/* This file is part of the Zebra server.
   Copyright (C) Index Data

Zebra is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2, or (at your option) any later
version.

Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA

*/

#if HAVE_CONFIG_H
#include <config.h>
#endif
#include <stdio.h>
#include <assert.h>
#include <stdlib.h>
#include <string.h>

#include <yaz/log.h>
#include <yaz/oid_db.h>
#include <yaz/snprintf.h>
#include <idzebra/data1.h>
#include <idzebra/recctrl.h>
#include <zebra_xpath.h>
#include <d1_absyn.h>

#define D1_MAX_NESTING  128

struct data1_hash_table {
    NMEM nmem;
    int size;
    struct data1_hash_entry **ar;
};

struct data1_hash_entry {
    void *clientData;
    char *str;
    struct data1_hash_entry *next;
};

unsigned data1_hash_calc(struct data1_hash_table *ht, const char *str)
{
    unsigned v = 0;
    assert(str);
    while (*str)
    {
        if (*str >= 'a' && *str <= 'z')
            v = v*65509 + *str -'a'+10;
        else if (*str >= 'A' && *str <= 'Z')
            v = v*65509 + *str -'A'+10;
        else if (*str >= '0' && *str <= '9')
            v = v*65509 + *str -'0';
        str++;
    }
    return v % ht->size;
}

struct data1_hash_table *data1_hash_open(int size, NMEM nmem)
{
    int i;
    struct data1_hash_table *ht = nmem_malloc(nmem, sizeof(*ht));
    ht->nmem = nmem;
    ht->size = size;
    if (ht->size <= 0)
        ht->size = 29;
    ht->ar = nmem_malloc(nmem, sizeof(*ht->ar) * ht->size);
    for (i = 0; i<ht->size; i++)
        ht->ar[i] = 0;
    return ht;
}

void data1_hash_insert(struct data1_hash_table *ht, const char *str,
                       void *clientData, int copy)
{
    char *dstr = copy ? nmem_strdup(ht->nmem, str) : (char*) str;
    if (strchr(str, '?') || strchr(str, '.'))
    {
        int i;
        for (i = 0; i<ht->size; i++)
        {
            struct data1_hash_entry **he = &ht->ar[i];
            for (; *he && strcmp(str, (*he)->str); he = &(*he)->next)
                ;
            if (!*he)
            {
                *he = nmem_malloc(ht->nmem, sizeof(**he));
                (*he)->str = dstr;
                (*he)->next = 0;
            }
            (*he)->clientData = clientData;
        }
    }
    else
    {
        struct data1_hash_entry **he = &ht->ar[data1_hash_calc(ht, str)];
        for (; *he && strcmp(str, (*he)->str); he = &(*he)->next)
            ;
        if (!*he)
        {
            *he = nmem_malloc(ht->nmem, sizeof(**he));
            (*he)->str = dstr;
            (*he)->next = 0;
        }
        (*he)->clientData = clientData;
    }
}

void *data1_hash_lookup(struct data1_hash_table *ht, const char *str)
{
    struct data1_hash_entry **he = &ht->ar[data1_hash_calc(ht, str)];

    for (; *he && yaz_matchstr(str, (*he)->str); he = &(*he)->next)
        ;
    if (*he)
        return (*he)->clientData;
    return 0;
}

struct data1_systag {
    char *name;
    char *value;
    struct data1_systag *next;
};

struct data1_absyn_cache_info
{
    char *name;
    data1_absyn *absyn;
    data1_absyn_cache next;
};

struct data1_attset_cache_info
{
    char *name;
    data1_attset *attset;
    data1_attset_cache next;
};

data1_element *data1_mk_element(data1_handle dh)
{
    data1_element *e = nmem_malloc(data1_nmem_get(dh), sizeof(*e));
    e->name = 0;
    e->tag = 0;
    e->termlists = 0;
    e->next = e->children = 0;
    e->sub_name = 0;
    e->hash = 0;
    return e;
}

data1_absyn *data1_absyn_search(data1_handle dh, const char *name)
{
    data1_absyn_cache p = *data1_absyn_cache_get(dh);

