X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=retrieval%2Fd1_read.c;h=130c51853d7dfcf2acef1e66f045b50a3f11ff45;hb=caa50534e057668bf2cded573d41c5b07034633b;hp=289a7514e47ed58eade509961b298c25c9a411e3;hpb=99fef536b882e1a7e8fae9955ff8c28bfe171e9c;p=yaz-moved-to-github.git
diff --git a/retrieval/d1_read.c b/retrieval/d1_read.c
index 289a751..130c518 100644
--- a/retrieval/d1_read.c
+++ b/retrieval/d1_read.c
@@ -1,173 +1,9 @@
/*
- * Copyright (c) 1995-2000, Index Data.
+ * Copyright (c) 1995-2002, Index Data.
* See the file LICENSE for details.
* Sebastian Hammer, Adam Dickmeiss
*
- * $Log: d1_read.c,v $
- * Revision 1.34 2000-12-05 10:06:23 adam
- * Added support for null-data rules like .
- *
- * Revision 1.33 2000/11/29 14:22:47 adam
- * Implemented XML/SGML attributes for data1 so that d1_read reads them
- * and d1_write generates proper attributes for XML/SGML records. Added
- * register locking for threaded version.
- *
- * Revision 1.32 2000/01/06 11:25:59 adam
- * Added case to prevent warning.
- *
- * Revision 1.31 1999/12/21 14:16:20 ian
- * Changed retrieval module to allow data1 trees with no associated absyn.
- * Also added a simple interface for extracting values from data1 trees using
- * a string based tagpath.
- *
- * Revision 1.30 1999/11/30 13:47:12 adam
- * Improved installation. Moved header files to include/yaz.
- *
- * Revision 1.29 1999/10/21 12:06:29 adam
- * Retrieval module no longer uses ctype.h - functions.
- *
- * Revision 1.28 1999/10/21 09:50:33 adam
- * SGML reader uses own isspace - it doesn't do 8-bit on WIN32!
- *
- * Revision 1.27 1999/08/27 09:40:32 adam
- * Renamed logf function to yaz_log. Removed VC++ project files.
- *
- * Revision 1.26 1999/07/13 13:23:48 adam
- * Non-recursive version of data1_read_node. data1_read_nodex reads
- * stream of bytes (instead of buffer in memory).
- *
- * Revision 1.25 1999/04/20 09:56:48 adam
- * Added 'name' paramter to encoder/decoder routines (typedef Odr_fun).
- * Modified all encoders/decoders to reflect this change.
- *
- * Revision 1.24 1998/10/28 15:10:09 adam
- * Added --with-yc option to configure. For the data1_node in data1.h:
- * decreased size of localdata and removed member "line" which wasn't useful.
- *
- * Revision 1.23 1998/03/12 11:28:45 adam
- * Fix: didn't set root member of tagged node in function.
- * data1_add_insert_taggeddata.
- *
- * Revision 1.22 1998/03/05 08:15:32 adam
- * Implemented data1_add_insert_taggeddata utility which is more flexible
- * than data1_insert_taggeddata.
- *
- * Revision 1.21 1998/02/27 14:08:05 adam
- * Added const to some char pointer arguments.
- * Reworked data1_read_node so that it doesn't create a tree with
- * pointers to original "SGML"-buffer.
- *
- * Revision 1.20 1998/02/11 11:53:35 adam
- * Changed code so that it compiles as C++.
- *
- * Revision 1.19 1997/12/09 16:17:09 adam
- * Fix bug regarding variants. Tags with prefix "var" was incorrectly
- * interpreted as "start of variants". Now, only "var" indicates such
- * start.
- * Cleaned up data1_read_node so tag names and variant names are
- * copied and not pointed to by the generated data1 tree. Data nodes
- * still point to old buffer.
- *
- * Revision 1.18 1997/11/18 09:51:09 adam
- * Removed element num_children from data1_node. Minor changes in
- * data1 to Explain.
- *
- * Revision 1.17 1997/11/05 09:20:51 adam
- * Minor change.
- *
- * Revision 1.16 1997/09/17 12:10:37 adam
- * YAZ version 1.4.
- *
- * Revision 1.15 1997/09/05 09:50:57 adam
- * Removed global data1_tabpath - uses data1_get_tabpath() instead.
