the identifier. */
-char tokid[LONG_NAME_LEN + 1];
+ int token; /* Current token. */
+ double tokval; /* T_POS_NUM, T_NEG_NUM: the token's value. */
-/* T_ID, T_STRING: token string value.
- For T_ID, this is not truncated as is tokid. */
-struct string tokstr;
-\f
-/* Static variables. */
+ char tokid [LONG_NAME_LEN + 1]; /* T_ID: the identifier. */
-/* Pointer to next token in getl_buf. */
-static char *prog;
+ struct string tokstr; /* T_ID, T_STRING: token string value.
+ For T_ID, this is not truncated as is
+ tokid. */
-/* Nonzero only if this line ends with a terminal dot. */
-static int dot;
+ char *prog; /* Pointer to next token in line_buffer. */
+ bool dot; /* True only if this line ends with a terminal dot. */
-/* Nonzero only if the last token returned was T_STOP. */
-static int eof;
+ int put_token ; /* If nonzero, next token returned by lex_get().
+ Used only in exceptional circumstances. */
+
+ struct string put_tokstr;
+ double put_tokval;
+};
-/* If nonzero, next token returned by lex_get().
- Used only in exceptional circumstances. */
-static int put_token;
-static struct string put_tokstr;
-static double put_tokval;
-static int parse_id (void);
+static int parse_id (struct lexer *);
/* How a string represents its contents. */
enum string_type
HEX_STRING /* Hexadecimal digits. */
};
-static int parse_string (enum string_type);
+static int parse_string (struct lexer *, enum string_type);
#if DUMP_TOKENS
-static void dump_token (void);
+static void dump_token (struct lexer *);
#endif
\f
/* Initialization. */
/* Initializes the lexer. */
-void
-lex_init (void)
+struct lexer *
+lex_create (struct source_stream *ss)
+{
+ struct lexer *lexer = xzalloc (sizeof (*lexer));
+
+ ds_init_empty (&lexer->tokstr);
+ ds_init_empty (&lexer->put_tokstr);
+ ds_init_empty (&lexer->line_buffer);
+ lexer->ss = ss;
+
+ return lexer;
+}
+
+struct source_stream *
+lex_get_source_stream (const struct lexer *lex)
{
- ds_init_empty (&tokstr);
- ds_init_empty (&put_tokstr);
- if (!lex_get_line ())
- eof = true;
+ return lex->ss;
}
+
void
-lex_done (void)
+lex_destroy (struct lexer *lexer)
{
- ds_destroy (&put_tokstr);
- ds_destroy (&tokstr);
+ if ( NULL != lexer )
+ {
+ ds_destroy (&lexer->put_tokstr);
+ ds_destroy (&lexer->tokstr);
+ ds_destroy (&lexer->line_buffer);
+
+ free (lexer);
+ }
}
\f
/* Common functions. */
-/* Copies put_token, put_tokstr, put_tokval into token, tokstr,
+/* Copies put_token, lexer->put_tokstr, put_tokval into token, tokstr,
tokval, respectively, and sets tokid appropriately. */
static void
-restore_token (void)
+restore_token (struct lexer *lexer)
{
- assert (put_token != 0);
- token = put_token;
- ds_assign_string (&tokstr, &put_tokstr);
- str_copy_trunc (tokid, sizeof tokid, ds_cstr (&tokstr));
- tokval = put_tokval;
- put_token = 0;
+ assert (lexer->put_token != 0);
+ lexer->token = lexer->put_token;
+ ds_assign_string (&lexer->tokstr, &lexer->put_tokstr);
+ str_copy_trunc (lexer->tokid, sizeof lexer->tokid, ds_cstr (&lexer->tokstr));
+ lexer->tokval = lexer->put_tokval;
+ lexer->put_token = 0;
}
-/* Copies token, tokstr, tokval into put_token, put_tokstr,
- put_tokval respectively. */
+/* Copies token, tokstr, lexer->tokval into lexer->put_token, put_tokstr,
+ put_lexer->tokval respectively. */
static void
-save_token (void)
+save_token (struct lexer *lexer)
{
- put_token = token;
- ds_assign_string (&put_tokstr, &tokstr);
- put_tokval = tokval;
+ lexer->put_token = lexer->token;
+ ds_assign_string (&lexer->put_tokstr, &lexer->tokstr);
+ lexer->put_tokval = lexer->tokval;
}
/* Parses a single token, setting appropriate global variables to
indicate the token's attributes. */
void
-lex_get (void)
+lex_get (struct lexer *lexer)
{
+ /* Find a token. */
+ for (;;)
+ {
+ if (NULL == lexer->prog && ! lex_get_line (lexer) )
+ {
+ lexer->token = T_STOP;
+ return;
+ }
+
/* If a token was pushed ahead, return it. */
- if (put_token)
+ if (lexer->put_token)