    while (p)
    {
        if (!yaz_matchstr(name, p->name))
            return p->absyn;
        p = p->next;
    }
    return 0;
}
/* *ostrich*
   We need to destroy DFAs, in xp_element (xelm) definitions
   pop, 2002-12-13
*/

void data1_absyn_destroy(data1_handle dh)
{
    data1_absyn_cache p = *data1_absyn_cache_get(dh);

    while (p)
    {
        data1_absyn *abs = p->absyn;
        if (abs)
        {
            data1_xpelement *xpe = abs->xp_elements;
            while (xpe) {
                yaz_log (YLOG_DEBUG,"Destroy xp element %s",xpe->xpath_expr);
                if (xpe->dfa)
                    dfa_delete (&xpe->dfa);
                xpe = xpe->next;
            }
        }
        p = p->next;
    }
}


void data1_absyn_trav(data1_handle dh, void *handle,
                      void (*fh)(data1_handle dh, void *h, data1_absyn *a))
{
    data1_absyn_cache p = *data1_absyn_cache_get(dh);

    while (p)
    {
        (*fh)(dh, handle, p->absyn);
        p = p->next;
    }
}

static data1_absyn *data1_read_absyn(data1_handle dh, const char *file,
                                     enum DATA1_XPATH_INDEXING en);

static data1_absyn *data1_absyn_add(data1_handle dh, const char *name,
                                    enum DATA1_XPATH_INDEXING en)
{
    char fname[512];
    NMEM mem = data1_nmem_get(dh);

    data1_absyn_cache p = (data1_absyn_cache)nmem_malloc(mem, sizeof(*p));
    data1_absyn_cache *pp = data1_absyn_cache_get(dh);

    yaz_snprintf(fname, sizeof(fname), "%s.abs", name);
    p->absyn = data1_read_absyn(dh, fname, en);
    p->name = nmem_strdup(mem, name);
    p->next = *pp;
    *pp = p;
    return p->absyn;
}

data1_absyn *data1_get_absyn(data1_handle dh, const char *name,
                             enum DATA1_XPATH_INDEXING en)
{
    data1_absyn *absyn;

    if (!(absyn = data1_absyn_search(dh, name)))
        absyn = data1_absyn_add(dh, name, en);
    return absyn;
}

data1_attset *data1_attset_search_name(data1_handle dh, const char *name)
{
    data1_attset_cache p = *data1_attset_cache_get(dh);

    while (p)
    {
        if (!yaz_matchstr(name, p->name))
            return p->attset;
        p = p->next;
    }
    return 0;
}

data1_attset *data1_attset_search_id(data1_handle dh, const Odr_oid *oid)
{
    data1_attset_cache p = *data1_attset_cache_get(dh);

    while (p)
    {
        if (p->attset->oid && !oid_oidcmp(oid, p->attset->oid))
            return p->attset;
        p = p->next;
    }
    return 0;
}

data1_attset *data1_attset_add(data1_handle dh, const char *name)
{
    NMEM mem = data1_nmem_get(dh);
    data1_attset *attset;

    attset = data1_read_attset(dh, name);
    if (!attset)
        yaz_log(YLOG_WARN|YLOG_ERRNO, "Couldn't load attribute set %s", name);
    else
    {
        data1_attset_cache p = (data1_attset_cache)
            nmem_malloc (mem, sizeof(*p));
        data1_attset_cache *pp = data1_attset_cache_get(dh);

        attset->name = p->name = nmem_strdup(mem, name);
        p->attset = attset;
        p->next = *pp;
        *pp = p;
    }
    return attset;
}

data1_attset *data1_get_attset(data1_handle dh, const char *name)
{
    data1_attset *attset;

    if (!(attset = data1_attset_search_name(dh, name)))
        attset = data1_attset_add(dh, name);
    return attset;
}

data1_esetname *data1_getesetbyname(data1_handle dh, data1_absyn *a,
                                    const char *name)
{
    data1_esetname *r;

    for (r = a->esetnames; r; r = r->next)
        if (!data1_matchstr(r->name, name))
            return r;
    return 0;
}

/* we have multiple versions of data1_getelementbyname */
#define DATA1_GETELEMENTBYTAGNAME_VERSION 1

data1_element *data1_getelementbytagname(data1_handle dh, data1_absyn *abs,
                                         data1_element *parent,
                                         const char *tagname)
{
    data1_element *r;
    struct data1_hash_table *ht;