- *
- * Revision 1.14 1997/05/14 06:54:04 adam
- * C++ support.
- *
- * Revision 1.13 1996/10/29 13:35:38 adam
- * Implemented data1_set_tabpath and data1_get_tabpath.
- *
- * Revision 1.12 1996/10/11 10:35:38 adam
- * Fixed a bug that caused data1_read_node to core dump when no abstract
- * syntax was defined in a "sgml"-record.
- *
- * Revision 1.11 1996/07/06 19:58:35 quinn
- * System headerfiles gathered in yconfig
- *
- * Revision 1.10 1996/01/19 15:41:47 quinn
- * Fixed uninitialized boolean.
- *
- * Revision 1.9 1996/01/17 14:52:47 adam
- * Changed prototype for reader function parsed to data1_read_record.
- *
- * Revision 1.8 1995/12/15 16:20:41 quinn
- * Added formatted text.
- *
- * Revision 1.7 1995/12/13 13:44:32 quinn
- * Modified Data1-system to use nmem
- *
- * Revision 1.6 1995/12/12 16:37:08 quinn
- * Added destroy element to data1_node.
- *
- * Revision 1.5 1995/12/11 15:22:37 quinn
- * Added last_child field to the node.
- * Rewrote schema-mapping.
- *
- * Revision 1.4 1995/11/13 09:27:36 quinn
- * Fiddling with the variant stuff.
- *
- * Revision 1.3 1995/11/01 16:34:57 quinn
- * Making data1 look for tables in data1_tabpath
- *
- * Revision 1.2 1995/11/01 13:54:48 quinn
- * Minor adjustments
- *
- * Revision 1.1 1995/11/01 11:56:09 quinn
- * Added Retrieval (data management) functions en masse.
- *
- * Revision 1.14 1995/10/30 12:40:55 quinn
- * Fixed a couple of bugs.
- *
- * Revision 1.13 1995/10/25 16:00:47 quinn
- * USMARC support is now almost operational
- *
- * Revision 1.12 1995/10/16 14:02:55 quinn
- * Changes to support element set names and espec1
- *
- * Revision 1.11 1995/10/13 16:05:08 quinn
- * Adding Espec1-processing
- *
- * Revision 1.10 1995/10/11 14:53:44 quinn
- * Work on variants.
- *
- * Revision 1.9 1995/10/06 16:56:50 quinn
- * Fixed ranked result.
- *
- * Revision 1.8 1995/10/06 16:44:13 quinn
- * Work on attribute set mapping, etc.
- *
- * Revision 1.7 1995/10/06 12:58:35 quinn
- * SUTRS support
- *
- * Revision 1.6 1995/10/04 09:29:49 quinn
- * Adjustments to support USGS test data
- *
- * Revision 1.5 1995/10/03 17:56:43 quinn
- * Fixing GRS code.
- *
- * Revision 1.4 1995/10/02 15:53:19 quinn
- * Work
- *
- * Revision 1.3 1995/10/02 14:55:21 quinn
- * *** empty log message ***
- *
- * Revision 1.2 1995/09/14 15:18:13 quinn
- * Work
- *
- * Revision 1.1 1995/09/12 11:24:30 quinn
- * Beginning to add code for structured records.
- *
- *
+ * $Id: d1_read.c,v 1.51 2002-08-28 07:54:11 adam Exp $
*/
#include
@@ -175,36 +11,78 @@
#include
#include
+#include
#include
#include
+data1_node *data1_get_root_tag (data1_handle dh, data1_node *n)
+{
+ if (!n)
+ return 0;
+ if (data1_is_xmlmode(dh))
+ {
+ n = n->child;
+ while (n && n->which != DATA1N_tag)
+ n = n->next;
+ }
+ return n;
+}
+
/*
* get the tag which is the immediate parent of this node (this may mean
* traversing intermediate things like variants and stuff.