{
- restore_token ();
+ restore_token (lexer);
#if DUMP_TOKENS
- dump_token ();
+ dump_token (lexer);
#endif
return;
}
- /* Find a token. */
for (;;)
{
/* Skip whitespace. */
- if (eof)
- {
- token = T_STOP;
- return;
- }
+ while (isspace ((unsigned char) *lexer->prog))
+ lexer->prog++;
- for (;;)
- {
- while (isspace ((unsigned char) *prog))
- prog++;
- if (*prog)
+ if (*lexer->prog)
break;
- if (dot)
+ if (lexer->dot)
{
- dot = 0;
- token = '.';
+ lexer->dot = 0;
+ lexer->token = '.';
#if DUMP_TOKENS
- dump_token ();
+ dump_token (lexer);
#endif
return;
}
- else if (!lex_get_line ())
+ else if (!lex_get_line (lexer))
{
- eof = 1;
- token = T_STOP;
+ lexer->prog = NULL;
+ lexer->token = T_STOP;
#if DUMP_TOKENS
- dump_token ();
+ dump_token (lexer);
#endif
return;
}
- if (put_token)
+ if (lexer->put_token)
{
- restore_token ();
+ restore_token (lexer);
#if DUMP_TOKENS
- dump_token ();
+ dump_token (lexer);
#endif
return;
}
/* Actually parse the token. */
- ds_clear (&tokstr);
+ ds_clear (&lexer->tokstr);
- switch (*prog)
+ switch (*lexer->prog)
{
case '-': case '.':
case '0': case '1': case '2': case '3': case '4':
want it as a number. When the syntax calls for a `-'
token, lex_negative_to_dash() must be used to break
negative numbers into two tokens. */
- if (*prog == '-')
+ if (*lexer->prog == '-')
{
- ds_put_char (&tokstr, *prog++);
- while (isspace ((unsigned char) *prog))
- prog++;
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
+ while (isspace ((unsigned char) *lexer->prog))
+ lexer->prog++;
- if (!isdigit ((unsigned char) *prog) && *prog != '.')
+ if (!isdigit ((unsigned char) *lexer->prog) && *lexer->prog != '.')
{
- token = '-';
+ lexer->token = '-';
break;
}
- token = T_NEG_NUM;
+ lexer->token = T_NEG_NUM;
}
else
- token = T_POS_NUM;
+ lexer->token = T_POS_NUM;
/* Parse the number, copying it into tokstr. */
- while (isdigit ((unsigned char) *prog))
- ds_put_char (&tokstr, *prog++);
- if (*prog == '.')
+ while (isdigit ((unsigned char) *lexer->prog))
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
+ if (*lexer->prog == '.')
{
- ds_put_char (&tokstr, *prog++);
- while (isdigit ((unsigned char) *prog))
- ds_put_char (&tokstr, *prog++);
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
+ while (isdigit ((unsigned char) *lexer->prog))
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
}
- if (*prog == 'e' || *prog == 'E')
+ if (*lexer->prog == 'e' || *lexer->prog == 'E')
{
- ds_put_char (&tokstr, *prog++);
- if (*prog == '+' || *prog == '-')
- ds_put_char (&tokstr, *prog++);
- while (isdigit ((unsigned char) *prog))
- ds_put_char (&tokstr, *prog++);
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
+ if (*lexer->prog == '+' || *lexer->prog == '-')
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
+ while (isdigit ((unsigned char) *lexer->prog))
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
}
/* Parse as floating point. */
- tokval = strtod (ds_cstr (&tokstr), &tail);
+ lexer->tokval = strtod (ds_cstr (&lexer->tokstr), &tail);
if (*tail)
{
msg (SE, _("%s does not form a valid number."),
- ds_cstr (&tokstr));
- tokval = 0.0;
+ ds_cstr (&lexer->tokstr));
+ lexer->tokval = 0.0;
- ds_clear (&tokstr);
- ds_put_char (&tokstr, '0');
+ ds_clear (&lexer->tokstr);
+ ds_put_char (&lexer->tokstr, '0');
}
break;
}
case '\'': case '"':
- token = parse_string (CHARACTER_STRING);
+ lexer->token = parse_string (lexer, CHARACTER_STRING);
break;
case '(': case ')': case ',': case '=': case '+': case '/':
- token = *prog++;
+ lexer->token = *lexer->prog++;
break;
case '*':
- if (*++prog == '*')
+ if (*++lexer->prog == '*')
{
- prog++;
- token = T_EXP;
+ lexer->prog++;
+ lexer->token = T_EXP;
}
else
- token = '*';
+ lexer->token = '*';
break;
case '<':
- if (*++prog == '=')
+ if (*++lexer->prog == '=')
{
- prog++;
- token = T_LE;
+ lexer->prog++;
+ lexer->token = T_LE;
}
- else if (*prog == '>')
+ else if (*lexer->prog == '>')
{
- prog++;