    /* It's now possible to have a data1 tree with no abstract syntax */
    if (abs == 0)
        return 0;

    if (!parent)
        r = abs->main_elements;
    else
        r = parent->children;

#if DATA1_GETELEMENTBYTAGNAME_VERSION==1
    /* using hash search */
    if (!r)
        return 0;

    ht = r->hash;
    if (!ht)
    {
        /* build hash table (the first time) */
        ht = r->hash = data1_hash_open(29, data1_nmem_get(dh));
        for (; r; r = r->next)
        {
            data1_name *n;

            for (n = r->tag->names; n; n = n->next)
                data1_hash_insert(ht, n->name, r, 0);
        }
    }
    return data1_hash_lookup(ht, tagname);
#else
    /* using linear search */
    for (; r; r = r->next)
    {
        data1_name *n;

        for (n = r->tag->names; n; n = n->next)
            if (!data1_matchstr(tagname, n->name))
                return r;
    }
    return 0;
#endif
}

data1_element *data1_getelementbyname(data1_handle dh, data1_absyn *absyn,
                                      const char *name)
{
    data1_element *r;

    /* It's now possible to have a data1 tree with no abstract syntax */
    if (absyn == 0)
        return 0;
    for (r = absyn->main_elements; r; r = r->next)
        if (!data1_matchstr(r->name, name))
            return r;
    return 0;
}


void fix_element_ref(data1_handle dh, data1_absyn *absyn, data1_element *e)
{
    /* It's now possible to have a data1 tree with no abstract syntax */
    if (absyn == 0)
        return;

    for (; e; e = e->next)
    {
        if (!e->sub_name)
        {
            if (e->children)
                fix_element_ref(dh, absyn, e->children);
        }
        else
        {
            data1_sub_elements *sub_e = absyn->sub_elements;
            while (sub_e && strcmp(e->sub_name, sub_e->name))
                sub_e = sub_e->next;
            if (sub_e)
                e->children = sub_e->elements;
            else
                yaz_log(YLOG_WARN, "Unresolved reference to sub-elements %s",
                        e->sub_name);
        }
    }
}
/* *ostrich*

   New function, a bit dummy now... I've seen it in zrpn.c... We should build
   more clever regexps...


      //a    ->    ^a/.*$
      //a/b  ->    ^b/a/.*$
      /a     ->    ^a/$
      /a/b   ->    ^b/a/$

      /      ->    none

   pop, 2002-12-13

   Now [] predicates are supported

   pop, 2003-01-17

 */

static const char * mk_xpath_regexp(data1_handle dh, const char *expr)
{
    const char *p = expr;
    int abs = 1;
    int e = 0;
    char *stack[32];
    char *res_p, *res = 0;
    size_t res_size = 1;

    if (*p != '/')
        return ("");
    p++;
    if (*p == '/')
    {
        abs = 0;
        p++;
    }
    while (*p && e < 30)
    {
        int is_predicate = 0;
        char *s;
        int i, j;
        for (i = 0; *p && !strchr("/",*p); i++, p++)
            ;
        res_size += (i+3); /* we'll add / between later .. */
        s = stack[e] = (char *) nmem_malloc(data1_nmem_get(dh), i + 1);
        for (j = 0; j < i; j++)
        {
            const char *pp = p-i+j;
            if (*pp == '[')
                is_predicate = 1;
            else if (*pp == ']')
                is_predicate = 0;
            else
            {
                if (!is_predicate) {
                    if (*pp == '*')
                        *s++ = '.';
                    *s++ = *pp;
                }
            }
        }
        *s = 0;
        e++;
        if (*p)
            p++;
    }
    res_p = res = nmem_malloc(data1_nmem_get(dh), res_size + 10);

    if (stack[e-1][0] == '@')  /* path/@attr spec (leaf is attribute) */
        strcpy(res_p, "/");
    else
        strcpy(res_p, "[^@]*/");  /* path .. (index all cdata below it) */
    res_p = res_p + strlen(res_p);
    while (--e >= 0) {
        strcpy(res_p, stack[e]);
        strcat(res_p, "/");
        res_p += strlen(stack[e]) + 1;
    }
    if (!abs)
    {
        strcpy(res_p, ".*");
        res_p += 2;
    }
    strcpy(res_p, "$");
    res_p++;
    yaz_log(YLOG_DEBUG, "Got regexp: %s", res);
    return res;
}

