/* PSPP - a program for statistical analysis.
- Copyright (C) 1997-9, 2000, 2009, 2010 Free Software Foundation, Inc.
+ Copyright (C) 1997-9, 2000, 2009, 2010, 2011, 2012 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 published by
#include "language/lexer/variable-parser.h"
#include <ctype.h>
+#include <limits.h>
#include <stdbool.h>
#include <stdlib.h>
+#include "data/dataset.h"
#include "data/dictionary.h"
-#include "data/procedure.h"
#include "data/variable.h"
#include "language/lexer/lexer.h"
#include "libpspp/assertion.h"
#include "libpspp/cast.h"
#include "libpspp/hash-functions.h"
+#include "libpspp/i18n.h"
#include "libpspp/hmapx.h"
#include "libpspp/message.h"
#include "libpspp/misc.h"
#include "libpspp/str.h"
#include "libpspp/stringi-set.h"
+#include "math/interaction.h"
+
+#include "gl/c-ctype.h"
#include "gl/xalloc.h"
#include "gettext.h"
#define _(msgid) gettext (msgid)
-static struct variable * var_set_get_var (const struct var_set *, size_t );
-
+static struct variable *var_set_get_var (const struct var_set *, size_t);
static struct variable *var_set_lookup_var (const struct var_set *,
const char *);
-
static bool var_set_lookup_var_idx (const struct var_set *, const char *,
size_t *);
+static bool var_set_get_names_must_be_ids (const struct var_set *);
+
+static bool
+is_name_token (const struct lexer *lexer, bool names_must_be_ids)
+{
+ return (lex_token (lexer) == T_ID
+ || (!names_must_be_ids && lex_token (lexer) == T_STRING));
+}
+static bool
+is_vs_name_token (const struct lexer *lexer, const struct var_set *vs)
+{
+ return is_name_token (lexer, var_set_get_names_must_be_ids (vs));
+}
+static bool
+is_dict_name_token (const struct lexer *lexer, const struct dictionary *d)
+{
+ return is_name_token (lexer, dict_get_names_must_be_ids (d));
+}
/* Parses a name as a variable within VS. Sets *IDX to the
variable's index and returns true if successful. On failure
emits an error message and returns false. */
static bool
parse_vs_variable_idx (struct lexer *lexer, const struct var_set *vs,
- size_t *idx)
+ size_t *idx)
{
assert (idx != NULL);
- if (lex_token (lexer) != T_ID)
+ if (!is_vs_name_token (lexer, vs))
{
lex_error (lexer, _("expecting variable name"));
return false;
}
- else if (var_set_lookup_var_idx (vs, lex_tokid (lexer), idx))
+ else if (var_set_lookup_var_idx (vs, lex_tokcstr (lexer), idx))
{
lex_get (lexer);
return true;
}
else
{
- msg (SE, _("%s is not a variable name."), lex_tokid (lexer));
+ msg (SE, _("%s is not a variable name."), lex_tokcstr (lexer));
return false;
}
}
"All variables in this variable list must have the "
"same width. %s will be omitted from the list."),
var_get_name ((*v)[0]), add_name, add_name);
- else if ((pv_opts & PV_NO_DUPLICATE) && included[idx])
+ else if ((pv_opts & PV_NO_DUPLICATE) && included && included[idx])
msg (SE, _("Variable %s appears twice in variable list."), add_name);
- else if ((pv_opts & PV_DUPLICATE) || !included[idx])
+ else if ((pv_opts & PV_DUPLICATE) || !included || !included[idx])