- token = T_NE;
+ lexer->prog++;
+ lexer->token = T_NE;
}
else
- token = T_LT;
+ lexer->token = T_LT;
break;
case '>':
- if (*++prog == '=')
+ if (*++lexer->prog == '=')
{
- prog++;
- token = T_GE;
+ lexer->prog++;
+ lexer->token = T_GE;
}
else
- token = T_GT;
+ lexer->token = T_GT;
break;
case '~':
- if (*++prog == '=')
+ if (*++lexer->prog == '=')
{
- prog++;
- token = T_NE;
+ lexer->prog++;
+ lexer->token = T_NE;
}
else
- token = T_NOT;
+ lexer->token = T_NOT;
break;
case '&':
- prog++;
- token = T_AND;
+ lexer->prog++;
+ lexer->token = T_AND;
break;
case '|':
- prog++;
- token = T_OR;
+ lexer->prog++;
+ lexer->token = T_OR;
break;
case 'b': case 'B':
- if (prog[1] == '\'' || prog[1] == '"')
- token = parse_string (BINARY_STRING);
+ if (lexer->prog[1] == '\'' || lexer->prog[1] == '"')
+ lexer->token = parse_string (lexer, BINARY_STRING);
else
- token = parse_id ();
+ lexer->token = parse_id (lexer);
break;
case 'o': case 'O':
- if (prog[1] == '\'' || prog[1] == '"')
- token = parse_string (OCTAL_STRING);
+ if (lexer->prog[1] == '\'' || lexer->prog[1] == '"')
+ lexer->token = parse_string (lexer, OCTAL_STRING);
else
- token = parse_id ();
+ lexer->token = parse_id (lexer);
break;
case 'x': case 'X':
- if (prog[1] == '\'' || prog[1] == '"')
- token = parse_string (HEX_STRING);
+ if (lexer->prog[1] == '\'' || lexer->prog[1] == '"')
+ lexer->token = parse_string (lexer, HEX_STRING);
else
- token = parse_id ();
+ lexer->token = parse_id (lexer);
break;
default:
- if (lex_is_id1 (*prog))
+ if (lex_is_id1 (*lexer->prog))
{
- token = parse_id ();
+ lexer->token = parse_id (lexer);
break;
}
else
{
- if (isgraph ((unsigned char) *prog))
- msg (SE, _("Bad character in input: `%c'."), *prog++);
+ if (isgraph ((unsigned char) *lexer->prog))
+ msg (SE, _("Bad character in input: `%c'."), *lexer->prog++);
else
- msg (SE, _("Bad character in input: `\\%o'."), *prog++);
+ msg (SE, _("Bad character in input: `\\%o'."), *lexer->prog++);
continue;
}
}
}
#if DUMP_TOKENS
- dump_token ();
+ dump_token (lexer);
#endif
}
tokstr.
Returns the correct token type. */
static int
-parse_id (void)
+parse_id (struct lexer *lexer)
{
- const char *start = prog;
- prog = lex_skip_identifier (start);
-
- ds_put_substring (&tokstr, ss_buffer (start, prog - start));
- str_copy_trunc (tokid, sizeof tokid, ds_cstr (&tokstr));
- return lex_id_to_token (ds_cstr (&tokstr), ds_length (&tokstr));
+ struct substring rest_of_line
+ = ss_substr (ds_ss (&lexer->line_buffer),
+ ds_pointer_to_position (&lexer->line_buffer, lexer->prog),
+ SIZE_MAX);
+ struct substring id = ss_head (rest_of_line,
+ lex_id_get_length (rest_of_line));
+ lexer->prog += ss_length (id);
+
+ ds_assign_substring (&lexer->tokstr, id);
+ str_copy_trunc (lexer->tokid, sizeof lexer->tokid, ds_cstr (&lexer->tokstr));
+ return lex_id_to_token (id);
}
/* Reports an error to the effect that subcommand SBC may only be
/* Reports an error to the effect that subcommand SBC is
missing. */
void
-lex_sbc_missing (const char *sbc)
+lex_sbc_missing (struct lexer *lexer, const char *sbc)
{
- lex_error (_("missing required subcommand %s"), sbc);
+ lex_error (lexer, _("missing required subcommand %s"), sbc);
}
/* Prints a syntax error message containing the current token and
given message MESSAGE (if non-null). */
void
-lex_error (const char *message, ...)
+lex_error (struct lexer *lexer, const char *message, ...)
{
char *token_rep;
char where[128];
- token_rep = lex_token_representation ();
- if (token == T_STOP)
+ token_rep = lex_token_representation (lexer);
+ if (lexer->token == T_STOP)
strcpy (where, "end of file");
- else if (token == '.')
+ else if (lexer->token == '.')
strcpy (where, "end of command");
else
snprintf (where, sizeof where, "`%s'", token_rep);
If not, flags a syntax error and returns an error command
completion code. */
int
-lex_end_of_command (void)
+lex_end_of_command (struct lexer *lexer)
{
- if (token != '.')
+ if (lexer->token != '.')