*/
data1_node *get_parent_tag (data1_handle dh, data1_node *n)
{
- for (; n && n->which != DATA1N_root; n = n->parent)
- if (n->which == DATA1N_tag)
- return n;
+ if (data1_is_xmlmode(dh))
+ {
+ for (; n && n->which != DATA1N_root; n = n->parent)
+ if (n->which == DATA1N_tag && n->parent &&
+ n->parent->which != DATA1N_root)
+ return n;
+ }
+ else
+ {
+ for (; n && n->which != DATA1N_root; n = n->parent)
+ if (n->which == DATA1N_tag)
+ return n;
+ }
return 0;
}
data1_node *data1_mk_node (data1_handle dh, NMEM m)
{
- data1_node *r;
-
- r = (data1_node *)nmem_malloc(m, sizeof(*r));
- r->next = r->child = r->last_child = r->parent = 0;
- r->destroy = 0;
- return r;
+ return data1_mk_node2 (dh, m, DATA1N_root, 0);
}
data1_node *data1_mk_node_type (data1_handle dh, NMEM m, int type)
{
+ return data1_mk_node2 (dh, m, type, 0);
+}
+
+data1_node *data1_mk_node2 (data1_handle dh, NMEM m, int type,
+ data1_node *parent)
+{
data1_node *r;
+
+ r = (data1_node *)nmem_malloc(m, sizeof(*r));
+ r->next = r->child = r->last_child = 0;
+ r->destroy = 0;
- r = data1_mk_node(dh, m);
+ if (!parent)
+ {
+ r->root = r;
+ }
+ else
+ {
+ r->root = parent->root;
+ r->parent = parent;
+ if (!parent->child)
+ parent->child = parent->last_child = r;
+ else
+ parent->last_child->next = r;
+ parent->last_child = r;
+ }
r->which = type;
switch(type)
{
@@ -215,9 +93,7 @@ data1_node *data1_mk_node_type (data1_handle dh, NMEM m, int type)
r->u.tag.node_selected = 0;
r->u.tag.make_variantlist = 0;
r->u.tag.get_bytes = -1;
-#if DATA1_USING_XATTR
r->u.tag.attributes = 0;
-#endif
break;
case DATA1N_root:
r->u.root.type = 0;
@@ -228,7 +104,21 @@ data1_node *data1_mk_node_type (data1_handle dh, NMEM m, int type)
r->u.data.len = 0;
r->u.data.what = 0;
r->u.data.formatted_text = 0;
+ break;
+ case DATA1N_comment:
+ r->u.data.data = 0;
+ r->u.data.len = 0;
+ r->u.data.what = 0;
+ r->u.data.formatted_text = 1;
+ break;
+ case DATA1N_variant:
+ r->u.variant.type = 0;
+ r->u.variant.value = 0;
break;
+ case DATA1N_preprocess:
+ r->u.preprocess.target = 0;
+ r->u.preprocess.attributes = 0;
+ break;
default:
logf (LOG_WARN, "data_mk_node_type. bad type = %d\n", type);
}
@@ -249,71 +139,240 @@ void data1_free_tree (data1_handle dh, data1_node *t)
(*t->destroy)(t);
}
-char *data1_insert_string (data1_handle dh, data1_node *res,
- NMEM m, const char *str)
+data1_node *data1_mk_root (data1_handle dh, NMEM nmem, const char *name)
+{
+ data1_absyn *absyn = data1_get_absyn (dh, name);
+ data1_node *res;
+ if (!absyn)
+ {
+ yaz_log(LOG_WARN, "Unable to acquire abstract syntax " "for '%s'",
+ name);
+ /* It's now OK for a record not to have an absyn */
+ }
+ res = data1_mk_node2 (dh, nmem, DATA1N_root, 0);
+ res->u.root.type = data1_insert_string (dh, res, nmem, name);
+ res->u.root.absyn = absyn;
+ return res;
+}
+
+void data1_set_root(data1_handle dh, data1_node *res,
+ NMEM nmem, const char *name)
{
- int len = strlen(str);
+ data1_absyn *absyn = data1_get_absyn (dh, name);
- if (len >= DATA1_LOCALDATA)
- return nmem_strdup (m, str);
+ res->u.root.type = data1_insert_string (dh, res, nmem, name);
+ res->u.root.absyn = absyn;
+}
+
+data1_node *data1_mk_preprocess (data1_handle dh, NMEM nmem,
+ const char *target, const char **attr,
+ data1_node *at)
+{
+ data1_xattr **p;