{
if (*nv >= *mv)
{
if (pv_opts & PV_SINGLE)
break;
- lex_match (lexer, ',');
+ lex_match (lexer, T_COMMA);
}
while (lex_token (lexer) == T_ALL
- || (lex_token (lexer) == T_ID && var_set_lookup_var (vs, lex_tokid (lexer)) != NULL));
+ || (is_vs_name_token (lexer, vs)
+ && var_set_lookup_var (vs, lex_tokcstr (lexer)) != NULL));
if (*nv == 0)
goto fail;
return 0;
}
-/* Extracts a numeric suffix from variable name S, copying it
- into string R. Sets *D to the length of R and *N to its
- value. */
-static int
-extract_num (char *s, char *r, int *n, int *d)
+char *
+parse_DATA_LIST_var (struct lexer *lexer, const struct dictionary *d)
{
- char *cp;
+ if (!is_dict_name_token (lexer, d))
+ {
+ lex_error (lexer, "expecting variable name");
+ return NULL;
+ }
+ if (!dict_id_is_valid (d, lex_tokcstr (lexer), true))
+ return NULL;
+
+ char *name = xstrdup (lex_tokcstr (lexer));
+ lex_get (lexer);
+ return name;
+}
- /* Find first digit. */
- cp = s + strlen (s) - 1;
- while (isdigit ((unsigned char) *cp) && cp > s)
- cp--;
- cp++;
+/* Attempts to break UTF-8 encoded NAME into a root (whose contents are
+ arbitrary except that it does not end in a digit) followed by an integer
+ numeric suffix. On success, stores the value of the suffix into *NUMBERP,
+ the number of digits in the suffix into *N_DIGITSP, and returns the number
+ of bytes in the root. On failure, returns 0. */
+static int
+extract_numeric_suffix (const char *name,
+ unsigned long int *numberp, int *n_digitsp)
+{
+ size_t root_len, n_digits;
+ size_t i;
- /* Extract root. */
- strncpy (r, s, cp - s);
- r[cp - s] = 0;
+ /* Count length of root. */
+ root_len = 1; /* Valid identifier never starts with digit. */
+ for (i = 1; name[i] != '\0'; i++)
+ if (!c_isdigit (name[i]))
+ root_len = i + 1;
+ n_digits = i - root_len;
- /* Count initial zeros. */
- *n = *d = 0;
- while (*cp == '0')
+ if (n_digits == 0)
{
- (*d)++;
- cp++;
+ msg (SE, _("`%s' cannot be used with TO because it does not end in "
+ "a digit."), name);
+ return 0;
}
- /* Extract value. */
- while (isdigit ((unsigned char) *cp))
+ *numberp = strtoull (name + root_len, NULL, 10);
+ if (*numberp == ULONG_MAX)
{
- (*d)++;
- *n = (*n * 10) + (*cp - '0');
- cp++;
+ msg (SE, _("Numeric suffix on `%s' is larger than supported with TO."),
+ name);
+ return 0;
}
+ *n_digitsp = n_digits;
+ return root_len;
+}
- /* Sanity check. */
- if (*n == 0 && *d == 0)
+static bool
+add_var_name (char *name,
+ char ***names, size_t *n_vars, size_t *allocated_vars,
+ struct stringi_set *set, int pv_opts)
+{
+ if (pv_opts & PV_NO_DUPLICATE && !stringi_set_insert (set, name))
{
- msg (SE, _("incorrect use of TO convention"));
- return 0;
+ msg (SE, _("Variable %s appears twice in variable list."),
+ name);
+ return false;
}
- return 1;
+
+ if (*n_vars >= *allocated_vars)
+ *names = x2nrealloc (*names, allocated_vars, sizeof **names);
+ (*names)[(*n_vars)++] = name;
+ return true;
}
/* Parses a list of variable names according to the DATA LIST version