{
- lex_error (_("expecting end of command"));
+ lex_error (lexer, _("expecting end of command"));
return CMD_FAILURE;
}
else
/* Returns true if the current token is a number. */
bool
-lex_is_number (void)
+lex_is_number (struct lexer *lexer)
{
- return token == T_POS_NUM || token == T_NEG_NUM;
+ return lexer->token == T_POS_NUM || lexer->token == T_NEG_NUM;
}
/* Returns the value of the current token, which must be a
floating point number. */
double
-lex_number (void)
+lex_number (struct lexer *lexer)
{
- assert (lex_is_number ());
- return tokval;
+ assert (lex_is_number (lexer));
+ return lexer->tokval;
}
/* Returns true iff the current token is an integer. */
bool
-lex_is_integer (void)
+lex_is_integer (struct lexer *lexer)
{
- return (lex_is_number ()
- && tokval != NOT_LONG
- && tokval >= LONG_MIN
- && tokval <= LONG_MAX
- && floor (tokval) == tokval);
+ return (lex_is_number (lexer)
+ && lexer->tokval != NOT_LONG
+ && lexer->tokval >= LONG_MIN
+ && lexer->tokval <= LONG_MAX
+ && floor (lexer->tokval) == lexer->tokval);
}
/* Returns the value of the current token, which must be an
integer. */
long
-lex_integer (void)
+lex_integer (struct lexer *lexer)
{
- assert (lex_is_integer ());
- return tokval;
+ assert (lex_is_integer (lexer));
+ return lexer->tokval;
}
\f
/* Token matching functions. */
-/* If TOK is the current token, skips it and returns nonzero.
- Otherwise, returns zero. */
-int
-lex_match (int t)
+/* If TOK is the current token, skips it and returns true
+ Otherwise, returns false. */
+bool
+lex_match (struct lexer *lexer, int t)
{
- if (token == t)
+ if (lexer->token == t)
{
- lex_get ();
- return 1;
+ lex_get (lexer);
+ return true;
}
else
- return 0;
+ return false;
}
/* If the current token is the identifier S, skips it and returns
- nonzero. The identifier may be abbreviated to its first three
+ true. The identifier may be abbreviated to its first three
letters.
- Otherwise, returns zero. */
-int
-lex_match_id (const char *s)
+ Otherwise, returns false. */
+bool
+lex_match_id (struct lexer *lexer, const char *s)
{
- if (token == T_ID && lex_id_match (s, tokid))
+ if (lexer->token == T_ID
+ && lex_id_match (ss_cstr (s), ss_cstr (lexer->tokid)))
{
- lex_get ();
- return 1;
+ lex_get (lexer);
+ return true;
}
else
- return 0;
+ return false;
}
-/* If the current token is integer N, skips it and returns nonzero.
- Otherwise, returns zero. */
-int
-lex_match_int (int x)
+/* If the current token is integer N, skips it and returns true.
+ Otherwise, returns false. */
+bool
+lex_match_int (struct lexer *lexer, int x)
{
- if (lex_is_integer () && lex_integer () == x)
+ if (lex_is_integer (lexer) && lex_integer (lexer) == x)
{
- lex_get ();
- return 1;
+ lex_get (lexer);
+ return true;
}
else
- return 0;
+ return false;
}
\f
/* Forced matches. */
/* If this token is identifier S, fetches the next token and returns
nonzero.
Otherwise, reports an error and returns zero. */
-int
-lex_force_match_id (const char *s)
+bool
+lex_force_match_id (struct lexer *lexer, const char *s)
{
- if (token == T_ID && lex_id_match (s, tokid))
- {
- lex_get ();
- return 1;
- }
+ if (lex_match_id (lexer, s))
+ return true;
else
{
- lex_error (_("expecting `%s'"), s);
- return 0;
+ lex_error (lexer, _("expecting `%s'"), s);
+ return false;
}
}
/* If the current token is T, skips the token. Otherwise, reports an
- error and returns from the current function with return value 0. */
-int
-lex_force_match (int t)
+ error and returns from the current function with return value false. */
+bool
+lex_force_match (struct lexer *lexer, int t)
{
- if (token == t)
+ if (lexer->token == t)
{
- lex_get ();
- return 1;
+ lex_get (lexer);
+ return true;
}
else
{
- lex_error (_("expecting `%s'"), lex_token_name (t));
- return 0;
+ lex_error (lexer, _("expecting `%s'"), lex_token_name (t));
+ return false;
}
}
-/* If this token is a string, does nothing and returns nonzero.
- Otherwise, reports an error and returns zero. */
-int
-lex_force_string (void)
+/* If this token is a string, does nothing and returns true.
+ Otherwise, reports an error and returns false. */
+bool
+lex_force_string (struct lexer *lexer)
{
- if (token == T_STRING)
- return 1;
+ if (lexer->token == T_STRING)
+ return true;
else
{
- lex_error (_("expecting string"));
- return 0;
+ lex_error (lexer, _("expecting string"));
+ return false;
}
}
-/* If this token is an integer, does nothing and returns nonzero.
- Otherwise, reports an error and returns zero. */
-int
-lex_force_int (void)
+/* If this token is an integer, does nothing and returns true.
+ Otherwise, reports an error and returns false. */
+bool
+lex_force_int (struct lexer *lexer)
{
- if (lex_is_integer ())
- return 1;
+ if (lex_is_integer (lexer))
+ return true;
else
{
- lex_error (_("expecting integer"));
- return 0;
+ lex_error (lexer, _("expecting integer"));
+ return false;
}
}
-/* If this token is a number, does nothing and returns nonzero.