+ data1_node *res = data1_mk_node2 (dh, nmem, DATA1N_preprocess, at);
+ res->u.preprocess.target = data1_insert_string (dh, res, nmem, target);
+
+ p = &res->u.preprocess.attributes;
+ while (attr && *attr)
+ {
+ *p = (data1_xattr*) nmem_malloc (nmem, sizeof(**p));
+ (*p)->name = nmem_strdup (nmem, *attr++);
+ (*p)->value = nmem_strdup (nmem, *attr++);
+ p = &(*p)->next;
+ }
+ *p = 0;
+ return res;
+}
+
+data1_node *data1_mk_tag_n (data1_handle dh, NMEM nmem,
+ const char *tag, size_t len, const char **attr,
+ data1_node *at)
+{
+ data1_node *partag = get_parent_tag(dh, at);
+ data1_node *res = data1_mk_node2 (dh, nmem, DATA1N_tag, at);
+ data1_xattr **p;
+ data1_element *e = 0;
+
+ res->u.tag.tag = data1_insert_string_n (dh, res, nmem, tag, len);
+
+ if (!partag) /* top tag? */
+ e = data1_getelementbytagname (dh, at->root->u.root.absyn,
+ 0 /* index as local */,
+ res->u.tag.tag);
else
{
- strcpy (res->lbuf, str);
- return res->lbuf;
+ /* only set element for known tags */
+ e = partag->u.tag.element;
+ if (e)
+ e = data1_getelementbytagname (dh, at->root->u.root.absyn,
+ e, res->u.tag.tag);
+ }
+ res->u.tag.element = e;
+ p = &res->u.tag.attributes;
+ while (attr && *attr)
+ {
+ *p = (data1_xattr*) nmem_malloc (nmem, sizeof(**p));
+ (*p)->name = nmem_strdup (nmem, *attr++);
+ (*p)->value = nmem_strdup (nmem, *attr++);
+ p = &(*p)->next;
}
+ *p = 0;
+ return res;
}
-data1_node *data1_add_insert_taggeddata(data1_handle dh, data1_node *root,
- data1_node *at,
- const char *tagname, NMEM m,
- int first_flag, int local_allowed)
+void data1_tag_add_attr (data1_handle dh, NMEM nmem,
+ data1_node *res, const char **attr)
{
- data1_node *partag = get_parent_tag (dh, at);
- data1_node *tagn = data1_mk_node_type (dh, m, DATA1N_tag);
- data1_element *e = NULL;
- data1_node *datn;
+ data1_xattr **p;
- tagn->u.tag.tag = data1_insert_string (dh, tagn, m, tagname);
+ if (res->which != DATA1N_tag)
+ return;
- if (partag)
- e = partag->u.tag.element;
- tagn->u.tag.element =
- data1_getelementbytagname (dh, root->u.root.absyn, e, tagname);
- if (!local_allowed && !tagn->u.tag.element)
- return NULL;
- tagn->last_child = tagn->child = datn = data1_mk_node_type (dh, m, DATA1N_data);
- tagn->root = root;
- datn->parent = tagn;
- datn->root = root;
- datn->u.data.formatted_text = 0;
- tagn->parent = at;
-
- if (first_flag)
+ p = &res->u.tag.attributes;
+ while (*p)
+ p = &(*p)->next;
+
+ while (attr && *attr)
{
- tagn->next = at->child;
- if (!tagn->next)
- at->last_child = tagn;
- at->child = tagn;
+ *p = (data1_xattr*) nmem_malloc (nmem, sizeof(**p));
+ (*p)->name = nmem_strdup (nmem, *attr++);
+ (*p)->value = nmem_strdup (nmem, *attr++);
+ p = &(*p)->next;
}
- else
+ *p = 0;
+}
+
+data1_node *data1_mk_tag (data1_handle dh, NMEM nmem,
+ const char *tag, const char **attr, data1_node *at)
+{
+ return data1_mk_tag_n (dh, nmem, tag, strlen(tag), attr, at);
+}
+
+data1_node *data1_search_tag (data1_handle dh, data1_node *n,
+ const char *tag)
+{
+ if (*tag == '/')
{
- if (!at->child)
- at->child = tagn;
- else
+ n = data1_get_root_tag (dh, n);
+ if (n)
+ n = n->child;
+ tag++;
+ }
+ for (; n; n = n->next)
+ if (n->which == DATA1N_tag && n->u.tag.tag &&
+ !yaz_matchstr (tag, n->u.tag.tag))
{
- assert (at->last_child);
- at->last_child->next = tagn;