static int parse_termlists(data1_handle dh, data1_termlist ***tpp,
                           char *cp, const char *file, int lineno,
                           const char *element_name, data1_absyn *res,
                           int xpelement,
                           data1_attset *attset)
{
    data1_termlist **tp = *tpp;
    while(1)
    {
        char attname[512], structure[512];
        char *source;
        int r, i;
        int level = 0;
        structure[0] = '\0';
        for (i = 0; cp[i] && i<sizeof(attname)-1; i++)
            if (strchr(":,", cp[i]))
                break;
            else
                attname[i] = cp[i];
        if (i == 0)
        {
            if (*cp)
                yaz_log(YLOG_WARN,
                        "%s:%d: Syntax error in termlistspec '%s'",
                        file, lineno, cp);
            break;
        }
        attname[i] = '\0';
        r = 1;
        cp += i;
        if (*cp == ':')
            cp++;

        for (i = 0; cp[i] && i<sizeof(structure)-1; i++)
            if (level == 0 && strchr(",", cp[i]))
                break;
            else
            {
                structure[i] = cp[i];
                if (cp[i] == '(')
                    level++;
                else if (cp[i] == ')')
                    level--;
            }
        structure[i] = '\0';
        if (i)
            r = 2;
        cp += i;
        if (*cp)
            cp++;  /* skip , */

        *tp = (data1_termlist *)
            nmem_malloc(data1_nmem_get(dh), sizeof(**tp));
        (*tp)->next = 0;

        if (*attname == '!')
        {
            if (!xpelement && element_name)
                strcpy(attname, element_name);
            else if (xpelement)
                strcpy(attname, ZEBRA_XPATH_CDATA);
        }
        if (attset)
        {
            if (!data1_getattbyname(dh, attset, attname))
            {
                yaz_log(YLOG_WARN, "Index '%s' not found in attset(s)",
                        attname);
            }
        }

        (*tp)->index_name = nmem_strdup(data1_nmem_get(dh), attname);
        assert (*(*tp)->index_name != '!');
        if (r == 2 && (source = strchr(structure, ':')))
            *source++ = '\0';   /* cut off structure .. */
        else
            source = "data";    /* ok: default is leaf data */
        (*tp)->source = nmem_strdup(data1_nmem_get (dh), source);

        if (r < 2) /* is the structure qualified? */
            (*tp)->structure = "w";
        else
            (*tp)->structure = nmem_strdup(data1_nmem_get (dh), structure);
        tp = &(*tp)->next;
    }

    *tpp = tp;
    return 0;
}

/* quinn
 * Converts a 'melm' field[$subfield] pattern to a simple xpath
 */
static int melm2xpath(char *melm, char *buf)
{
    char *dollar;
    char *field = melm;
    char *subfield;
    char *fieldtype;
    if ((dollar = strchr(melm, '$'))) {
        *dollar = '\0';
        subfield = ++dollar;
    } else
        subfield = "";
    if (field[0] == '0' && field[1] == '0')
        fieldtype = "controlfield";
    else
        fieldtype = "datafield";
    yaz_snprintf(buf, 60, "/*/%s[@tag=\"%s\"]", fieldtype, field);
    if (*subfield)
        yaz_snprintf(buf + strlen(buf), 60, "/subfield[@code=\"%s\"]", subfield);
    else if (field[0] != '0' || field[1] != '0')
        strcat(buf, "/subfield");
    yaz_log(YLOG_DEBUG, "Created xpath: '%s'", buf);
    return 0;
}

const char *data1_systag_lookup(data1_absyn *absyn, const char *tag,
                                const char *default_value)
{
    struct data1_systag *p = absyn->systags;
    for (; p; p = p->next)
        if (!strcmp(p->name, tag))
            return p->value;
    return default_value;
}