- Otherwise, reports an error and returns zero. */
-int
-lex_force_num (void)
+/* If this token is a number, does nothing and returns true.
+ Otherwise, reports an error and returns false. */
+bool
+lex_force_num (struct lexer *lexer)
{
- if (lex_is_number ())
- return 1;
- else
- {
- lex_error (_("expecting number"));
- return 0;
- }
+ if (lex_is_number (lexer))
+ return true;
+
+ lex_error (lexer, _("expecting number"));
+ return false;
}
-/* If this token is an identifier, does nothing and returns nonzero.
- Otherwise, reports an error and returns zero. */
-int
-lex_force_id (void)
+/* If this token is an identifier, does nothing and returns true.
+ Otherwise, reports an error and returns false. */
+bool
+lex_force_id (struct lexer *lexer)
{
- if (token == T_ID)
- return 1;
- else
- {
- lex_error (_("expecting identifier"));
- return 0;
- }
+ if (lexer->token == T_ID)
+ return true;
+
+ lex_error (lexer, _("expecting identifier"));
+ return false;
}
+
/* Weird token functions. */
/* Returns the first character of the next token, except that if the
alphanumeric return value doesn't guarantee an ID token, it could
also be a reserved-word token. */
int
-lex_look_ahead (void)
+lex_look_ahead (struct lexer *lexer)
{
- if (put_token)
- return put_token;
+ if (lexer->put_token)
+ return lexer->put_token;
for (;;)
{
- if (eof)
+ if (NULL == lexer->prog && ! lex_get_line (lexer) )
return 0;
for (;;)
{
- while (isspace ((unsigned char) *prog))
- prog++;
- if (*prog)
+ while (isspace ((unsigned char) *lexer->prog))
+ lexer->prog++;
+ if (*lexer->prog)
break;
- if (dot)
+ if (lexer->dot)
return '.';
- else if (!lex_get_line ())
+ else if (!lex_get_line (lexer))
return 0;
- if (put_token)
- return put_token;
+ if (lexer->put_token)
+ return lexer->put_token;
}
- if ((toupper ((unsigned char) *prog) == 'X'
- || toupper ((unsigned char) *prog) == 'B'
- || toupper ((unsigned char) *prog) == 'O')
- && (prog[1] == '\'' || prog[1] == '"'))
+ if ((toupper ((unsigned char) *lexer->prog) == 'X'
+ || toupper ((unsigned char) *lexer->prog) == 'B'
+ || toupper ((unsigned char) *lexer->prog) == 'O')
+ && (lexer->prog[1] == '\'' || lexer->prog[1] == '"'))
return '\'';
- return *prog;
+ return *lexer->prog;
}
}
/* Makes the current token become the next token to be read; the
current token is set to T. */
void
-lex_put_back (int t)
+lex_put_back (struct lexer *lexer, int t)
{
- save_token ();
- token = t;
+ save_token (lexer);
+ lexer->token = t;
}
/* Makes the current token become the next token to be read; the
current token is set to the identifier ID. */
void
-lex_put_back_id (const char *id)
+lex_put_back_id (struct lexer *lexer, const char *id)
{
- assert (lex_id_to_token (id, strlen (id)) == T_ID);
- save_token ();
- token = T_ID;
- ds_assign_cstr (&tokstr, id);
- str_copy_trunc (tokid, sizeof tokid, ds_cstr (&tokstr));
+ assert (lex_id_to_token (ss_cstr (id)) == T_ID);
+ save_token (lexer);
+ lexer->token = T_ID;
+ ds_assign_cstr (&lexer->tokstr, id);
+ str_copy_trunc (lexer->tokid, sizeof lexer->tokid, ds_cstr (&lexer->tokstr));
}
\f
/* Weird line processing functions. */
/* Returns the entire contents of the current line. */
const char *
-lex_entire_line (void)
+lex_entire_line (struct lexer *lexer)
{
- return ds_cstr (&getl_buf);
+ return ds_cstr (&lexer->line_buffer);
+}
+
+const struct string *
+lex_entire_line_ds (struct lexer *lexer)
+{
+ return &lexer->line_buffer;
}
/* As lex_entire_line(), but only returns the part of the current line
If END_DOT is non-null, stores nonzero into *END_DOT if the line
ends with a terminal dot, or zero if it doesn't. */
const char *
-lex_rest_of_line (int *end_dot)
+lex_rest_of_line (struct lexer *lexer, int *end_dot)
{
if (end_dot)
- *end_dot = dot;
- return prog;
+ *end_dot = lexer->dot;
+ return lexer->prog;
}
/* Causes the rest of the current input line to be ignored for
tokenization purposes. */
void
-lex_discard_line (void)
+lex_discard_line (struct lexer *lexer)
{
- prog = ds_end (&getl_buf);
- dot = put_token = 0;
+ ds_cstr (&lexer->line_buffer); /* Ensures ds_end points to something valid */
+ lexer->prog = ds_end (&lexer->line_buffer);
+ lexer->dot = false;
+ lexer->put_token = 0;
}
-/* Sets the current position in the current line to P, which must be
- in getl_buf. */
-void
-lex_set_prog (char *p)
-{
- prog = p;
-}
/* Discards the rest of the current command.