+ return n;
}
- at->last_child = tagn;
+ return 0;
+}
+
+data1_node *data1_mk_tag_uni (data1_handle dh, NMEM nmem,
+ const char *tag, data1_node *at)
+{
+ data1_node *node = data1_search_tag (dh, at->child, tag);
+ if (!node)
+ node = data1_mk_tag (dh, nmem, tag, 0 /* attr */, at);
+ else
+ node->child = node->last_child = 0;
+ return node;
+}
+
+data1_node *data1_mk_text_n (data1_handle dh, NMEM mem,
+ const char *buf, size_t len, data1_node *parent)
+{
+ data1_node *res = data1_mk_node2 (dh, mem, DATA1N_data, parent);
+ res->u.data.what = DATA1I_text;
+ res->u.data.len = len;
+
+ res->u.data.data = data1_insert_string_n (dh, res, mem, buf, len);
+ return res;
+}
+
+data1_node *data1_mk_text_nf (data1_handle dh, NMEM mem,
+ const char *buf, size_t len, data1_node *parent)
+{
+ data1_node *res = data1_mk_text_n (dh, mem, buf, len, parent);
+ res->u.data.formatted_text = 1;
+ return res;
+}
+
+data1_node *data1_mk_text (data1_handle dh, NMEM mem,
+ const char *buf, data1_node *parent)
+{
+ return data1_mk_text_n (dh, mem, buf, strlen(buf), parent);
+}
+
+data1_node *data1_mk_comment_n (data1_handle dh, NMEM mem,
+ const char *buf, size_t len,
+ data1_node *parent)
+{
+ data1_node *res = data1_mk_node2 (dh, mem, DATA1N_comment, parent);
+ res->u.data.what = DATA1I_text;
+ res->u.data.len = len;
+
+ res->u.data.data = data1_insert_string_n (dh, res, mem, buf, len);
+ return res;
+}
+
+data1_node *data1_mk_comment (data1_handle dh, NMEM mem,
+ const char *buf, data1_node *parent)
+{
+ return data1_mk_comment_n (dh, mem, buf, strlen(buf), parent);
+}
+
+char *data1_insert_string_n (data1_handle dh, data1_node *res,
+ NMEM m, const char *str, size_t len)
+{
+ char *b;
+ if (len >= DATA1_LOCALDATA)
+ b = nmem_malloc (m, len+1);
+ else
+ b = res->lbuf;
+ memcpy (b, str, len);
+ b[len] = 0;
+ return b;
+}
+
+char *data1_insert_string (data1_handle dh, data1_node *res,
+ NMEM m, const char *str)
+{
+ return data1_insert_string_n (dh, res, m, str, strlen(str));
+}
+
+static data1_node *data1_add_insert_taggeddata(data1_handle dh,
+ data1_node *at,
+ const char *tagname, NMEM m,
+ int local_allowed)
+{
+ data1_node *root = at->root;
+ data1_node *partag = get_parent_tag (dh, at);
+ data1_element *e = NULL;
+ data1_node *datn = 0;
+ data1_node *tagn = 0;
+
+ if (!partag)
+ e = data1_getelementbytagname (dh, root->u.root.absyn, 0, tagname);
+ else
+ {
+ e = partag->u.tag.element;
+ if (e)
+ e = data1_getelementbytagname (dh, root->u.root.absyn, e, tagname);
+ }
+ if (local_allowed || e)
+ {
+ tagn = data1_mk_node2 (dh, m, DATA1N_tag, at);
+ tagn->u.tag.tag = data1_insert_string (dh, tagn, m, tagname);
+ tagn->u.tag.element = e;
+ datn = data1_mk_node2 (dh, m, DATA1N_data, tagn);
}
return datn;
}
-data1_node *data1_add_taggeddata(data1_handle dh, data1_node *root,
- data1_node *at,
- const char *tagname, NMEM m)
+data1_node *data1_mk_tag_data(data1_handle dh, data1_node *at,
+ const char *tagname, NMEM m)
{
- return data1_add_insert_taggeddata (dh, root, at, tagname, m, 0, 1);
+ return data1_add_insert_taggeddata (dh, at, tagname, m, 1);
}
@@ -322,17 +381,145 @@ data1_node *data1_add_taggeddata(data1_handle dh, data1_node *root,
* which should be root or tag itself). Returns pointer to the data node,
* which can then be modified.