of the TO convention. */
bool
-parse_DATA_LIST_vars (struct lexer *lexer, char ***names,
- size_t *nnames, int pv_opts)
-{
- int n1, n2;
- int d1, d2;
- int n;
- size_t nvar, mvar;
- char name1[VAR_NAME_LEN + 1], name2[VAR_NAME_LEN + 1];
- char root1[VAR_NAME_LEN + 1], root2[VAR_NAME_LEN + 1];
+parse_DATA_LIST_vars (struct lexer *lexer, const struct dictionary *dict,
+ char ***namesp, size_t *n_varsp, int pv_opts)
+{
+ char **names;
+ size_t n_vars;
+ size_t allocated_vars;
+
struct stringi_set set;
- int success = 0;
- assert (names != NULL);
- assert (nnames != NULL);
+ char *name1 = NULL;
+
+ bool ok = false;
+
assert ((pv_opts & ~(PV_APPEND | PV_SINGLE
| PV_NO_SCRATCH | PV_NO_DUPLICATE)) == 0);
stringi_set_init (&set);
if (pv_opts & PV_APPEND)
{
- nvar = mvar = *nnames;
+ n_vars = allocated_vars = *n_varsp;
+ names = *namesp;
if (pv_opts & PV_NO_DUPLICATE)
{
size_t i;
- for (i = 0; i < nvar; i++)
- stringi_set_insert (&set, (*names)[i]);
+ for (i = 0; i < n_vars; i++)
+ stringi_set_insert (&set, names[i]);
}
}
else
{
- nvar = mvar = 0;
- *names = NULL;
+ n_vars = allocated_vars = 0;
+ names = NULL;
}
do
{
- if (lex_token (lexer) != T_ID)
- {
- lex_error (lexer, "expecting variable name");
- goto fail;
- }
- if (dict_class_from_id (lex_tokid (lexer)) == DC_SCRATCH
- && (pv_opts & PV_NO_SCRATCH))
+ name1 = parse_DATA_LIST_var (lexer, dict);
+ if (!name1)
+ goto exit;
+ if (dict_class_from_id (name1) == DC_SCRATCH && pv_opts & PV_NO_SCRATCH)
{
msg (SE, _("Scratch variables not allowed here."));
- goto fail;
+ goto exit;
}
- strcpy (name1, lex_tokid (lexer));
- lex_get (lexer);
- if (lex_token (lexer) == T_TO)
+ if (lex_match (lexer, T_TO))
{
- lex_get (lexer);
- if (lex_token (lexer) != T_ID)
- {
- lex_error (lexer, "expecting variable name");
- goto fail;
- }
- strcpy (name2, lex_tokid (lexer));
- lex_get (lexer);
+ unsigned long int num1, num2;
+ int n_digits1, n_digits2;
+ int root_len1, root_len2;
+ unsigned long int number;
- if (!extract_num (name1, root1, &n1, &d1)
- || !extract_num (name2, root2, &n2, &d2))
- goto fail;
+ char *name2 = parse_DATA_LIST_var (lexer, dict);
- if (strcasecmp (root1, root2))
+ root_len1 = extract_numeric_suffix (name1, &num1, &n_digits1);
+ if (root_len1 == 0)
+ goto exit;
+
+ root_len2 = extract_numeric_suffix (name2, &num2, &n_digits2);
+ if (root_len2 == 0)
+ goto exit;
+
+ if (root_len1 != root_len2 || memcasecmp (name1, name2, root_len1))
{
msg (SE, _("Prefixes don't match in use of TO convention."));
- goto fail;
+ goto exit;
}
- if (n1 > n2)
+ if (num1 > num2)
{
msg (SE, _("Bad bounds in use of TO convention."));
- goto fail;
+ goto exit;
}
- if (d2 > d1)
- d2 = d1;
- if (mvar < nvar + (n2 - n1 + 1))
+ for (number = num1; number <= num2; number++)
{
- mvar += ROUND_UP (n2 - n1 + 1, 16);
- *names = xnrealloc (*names, mvar, sizeof **names);
- }
-
- for (n = n1; n <= n2; n++)
- {
- char name[VAR_NAME_LEN + 1];
- sprintf (name, "%s%0*d", root1, d1, n);
-
- if (pv_opts & PV_NO_DUPLICATE && !stringi_set_insert (&set, name))