#define l_isspace(c) ((c) == '\t' || (c) == ' ' || (c) == '\n' || (c) == '\r')

int read_absyn_line(FILE *f, int *lineno, char *line, int len,
                    char *argv[], int num)
{
    char *p;
    int argc;
    int quoted = 0;

    while ((p = fgets(line, len, f)))
    {
        (*lineno)++;
        while (*p && l_isspace(*p))
            p++;
        if (*p && *p != '#')
            break;
    }
    if (!p)
        return 0;

    for (argc = 0; *p ; argc++)
    {
        if (*p == '#')  /* trailing comment */
            break;
        argv[argc] = p;
        while (*p && !(l_isspace(*p) && !quoted)) {
          if (*p =='"') quoted = 1 - quoted;
          if (*p =='[') quoted = 1;
          if (*p ==']') quoted = 0;
          p++;
        }
        if (*p)
        {
            *(p++) = '\0';
            while (*p && l_isspace(*p))
                p++;
        }
    }
    return argc;
}

data1_marctab *data1_absyn_getmarctab(data1_handle dh, data1_node *root)
{
    if (root->u.root.absyn)
        return root->u.root.absyn->marc;
    return 0;
}

data1_element *data1_absyn_getelements(data1_handle dh,
                                       data1_node *root)
{
    if (root->u.root.absyn)
        return root->u.root.absyn->main_elements;
    return 0;
}

static data1_absyn *data1_read_absyn(data1_handle dh, const char *file,
                                     enum DATA1_XPATH_INDEXING default_xpath)
{
    data1_sub_elements *cur_elements = NULL;
    data1_xpelement **cur_xpelement = NULL;
    data1_attset *attset_list = data1_empty_attset(dh);
    data1_attset_child **attset_childp = &attset_list->children;

    data1_absyn *res = 0;
    FILE *f;
    data1_element **ppl[D1_MAX_NESTING];
    data1_esetname **esetpp;
    data1_maptab **maptabp;
    data1_marctab **marcp;
    data1_termlist *all = 0;
    data1_tagset **tagset_childp;
    struct data1_systag **systagsp;
    int level = 0;
    int lineno = 0;
    int argc;
    char *argv[50], line[512];

    f = data1_path_fopen(dh, file, "r");

    res = (data1_absyn *) nmem_malloc(data1_nmem_get(dh), sizeof(*res));
    res->name = 0;
    res->oid = 0;
    res->tagset = 0;
    res->encoding = 0;
    res->xpath_indexing =
        (f ? DATA1_XPATH_INDEXING_DISABLE : default_xpath);
    res->systags = 0;
    systagsp = &res->systags;
    tagset_childp = &res->tagset;

    res->varset = 0;
    res->esetnames = 0;
    esetpp = &res->esetnames;
    res->maptabs = 0;
    maptabp = &res->maptabs;
    res->marc = 0;
    marcp = &res->marc;
    res->sub_elements = NULL;
    res->main_elements = NULL;
    res->xp_elements = NULL;
    cur_xpelement = &res->xp_elements;

    while (f && (argc = read_absyn_line(f, &lineno, line, 512, argv, 50)))
    {
        char *cmd = *argv;
        if (!strcmp(cmd, "elm") || !strcmp(cmd, "element"))
        {
            data1_element *new_element;
            int i;
            char *p, *sub_p, *path, *name, *termlists;
            int type, value;
            data1_termlist **tp;

            if (argc < 4)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to elm", file, lineno);
                continue;
            }
            path = argv[1];
            name = argv[2];
            termlists = argv[3];

            if (!cur_elements)
            {
                cur_elements = (data1_sub_elements *)
                    nmem_malloc(data1_nmem_get(dh), sizeof(*cur_elements));
                cur_elements->next = res->sub_elements;
                cur_elements->elements = NULL;
                cur_elements->name = "main";
                res->sub_elements = cur_elements;

                level = 0;
                ppl[level] = &cur_elements->elements;
            }
            p = path;
            for (i = 1;; i++)
            {
                char *e;

                if ((e = strchr(p, '/')))
                    p = e+1;
                else
                    break;
            }
            if (i > level+1)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad level increase", file, lineno);
                fclose(f);
                return 0;
            }
            level = i;
            new_element = *ppl[level-1] = data1_mk_element(dh);

            tp = &new_element->termlists;
            ppl[level-1] = &new_element->next;
            ppl[level] = &new_element->children;