*/
-data1_node *data1_insert_taggeddata(data1_handle dh, data1_node *root,
- data1_node *at,
- const char *tagname, NMEM m)
+data1_node *data1_mk_tag_data_wd(data1_handle dh, data1_node *at,
+ const char *tagname, NMEM m)
{
- return data1_add_insert_taggeddata (dh, root, at, tagname, m, 1, 0);
+ return data1_add_insert_taggeddata (dh, at, tagname, m, 0);
+}
+
+data1_node *data1_insert_taggeddata (data1_handle dh, data1_node *root,
+ data1_node *at, const char *tagname,
+ NMEM m)
+{
+ return data1_add_insert_taggeddata (dh, at, tagname, m, 0);
+}
+
+data1_node *data1_add_taggeddata (data1_handle dh, data1_node *root,
+ data1_node *at, const char *tagname,
+ NMEM m)
+{
+ return data1_add_insert_taggeddata (dh, at, tagname, m, 1);
+}
+
+data1_node *data1_mk_tag_data_int (data1_handle dh, data1_node *at,
+ const char *tag, int num,
+ NMEM nmem)
+{
+ data1_node *node_data;
+
+ node_data = data1_mk_tag_data (dh, at, tag, nmem);
+ if (!node_data)
+ return 0;
+ node_data->u.data.what = DATA1I_num;
+ node_data->u.data.data = node_data->lbuf;
+ sprintf (node_data->u.data.data, "%d", num);
+ node_data->u.data.len = strlen (node_data->u.data.data);
+ return node_data;
+}
+
+data1_node *data1_mk_tag_data_oid (data1_handle dh, data1_node *at,
+ const char *tag, Odr_oid *oid,
+ NMEM nmem)
+{
+ data1_node *node_data;
+ char str[128], *p = str;
+ Odr_oid *ii;
+
+ node_data = data1_mk_tag_data (dh, at, tag, nmem);
+ if (!node_data)
+ return 0;
+
+ for (ii = oid; *ii >= 0; ii++)
+ {
+ if (ii != oid)
+ *p++ = '.';
+ sprintf (p, "%d", *ii);
+ p += strlen (p);
+ }
+ node_data->u.data.what = DATA1I_oid;
+ node_data->u.data.len = strlen (str);
+ node_data->u.data.data = data1_insert_string (dh, node_data, nmem, str);
+ return node_data;
+}
+
+
+data1_node *data1_mk_tag_data_text (data1_handle dh, data1_node *at,
+ const char *tag, const char *str,
+ NMEM nmem)
+{
+ data1_node *node_data;
+
+ node_data = data1_mk_tag_data (dh, at, tag, nmem);
+ if (!node_data)
+ return 0;
+ node_data->u.data.what = DATA1I_text;
+ node_data->u.data.len = strlen (str);
+ node_data->u.data.data = data1_insert_string (dh, node_data, nmem, str);
+ return node_data;
+}
+
+
+data1_node *data1_mk_tag_data_text_uni (data1_handle dh, data1_node *at,
+ const char *tag, const char *str,
+ NMEM nmem)
+{
+ data1_node *node = data1_search_tag (dh, at->child, tag);
+ if (!node)
+ return data1_mk_tag_data_text (dh, at, tag, str, nmem);
+ else
+ {
+ data1_node *node_data = node->child;
+ node_data->u.data.what = DATA1I_text;
+ node_data->u.data.len = strlen (str);
+ node_data->u.data.data = data1_insert_string (dh, node_data,
+ nmem, str);
+ node_data->child = node_data->last_child = 0;
+ return node_data;
+ }
+}
+
+static int ampr (int (*get_byte)(void *fh), void *fh, int *amp)
+{
+ int c = (*get_byte)(fh);
+ *amp = 0;
+ if (c == '&')
+ {
+ char ent[20];
+ int i = 0;
+
+ while (1)
+ {
+ c = (*get_byte)(fh);
+ if (c == ';')
+ {
+ ent[i] = 0;
+
+ c = ' ';
+ if (!strcmp (ent, "quot"))
+ c = '"';
+ if (!strcmp (ent, "apos"))
+ c = '\'';
+ if (!strcmp (ent, "gt"))
+ c = '>';
+ if (!strcmp (ent, "lt"))
+ c = '<';
+ if (!strcmp (ent, "amp"))
+ c = '&';
+ *amp = 1;
+ break;
+ }
+ else if (c == 0 || d1_isspace(c))
+ break;
+ if (i < 19)
+ ent[i++] = c;
+ }
+ }
+ return c;
}
-#if DATA1_USING_XATTR