+ char *name = xasprintf ("%.*s%0*lu",
+ root_len1, name1,
+ n_digits1, number);
+ if (!add_var_name (name, &names, &n_vars, &allocated_vars,
+ &set, pv_opts))
{
- msg (SE, _("Variable %s appears twice in variable list."),
- name);
- goto fail;
+ free (name);
+ goto exit;
}
- (*names)[nvar] = xstrdup (name);
- nvar++;
}
+
+ free (name1);
+ name1 = NULL;
+ free (name2);
}
else
{
- if (nvar >= mvar)
- {
- mvar += 16;
- *names = xnrealloc (*names, mvar, sizeof **names);
- }
- (*names)[nvar++] = xstrdup (name1);
+ if (!add_var_name (name1, &names, &n_vars, &allocated_vars,
+ &set, pv_opts))
+ goto exit;
+ name1 = NULL;
}
- lex_match (lexer, ',');
+ lex_match (lexer, T_COMMA);
if (pv_opts & PV_SINGLE)
break;
}
while (lex_token (lexer) == T_ID);
- success = 1;
+ ok = true;
-fail:
- *nnames = nvar;
+exit:
stringi_set_destroy (&set);
- if (!success)
+ if (ok)
+ {
+ *namesp = names;
+ *n_varsp = n_vars;
+ }
+ else
{
int i;
- for (i = 0; i < nvar; i++)
- free ((*names)[i]);
- free (*names);
- *names = NULL;
- *nnames = 0;
+ for (i = 0; i < n_vars; i++)
+ free (names[i]);
+ free (names);
+ *namesp = NULL;
+ *n_varsp = 0;
+
+ free (name1);
}
- return success;
+ return ok;
}
/* Registers each of the NAMES[0...NNAMES - 1] in POOL, as well
parse_DATA_LIST_vars(), except that all allocations are taken
from the given POOL. */
bool
-parse_DATA_LIST_vars_pool (struct lexer *lexer, struct pool *pool,
+parse_DATA_LIST_vars_pool (struct lexer *lexer, const struct dictionary *dict,
+ struct pool *pool,
char ***names, size_t *nnames, int pv_opts)
{
int retval;
re-free it later. */
assert (!(pv_opts & PV_APPEND));
- retval = parse_DATA_LIST_vars (lexer, names, nnames, pv_opts);
+ retval = parse_DATA_LIST_vars (lexer, dict, names, nnames, pv_opts);
if (retval)
register_vars_pool (pool, *names, *nnames);
return retval;
assert (names != NULL);
assert (nnames != NULL);
- assert ((pv_opts & ~PV_APPEND) == 0);
if (!(pv_opts & PV_APPEND))
{
*names = NULL;
*nnames = 0;
}
- while (lex_token (lexer) == T_ID || lex_token (lexer) == T_ALL)
+ while (is_dict_name_token (lexer, dict) || lex_token (lexer) == T_ALL)
{
- if (lex_token (lexer) == T_ALL || dict_lookup_var (dict, lex_tokid (lexer)) != NULL)
+ if (lex_token (lexer) == T_ALL || dict_lookup_var (dict, lex_tokcstr (lexer)) != NULL)
{
struct variable **v;
size_t nv;
- if (!parse_variables (lexer, dict, &v, &nv, PV_NONE))
+ if (!parse_variables (lexer, dict, &v, &nv, pv_opts))
goto fail;
*names = xnrealloc (*names, *nnames + nv, sizeof **names);
for (i = 0; i < nv; i++)
free (v);
*nnames += nv;
}
- else if (!parse_DATA_LIST_vars (lexer, names, nnames, PV_APPEND))
+ else if (!parse_DATA_LIST_vars (lexer, dict, names, nnames, PV_APPEND | pv_opts))
goto fail;
}
- return 1;
+ if (*nnames == 0)
+ goto fail;
+
+ return true;
fail:
for (i = 0; i < *nnames; i++)
free (*names);
*names = NULL;
*nnames = 0;
- return 0;
+ return false;
}
/* Parses a list of variables where some of the variables may be
/* A set of variables. */
struct var_set
{