            /* consider subtree (if any) ... */
            if ((sub_p = strchr (p, ':')) && sub_p[1])
            {
                *sub_p++ = '\0';
                new_element->sub_name =
                    nmem_strdup (data1_nmem_get(dh), sub_p);
            }
            /* well-defined tag */
            if (sscanf(p, "(%d,%d)", &type, &value) == 2)
            {
                if (!res->tagset)
                {
                    yaz_log(YLOG_WARN, "%s:%d: No tagset loaded", file, lineno);
                    fclose(f);
                    return 0;
                }
                if (!(new_element->tag = data1_gettagbynum(dh, res->tagset,
                                                           type, value)))
                {
                    yaz_log(YLOG_WARN, "%s:%d: Couldn't find tag %s in tagset",
                         file, lineno, p);
                    fclose(f);
                    return 0;
                }
            }
            /* private tag */
            else if (*p)
            {
                data1_tag *nt =
                    new_element->tag = (data1_tag *)
                    nmem_malloc(data1_nmem_get (dh),
                                sizeof(*new_element->tag));
                nt->which = DATA1T_string;
                nt->value.string = nmem_strdup(data1_nmem_get(dh), p);
                nt->names = (data1_name *)
                    nmem_malloc(data1_nmem_get(dh),
                                sizeof(*new_element->tag->names));
                nt->names->name = nt->value.string;
                nt->names->next = 0;
                nt->kind = DATA1K_string;
                nt->next = 0;
                nt->tagset = 0;
            }
            else
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad element", file, lineno);
                fclose(f);
                return 0;
            }
            /* parse termList definitions */
            p = termlists;
            if (*p != '-')
            {
                if (parse_termlists(dh, &tp, p, file, lineno, name, res, 0,
                                    attset_list))
                {
                    fclose (f);
                    return 0;
                }
                *tp = all; /* append any ALL entries to the list */
            }
            new_element->name = nmem_strdup(data1_nmem_get(dh), name);
        }
        /* *ostrich*
           New code to support xelm directive
           for each xelm a dfa is built. xelms are stored in res->xp_elements

           maybe we should use a simple sscanf instead of dfa?

           pop, 2002-12-13

           Now [] predicates are supported. regexps and xpath structure is
           a bit redundant, however it's comfortable later...

           pop, 2003-01-17
        */

        else if (!strcmp(cmd, "xelm") || !strcmp(cmd, "melm")) {

            int i;
            char *p, *xpath_expr, *termlists;
            const char *regexp;
            struct DFA *dfa = 0;
            data1_termlist **tp;
            char melm_xpath[128];
            data1_xpelement *xp_ele = 0;
            data1_xpelement *last_match = 0;

            if (argc != 3)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to %s",
                        file, lineno, cmd);
                continue;
            }

            if (!strcmp(cmd, "melm")) {
                if (melm2xpath(argv[1], melm_xpath) < 0)
                    continue;
                xpath_expr = melm_xpath;
            } else {
                xpath_expr = argv[1];
            }
            termlists = argv[2];
            regexp = mk_xpath_regexp(dh, xpath_expr);

#if OPTIMIZE_MELM
            /* get last of existing regulars with same regexp */
            for (xp_ele = res->xp_elements; xp_ele; xp_ele = xp_ele->next)
                if (!strcmp(xp_ele->regexp, regexp))
                    last_match = xp_ele;
#endif
            if (!last_match)
            {
                /* new regular expression . Parse + generate */
                const char *regexp_ptr = regexp;

                dfa = dfa_init();
                i = dfa_parse (dfa, &regexp_ptr);
                if (i || *regexp_ptr) {
                    yaz_log(YLOG_WARN, "%s:%d: Bad xpath to xelm", file, lineno);
                    dfa_delete (&dfa);
                    continue;
                }
            }
            *cur_xpelement = (data1_xpelement *)
                nmem_malloc(data1_nmem_get(dh), sizeof(**cur_xpelement));
            (*cur_xpelement)->next = 0;
            (*cur_xpelement)->match_next = 0;
            if (last_match)
                last_match->match_next = *cur_xpelement;
#if OPTIMIZE_MELM
            (*cur_xpelement)->regexp = regexp;
#endif
            (*cur_xpelement)->xpath_expr = nmem_strdup(data1_nmem_get (dh),
                                                       xpath_expr);

            if (dfa)
                dfa_mkstate (dfa);
            (*cur_xpelement)->dfa = dfa;