data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
int (*get_byte)(void *fh), void *fh,
- WRBUF wrbuf, int *ch)
+ WRBUF wrbuf, int *ch, int *amp)
{
data1_xattr *p_first = 0;
data1_xattr **pp = &p_first;
@@ -341,11 +528,11 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
{
data1_xattr *p;
int len;
- while (c && d1_isspace(c))
- c = (*get_byte)(fh);
- if (!c || c == '>' || c == '/')
+ while (*amp || (c && d1_isspace(c)))
+ c = ampr (get_byte, fh, amp);
+ if (*amp == 0 && (c == 0 || c == '>' || c == '/'))
break;
- *pp = p = nmem_malloc (m, sizeof(*p));
+ *pp = p = (data1_xattr *) nmem_malloc (m, sizeof(*p));
p->next = 0;
pp = &p->next;
p->value = 0;
@@ -354,35 +541,57 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
while (c && c != '=' && c != '>' && c != '/' && !d1_isspace(c))
{
wrbuf_putc (wrbuf, c);
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, amp);
}
wrbuf_putc (wrbuf, '\0');
len = wrbuf_len(wrbuf);
- p->name = nmem_malloc (m, len);
+ p->name = (char*) nmem_malloc (m, len);
strcpy (p->name, wrbuf_buf(wrbuf));
if (c == '=')
{
- c = (*get_byte)(fh);
- if (c == '"')
- c = (*get_byte)(fh);
- wrbuf_rewind(wrbuf);
- while (c && c != '"' && c != '>' && c != '/')
+ c = ampr (get_byte, fh, amp);
+ if (*amp == 0 && c == '"')
{
- wrbuf_putc (wrbuf, c);
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, amp);
+ wrbuf_rewind(wrbuf);
+ while (*amp || (c && c != '"'))
+ {
+ wrbuf_putc (wrbuf, c);
+ c = ampr (get_byte, fh, amp);
+ }
+ if (c)
+ c = ampr (get_byte, fh, amp);
}
+ else if (*amp == 0 && c == '\'')
+ {
+ c = ampr (get_byte, fh, amp);
+ wrbuf_rewind(wrbuf);
+ while (*amp || (c && c != '\''))
+ {
+ wrbuf_putc (wrbuf, c);
+ c = ampr (get_byte, fh, amp);
+ }
+ if (c)
+ c = ampr (get_byte, fh, amp);
+ }
+ else
+ {
+ wrbuf_rewind(wrbuf);
+ while (*amp || (c && c != '>' && c != '/'))
+ {
+ wrbuf_putc (wrbuf, c);
+ c = ampr (get_byte, fh, amp);
+ }
+ }
wrbuf_putc (wrbuf, '\0');
len = wrbuf_len(wrbuf);
- p->value = nmem_malloc (m, len);
+ p->value = (char*) nmem_malloc (m, len);
strcpy (p->value, wrbuf_buf(wrbuf));
- if (c == '"')
- c = (*get_byte)(fh);
}
}
*ch = c;
return p_first;
}
-#endif
/*
* Ugh. Sometimes functions just grow and grow on you. This one reads a
@@ -391,74 +600,64 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
data1_node *data1_read_nodex (data1_handle dh, NMEM m,
int (*get_byte)(void *fh), void *fh, WRBUF wrbuf)
{
- data1_absyn *absyn = 0;
data1_node *d1_stack[256];
data1_node *res;
- int c;
+ int c, amp;
int level = 0;
int line = 1;
d1_stack[level] = 0;
- c = (*get_byte)(fh);
- while (1)
+ c = ampr (get_byte, fh, &);
+ while (c != '\0')
{
data1_node *parent = level ? d1_stack[level-1] : 0;
- while (c != '\0' && d1_isspace(c))
- {
- if (c == '\n')
- line++;
- c = (*get_byte)(fh);
- }
- if (c == '\0')
- break;
-
- if (c == '<') /* beginning of tag */
+
+ if (amp == 0 && c == '<') /* beginning of tag */
{
-#if DATA1_USING_XATTR
data1_xattr *xattr;
-#endif
+
char tag[64];
char args[256];
int null_tag = 0;
int end_tag = 0;
size_t i = 0;
- c = (*get_byte)(fh);
- if (c == '/')
+ c = ampr (get_byte, fh, &);
+ if (amp == 0 && c == '/')
{
end_tag = 1;
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
}
- while (c && c != '>' && c != '/' && !d1_isspace(c))
+ else if (amp == 0 && c == '!')