When we're reading commands from a file, we skip tokens until
the user doesn't want to finish typing a command that will be
ignored anyway. */
void
-lex_discard_rest_of_command (void)
+lex_discard_rest_of_command (struct lexer *lexer)
{
- if (!getl_is_interactive ())
+ if (!getl_is_interactive (lexer->ss))
{
- while (token != T_STOP && token != '.')
- lex_get ();
+ while (lexer->token != T_STOP && lexer->token != '.')
+ lex_get (lexer);
}
else
- lex_discard_line ();
+ lex_discard_line (lexer);
}
\f
/* Weird line reading functions. */
}
}
+/* Prepares LINE, which is subject to the given SYNTAX rules, for
+ tokenization by stripping comments and determining whether it
+ is the beginning or end of a command and storing into
+ *LINE_STARTS_COMMAND and *LINE_ENDS_COMMAND appropriately. */
+void
+lex_preprocess_line (struct string *line,
+ enum getl_syntax syntax,
+ bool *line_starts_command,
+ bool *line_ends_command)
+{
+ strip_comments (line);
+ ds_rtrim (line, ss_cstr (CC_SPACES));
+ *line_ends_command = (ds_chomp (line, get_endcmd ())
+ || (ds_is_empty (line) && get_nulline ()));
+ *line_starts_command = false;
+ if (syntax == GETL_BATCH)
+ {
+ int first = ds_first (line);
+ *line_starts_command = !isspace (first);
+ if (first == '+' || first == '-')
+ *ds_data (line) = ' ';
+ }
+}
+
+/* Reads a line, without performing any preprocessing.
+ Sets *SYNTAX, if SYNTAX is non-null, to the line's syntax
+ mode. */
+bool
+lex_get_line_raw (struct lexer *lexer, enum getl_syntax *syntax)
+{
+ enum getl_syntax dummy;
+ bool ok = getl_read_line (lexer->ss, &lexer->line_buffer,
+ syntax != NULL ? syntax : &dummy);
+ return ok;
+}
+
/* Reads a line for use by the tokenizer, and preprocesses it by
removing comments, stripping trailing whitespace and the
terminal dot, and removing leading indentors. */
bool
-lex_get_line (void)
+lex_get_line (struct lexer *lexer)
{
- struct string *line = &getl_buf;
- bool interactive;
-
- if (!getl_read_line (&interactive))
- return false;
+ bool line_starts_command;
+ enum getl_syntax syntax;
- strip_comments (line);
- ds_rtrim (line, ss_cstr (CC_SPACES));
-
- /* Check for and remove terminal dot. */
- dot = (ds_chomp (line, get_endcmd ())
- || (ds_is_empty (line) && get_nulline ()));
-
- /* Strip leading indentors or insert a terminal dot (unless the
- line was obtained interactively). */
- if (!interactive)
+ if (!lex_get_line_raw (lexer, &syntax))
{
- int first = ds_first (line);
-
- if (first == '+' || first == '-')
- *ds_data (line) = ' ';
- else if (first != EOF && !isspace (first))
- put_token = '.';
+ lexer->prog = NULL;
+ return false;
}
- prog = ds_cstr (line);
+ lex_preprocess_line (&lexer->line_buffer, syntax,
+ &line_starts_command, &lexer->dot);
+
+ if (line_starts_command)
+ lexer->put_token = '.';
+ lexer->prog = ds_cstr (&lexer->line_buffer);
return true;
}
\f
/* Token names. */
-/* Returns the name of a token in a static buffer. */
+/* Returns the name of a token. */
const char *
lex_token_name (int token)
{
- if (token >= T_FIRST_KEYWORD && token <= T_LAST_KEYWORD)
- return keywords[token - T_FIRST_KEYWORD];
-
- if (token < 256)
+ if (lex_is_keyword (token))
+ return lex_id_name (token);
+ else if (token < 256)
{
- static char t[2];
- t[0] = token;
- return t;
+ static char t[256][2];
+ char *s = t[token];
+ s[0] = token;
+ s[1] = '\0';
+ return s;
}
-
- return _("<ERROR>");
+ else
+ NOT_REACHED ();
}
/* Returns an ASCII representation of the current token as a
malloc()'d string. */
char *
-lex_token_representation (void)
+lex_token_representation (struct lexer *lexer)
{
char *token_rep;
- switch (token)
+ switch (lexer->token)
{
case T_ID:
case T_POS_NUM:
case T_NEG_NUM:
- return ds_xstrdup (&tokstr);