#ifdef ENHANCED_XELM
            (*cur_xpelement)->xpath_len =
                zebra_parse_xpath_str(
                    xpath_expr,
                    (*cur_xpelement)->xpath, XPATH_STEP_COUNT,
                    data1_nmem_get(dh));
#endif
            (*cur_xpelement)->termlists = 0;
            tp = &(*cur_xpelement)->termlists;

            /* parse termList definitions */
            p = termlists;
            if (*p != '-')
            {
                if (parse_termlists(dh, &tp, p, file, lineno,
                                    xpath_expr, res, 1, attset_list))
                {
                    fclose (f);
                    return 0;
                }
                *tp = all; /* append any ALL entries to the list */
            }
            cur_xpelement = &(*cur_xpelement)->next;
        }
        else if (!strcmp(cmd, "section"))
        {
            char *name;

            if (argc < 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to section",
                        file, lineno);
                continue;
            }
            name = argv[1];

            cur_elements = (data1_sub_elements *)
                nmem_malloc(data1_nmem_get(dh), sizeof(*cur_elements));
            cur_elements->next = res->sub_elements;
            cur_elements->elements = NULL;
            cur_elements->name = nmem_strdup(data1_nmem_get(dh), name);
            res->sub_elements = cur_elements;

            level = 0;
            ppl[level] = &cur_elements->elements;
        }
        else if (!strcmp(cmd, "xpath"))
        {
            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to 'xpath' directive",
                     file, lineno);
                continue;
            }
            if (!strcmp(argv[1], "enable"))
                res->xpath_indexing = DATA1_XPATH_INDEXING_ENABLE;
            else if (!strcmp (argv[1], "disable"))
                res->xpath_indexing = DATA1_XPATH_INDEXING_DISABLE;
            else
            {
                yaz_log(YLOG_WARN, "%s:%d: Expecting disable/enable "
                        "after 'xpath' directive", file, lineno);
            }
        }
        else if (!strcmp(cmd, "all"))
        {
            data1_termlist **tp = &all;
            if (all)
            {
                yaz_log(YLOG_WARN, "%s:%d: Too many 'all' directives - ignored",
                     file, lineno);
                continue;
            }
            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to 'all' directive",
                     file, lineno);
                continue;
            }
            if (parse_termlists(dh, &tp, argv[1], file, lineno, 0, res, 0,
                                attset_list))
            {
                fclose (f);
                return 0;
            }
        }
        else if (!strcmp(cmd, "name"))
        {
            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to name directive",
                     file, lineno);
                continue;
            }
            res->name = nmem_strdup(data1_nmem_get(dh), argv[1]);
        }
        else if (!strcmp(cmd, "reference"))
        {
            char *name;

            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to reference",
                     file, lineno);
                continue;
            }
            name = argv[1];
            res->oid = yaz_string_to_oid_nmem(yaz_oid_std(),
                                              CLASS_SCHEMA, name,
                                              data1_nmem_get(dh));
            if (!res->oid)
            {
                yaz_log(YLOG_WARN, "%s:%d: Unknown tagset ref '%s'",
                     file, lineno, name);
                continue;
            }
        }
        else if (!strcmp(cmd, "attset"))
        {
           char *name;
           data1_attset *attset;