+ bool names_must_be_ids;
size_t (*get_cnt) (const struct var_set *);
struct variable *(*get_var) (const struct var_set *, size_t idx);
bool (*lookup_var_idx) (const struct var_set *, const char *, size_t *);
{
assert (vs != NULL);
assert (name != NULL);
- assert (strlen (name) <= VAR_NAME_LEN);
return vs->lookup_var_idx (vs, name, idx);
}
if (vs != NULL)
vs->destroy (vs);
}
+
+static bool
+var_set_get_names_must_be_ids (const struct var_set *vs)
+{
+ return vs->names_must_be_ids;
+}
\f
/* Returns the number of variables in VS. */
static size_t
var_set_create_from_dict (const struct dictionary *d)
{
struct var_set *vs = xmalloc (sizeof *vs);
+ vs->names_must_be_ids = dict_get_names_must_be_ids (d);
vs->get_cnt = dict_var_set_get_cnt;
vs->get_var = dict_var_set_get_var;
vs->lookup_var_idx = dict_var_set_lookup_var_idx;
struct hmapx_node *node;
struct variable **varp;
- HMAPX_FOR_EACH_WITH_HASH (varp, node, hash_case_string (name, 0),
+ HMAPX_FOR_EACH_WITH_HASH (varp, node, utf8_hash_case_string (name, 0),
&avs->vars_by_name)
- if (!strcasecmp (name, var_get_name (*varp)))
+ if (!utf8_strcasecmp (name, var_get_name (*varp)))
{
*idx = varp - avs->var;
return true;
size_t i;
vs = xmalloc (sizeof *vs);
+ vs->names_must_be_ids = true;
vs->get_cnt = array_var_set_get_cnt;
vs->get_var = array_var_set_get_var;
vs->lookup_var_idx = array_var_set_lookup_var_idx;
return NULL;
}
hmapx_insert (&avs->vars_by_name, CONST_CAST (void *, &avs->var[i]),
- hash_case_string (name, 0));
+ utf8_hash_case_string (name, 0));
}
return vs;
}
+
+/* Match a variable.
+ If the match succeeds, the variable will be placed in VAR.
+ Returns true if successful */
+bool
+lex_match_variable (struct lexer *lexer, const struct dictionary *dict, const struct variable **var)
+{
+ if (lex_token (lexer) != T_ID)
+ return false;
+
+ *var = parse_variable_const (lexer, dict);
+
+ if ( *var == NULL)
+ return false;
+ return true;
+}
+
+/* An interaction is a variable followed by {*, BY} followed by an interaction */
+static bool
+parse_internal_interaction (struct lexer *lexer, const struct dictionary *dict, struct interaction **iact, struct interaction **it)
+{
+ const struct variable *v = NULL;
+ assert (iact);
+
+ switch (lex_next_token (lexer, 1))
+ {
+ case T_ENDCMD:
+ case T_SLASH:
+ case T_COMMA:
+ case T_ID:
+ case T_BY:
+ case T_ASTERISK:
+ break;
+ default:
+ return false;
+ break;
+ }
+
+ if (! lex_match_variable (lexer, dict, &v))
+ {
+ if (it)
+ interaction_destroy (*it);
+ *iact = NULL;
+ return false;
+ }
+
+ assert (v);
+
+ if ( *iact == NULL)
+ *iact = interaction_create (v);
+ else
+ interaction_add_variable (*iact, v);
+
+ if ( lex_match (lexer, T_ASTERISK) || lex_match (lexer, T_BY))
+ {
+ return parse_internal_interaction (lexer, dict, iact, iact);
+ }
+
+ return true;
+}
+
+bool
+parse_design_interaction (struct lexer *lexer, const struct dictionary *dict, struct interaction **iact)
+{
+ return parse_internal_interaction (lexer, dict, iact, NULL);
+}
+
projects / pspp / blobdiff