+ return ds_xstrdup (&lexer->tokstr);
break;
case T_STRING:
int hexstring = 0;
char *sp, *dp;
- for (sp = ds_cstr (&tokstr); sp < ds_end (&tokstr); sp++)
+ for (sp = ds_cstr (&lexer->tokstr); sp < ds_end (&lexer->tokstr); sp++)
if (!isprint ((unsigned char) *sp))
{
hexstring = 1;
break;
}
- token_rep = xmalloc (2 + ds_length (&tokstr) * 2 + 1 + 1);
+ token_rep = xmalloc (2 + ds_length (&lexer->tokstr) * 2 + 1 + 1);
dp = token_rep;
if (hexstring)
*dp++ = '\'';
if (!hexstring)
- for (sp = ds_cstr (&tokstr); *sp; )
+ for (sp = ds_cstr (&lexer->tokstr); *sp; )
{
if (*sp == '\'')
*dp++ = '\'';
*dp++ = (unsigned char) *sp++;
}
else
- for (sp = ds_cstr (&tokstr); sp < ds_end (&tokstr); sp++)
+ for (sp = ds_cstr (&lexer->tokstr); sp < ds_end (&lexer->tokstr); sp++)
{
*dp++ = (((unsigned char) *sp) >> 4)["0123456789ABCDEF"];
*dp++ = (((unsigned char) *sp) & 15)["0123456789ABCDEF"];
return xstrdup ("**");
default:
- if (token >= T_FIRST_KEYWORD && token <= T_LAST_KEYWORD)
- return xstrdup (keywords [token - T_FIRST_KEYWORD]);
- else
- {
- token_rep = xmalloc (2);
- token_rep[0] = token;
- token_rep[1] = '\0';
- return token_rep;
- }
+ return xstrdup (lex_token_name (lexer->token));
}
- assert (0);
+ NOT_REACHED ();
}
\f
/* Really weird functions. */
of syntax then this function is called to rip it off of a
number. */
void
-lex_negative_to_dash (void)
+lex_negative_to_dash (struct lexer *lexer)
{
- if (token == T_NEG_NUM)
+ if (lexer->token == T_NEG_NUM)
{
- token = T_POS_NUM;
- tokval = -tokval;
- ds_assign_substring (&tokstr, ds_substr (&tokstr, 1, SIZE_MAX));
- save_token ();
- token = '-';
+ lexer->token = T_POS_NUM;
+ lexer->tokval = -lexer->tokval;
+ ds_assign_substring (&lexer->tokstr, ds_substr (&lexer->tokstr, 1, SIZE_MAX));
+ save_token (lexer);
+ lexer->token = '-';
}
}
-/* We're not at eof any more. */
-void
-lex_reset_eof (void)
-{
- eof = 0;
-}
-
/* Skip a COMMENT command. */
void
-lex_skip_comment (void)
+lex_skip_comment (struct lexer *lexer)
{
for (;;)
{
- if (!lex_get_line ())
+ if (!lex_get_line (lexer))
{
- put_token = T_STOP;
- eof = 1;
+ lexer->put_token = T_STOP;
+ lexer->prog = NULL;
return;
}
- if (put_token == '.')
+ if (lexer->put_token == '.')
break;
- prog = ds_end (&getl_buf);
- if (dot)
+ ds_cstr (&lexer->line_buffer); /* Ensures ds_end will point to a valid char */
+ lexer->prog = ds_end (&lexer->line_buffer);
+ if (lexer->dot)
break;
}
}
hex digits, according to TYPE. The string is converted to
characters having the specified values. */
static void
-convert_numeric_string_to_char_string (enum string_type type)
+convert_numeric_string_to_char_string (struct lexer *lexer,
+ enum string_type type)
{
const char *base_name;
int base;
chars_per_byte = 2;
break;
default:
- abort ();
+ NOT_REACHED ();
}
- byte_cnt = ds_length (&tokstr) / chars_per_byte;
- if (ds_length (&tokstr) % chars_per_byte)
+ byte_cnt = ds_length (&lexer->tokstr) / chars_per_byte;
+ if (ds_length (&lexer->tokstr) % chars_per_byte)
msg (SE, _("String of %s digits has %d characters, which is not a "
"multiple of %d."),
- base_name, ds_length (&tokstr), chars_per_byte);
+ base_name, ds_length (&lexer->tokstr), chars_per_byte);
- p = ds_cstr (&tokstr);
+ p = ds_cstr (&lexer->tokstr);
for (i = 0; i < byte_cnt; i++)
{
int value;
value = value * base + v;
}
- ds_cstr (&tokstr)[i] = (unsigned char) value;
+ ds_cstr (&lexer->tokstr)[i] = (unsigned char) value;
}
- ds_truncate (&tokstr, byte_cnt);
+ ds_truncate (&lexer->tokstr, byte_cnt);
}
/* Parses a string from the input buffer into tokstr. The input
- buffer pointer prog must point to the initial single or double
+ buffer pointer lexer->prog must point to the initial single or double
quote. TYPE indicates the type of string to be parsed.