           if (argc != 2)
           {
               yaz_log(YLOG_WARN, "%s:%d: Bad # of args to attset",
                    file, lineno);
               continue;
           }
           name = argv[1];
           if (!(attset = data1_get_attset(dh, name)))
           {
               yaz_log(YLOG_WARN, "%s:%d: Couldn't find attset  %s",
                       file, lineno, name);
               continue;
           }
           *attset_childp = (data1_attset_child *)
               nmem_malloc(data1_nmem_get(dh), sizeof(**attset_childp));
           (*attset_childp)->child = attset;
           (*attset_childp)->next = 0;
           attset_childp = &(*attset_childp)->next;
        }
        else if (!strcmp(cmd, "tagset"))
        {
            char *name;
            int type = 0;
            if (argc < 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args to tagset",
                     file, lineno);
                continue;
            }
            name = argv[1];
            if (argc == 3)
                type = atoi(argv[2]);
            *tagset_childp = data1_read_tagset(dh, name, type);
            if (!(*tagset_childp))
            {
                yaz_log(YLOG_WARN, "%s:%d: Couldn't load tagset %s",
                     file, lineno, name);
                continue;
            }
            tagset_childp = &(*tagset_childp)->next;
        }
        else if (!strcmp(cmd, "varset"))
        {
            char *name;

            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args in varset",
                     file, lineno);
                continue;
            }
            name = argv[1];
            if (!(res->varset = data1_read_varset(dh, name)))
            {
                yaz_log(YLOG_WARN, "%s:%d: Couldn't load Varset %s",
                     file, lineno, name);
                continue;
            }
        }
        else if (!strcmp(cmd, "esetname"))
        {
            char *name, *fname;

            if (argc != 3)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args in esetname",
                     file, lineno);
                continue;
            }
            name = argv[1];
            fname = argv[2];

            *esetpp = (data1_esetname *)
                nmem_malloc(data1_nmem_get(dh), sizeof(**esetpp));
            (*esetpp)->name = nmem_strdup(data1_nmem_get(dh), name);
            (*esetpp)->next = 0;
            if (*fname == '@')
                (*esetpp)->spec = 0;
            else if (!((*esetpp)->spec = data1_read_espec1(dh, fname)))
            {
                yaz_log(YLOG_WARN, "%s:%d: Espec-1 read failed for %s",
                     file, lineno, fname);
                continue;
            }
            esetpp = &(*esetpp)->next;
        }
        else if (!strcmp(cmd, "maptab"))
        {
            char *name;

            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # of args for maptab",
                     file, lineno);
                continue;
            }
            name = argv[1];
            if (!(*maptabp = data1_read_maptab(dh, name)))
            {
                yaz_log(YLOG_WARN, "%s:%d: Couldn't load maptab %s",
                     file, lineno, name);
                continue;
            }
            maptabp = &(*maptabp)->next;
        }
        else if (!strcmp(cmd, "marc"))
        {
            char *name;

            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # or args for marc",
                     file, lineno);
                continue;
            }
            name = argv[1];
            if (!(*marcp = data1_read_marctab(dh, name)))
            {
                yaz_log(YLOG_WARN, "%s:%d: Couldn't read marctab %s",
                     file, lineno, name);
                continue;
            }
            marcp = &(*marcp)->next;
        }
        else if (!strcmp(cmd, "encoding"))
        {
            if (argc != 2)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # or args for encoding",
                     file, lineno);
                continue;
            }
            res->encoding = nmem_strdup(data1_nmem_get(dh), argv[1]);
        }
        else if (!strcmp(cmd, "systag"))
        {
            if (argc != 3)
            {
                yaz_log(YLOG_WARN, "%s:%d: Bad # or args for systag",
                     file, lineno);
                continue;
            }
            *systagsp = nmem_malloc(data1_nmem_get(dh), sizeof(**systagsp));

            (*systagsp)->name = nmem_strdup(data1_nmem_get(dh), argv[1]);
            (*systagsp)->value = nmem_strdup(data1_nmem_get(dh), argv[2]);
            systagsp = &(*systagsp)->next;
        }
        else
        {
            yaz_log(YLOG_WARN, "%s:%d: Unknown directive '%s'", file,
                    lineno, cmd);
            continue;
        }
    }
    if (f)
        fclose(f);

    for (cur_elements = res->sub_elements; cur_elements;
         cur_elements = cur_elements->next)
    {
        if (!strcmp(cur_elements->name, "main"))
            res->main_elements = cur_elements->elements;
        fix_element_ref(dh, res, cur_elements->elements);
    }
    *systagsp = 0;
    return res;
}

/*
 * Local variables:
 * c-basic-offset: 4
 * c-file-style: "Stroustrup"
 * indent-tabs-mode: nil
 * End:
 * vim: shiftwidth=4 tabstop=8 expandtab
 */