Returns token type. */
static int
-parse_string (enum string_type type)
+parse_string (struct lexer *lexer, enum string_type type)
{
+ if (type != CHARACTER_STRING)
+ lexer->prog++;
+
/* Accumulate the entire string, joining sections indicated by +
signs. */
for (;;)
{
/* Single or double quote. */
- int c = *prog++;
+ int c = *lexer->prog++;
/* Accumulate section. */
for (;;)
{
/* Check end of line. */
- if (*prog == '\0')
+ if (*lexer->prog == '\0')
{
msg (SE, _("Unterminated string constant."));
goto finish;
}
/* Double quote characters to embed them in strings. */
- if (*prog == c)
+ if (*lexer->prog == c)
{
- if (prog[1] == c)
- prog++;
+ if (lexer->prog[1] == c)
+ lexer->prog++;
else
break;
}
- ds_put_char (&tokstr, *prog++);
+ ds_put_char (&lexer->tokstr, *lexer->prog++);
}
- prog++;
+ lexer->prog++;
/* Skip whitespace after final quote mark. */
- if (eof)
+ if (lexer->prog == NULL)
break;
for (;;)
{
- while (isspace ((unsigned char) *prog))
- prog++;
- if (*prog)
+ while (isspace ((unsigned char) *lexer->prog))
+ lexer->prog++;
+ if (*lexer->prog)
break;
- if (dot)
+ if (lexer->dot)
goto finish;
- if (!lex_get_line ())
+ if (!lex_get_line (lexer))
goto finish;
}
/* Skip plus sign. */
- if (*prog != '+')
+ if (*lexer->prog != '+')
break;
- prog++;
+ lexer->prog++;
/* Skip whitespace after plus sign. */
- if (eof)
+ if (lexer->prog == NULL)
break;
for (;;)
{
- while (isspace ((unsigned char) *prog))
- prog++;
- if (*prog)
+ while (isspace ((unsigned char) *lexer->prog))
+ lexer->prog++;
+ if (*lexer->prog)
break;
- if (dot)
+ if (lexer->dot)
goto finish;
- if (!lex_get_line ())
+ if (!lex_get_line (lexer))
{
msg (SE, _("Unexpected end of file in string concatenation."));
goto finish;
}
/* Ensure that a valid string follows. */
- if (*prog != '\'' && *prog != '"')
+ if (*lexer->prog != '\'' && *lexer->prog != '"')
{
msg (SE, _("String expected following `+'."));
goto finish;
into one large string. */
finish:
if (type != CHARACTER_STRING)
- convert_numeric_string_to_char_string (type);
+ convert_numeric_string_to_char_string (lexer, type);
- if (ds_length (&tokstr) > 255)
+ if (ds_length (&lexer->tokstr) > 255)
{
msg (SE, _("String exceeds 255 characters in length (%d characters)."),
- ds_length (&tokstr));
- ds_truncate (&tokstr, 255);
+ ds_length (&lexer->tokstr));
+ ds_truncate (&lexer->tokstr, 255);
}
- {
- /* FIXME. */
- size_t i;
- int warned = 0;
-
- for (i = 0; i < ds_length (&tokstr); i++)
- if (ds_cstr (&tokstr)[i] == 0)
- {
- if (!warned)
- {
- msg (SE, _("Sorry, literal strings may not contain null "
- "characters. Replacing with spaces."));
- warned = 1;
- }
- ds_cstr (&tokstr)[i] = ' ';
- }
- }
-
return T_STRING;
}
\f
/* Reads one token from the lexer and writes a textual representation
on stdout for debugging purposes. */
static void
-dump_token (void)
+dump_token (struct lexer *lexer)
{
{
const char *curfn;
int curln;
- getl_location (&curfn, &curln);
+ curln = getl_source_location (lexer->ss);
+ curfn = getl_source_name (lexer->ss);
if (curfn)
fprintf (stderr, "%s:%d\t", curfn, curln);
}
- switch (token)
+ switch (lexer->token)
{
case T_ID:
- fprintf (stderr, "ID\t%s\n", tokid);
+ fprintf (stderr, "ID\t%s\n", lexer->tokid);
break;
case T_POS_NUM:
case T_NEG_NUM:
- fprintf (stderr, "NUM\t%f\n", tokval);
+ fprintf (stderr, "NUM\t%f\n", lexer->tokval);
break;
case T_STRING:
- fprintf (stderr, "STRING\t\"%s\"\n", ds_cstr (&tokstr));
+ fprintf (stderr, "STRING\t\"%s\"\n", ds_cstr (&lexer->tokstr));
break;
case T_STOP:
break;
default:
- if (token >= T_FIRST_KEYWORD && token <= T_LAST_KEYWORD)
- fprintf (stderr, "KEYWORD\t%s\n", lex_token_name (token));
+ if (lex_is_keyword (lexer->token))
+ fprintf (stderr, "KEYWORD\t%s\n", lex_token_name (lexer->token));
else
- fprintf (stderr, "PUNCT\t%c\n", token);
+ fprintf (stderr, "PUNCT\t%c\n", lexer->token);
break;
}
}
#endif /* DUMP_TOKENS */
+
+
+/* Token Accessor Functions */
+
+int
+lex_token (const struct lexer *lexer)
+{
+ return lexer->token;
+}
+
+double
+lex_tokval (const struct lexer *lexer)
+{
+ return lexer->tokval;
+}
+
+const char *
+lex_tokid (const struct lexer *lexer)
+{
+ return lexer->tokid;
+}
+
+const struct string *
+lex_tokstr (const struct lexer *lexer)
+{
+ return &lexer->tokstr;
+}