+ /* tags/comments that we don't deal with yet */
+ {
+ while (amp || (c && c != '>'))
+ c = ampr (get_byte, fh, &);
+ if (c)
+ c = ampr (get_byte, fh, &);
+ continue;
+ }
+ while (amp || (c && c != '>' && c != '/' && !d1_isspace(c)))
{
if (i < (sizeof(tag)-1))
tag[i++] = c;
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
}
tag[i] = '\0';
-#if DATA1_USING_XATTR
- xattr = data1_read_xattr (dh, m, get_byte, fh, wrbuf, &c);
+ xattr = data1_read_xattr (dh, m, get_byte, fh, wrbuf, &c, &);
args[0] = '\0';
-#else
- while (d1_isspace(c))
- c = (*get_byte)(fh);
- for (i = 0; c && c != '>' && c != '/'; c = (*get_byte)(fh))
- if (i < (sizeof(args)-1))
- args[i++] = c;
- args[i] = '\0';
-#endif
- if (c == '/')
+ if (amp == 0 && c == '/')
{ /* or */
null_tag = 1;
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
}
- if (c != '>')
+ if (amp || c != '>')
{
yaz_log(LOG_WARN, "d1: %d: Malformed tag", line);
return 0;
}
else
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
/* End tag? */
if (end_tag)
@@ -487,22 +686,18 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
break;
}
}
- if (level == 0)
- return d1_stack[0];
+ if (data1_is_xmlmode(dh))
+ {
+ if (level <= 1)
+ return d1_stack[0];
+ }
+ else
+ {
+ if (level <= 0)
+ return d1_stack[0];
+ }
continue;
}
- if (level == 0) /* root ? */
- {
- if (!(absyn = data1_get_absyn (dh, tag)))
- {
- yaz_log(LOG_WARN, "Unable to acquire abstract syntax " "for '%s'", tag);
- /* It's now OK for a record not to have an absyn */
- }
- res = data1_mk_node_type (dh, m, DATA1N_root);
- res->u.root.type = data1_insert_string (dh, res, m, tag);
- res->u.root.absyn = absyn;
- res->root = res;
- }
else if (!strcmp(tag, "var"))
{
char tclass[DATA1_MAX_SYMBOL], type[DATA1_MAX_SYMBOL];
@@ -525,10 +720,7 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
*/
if (parent->which != DATA1N_variant)
{
- res = data1_mk_node (dh, m);
- res->which = DATA1N_variant;
- res->u.variant.type = 0;
- res->u.variant.value = 0;
+ res = data1_mk_node2 (dh, m, DATA1N_variant, parent);
}
else
{
@@ -543,77 +735,59 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
level = i;
break;
}
- res = data1_mk_node (dh, m);
- res->which = DATA1N_variant;
+ res = data1_mk_node2 (dh, m, DATA1N_variant, parent);
res->u.variant.type = tp;
res->u.variant.value =
data1_insert_string (dh, res, m, args + val_offset);
}
}
- else /* tag.. acquire our element in the abstract syntax */
- {
- data1_node *partag = get_parent_tag (dh, parent);
- data1_element *elem, *e = 0;
- int localtag = 0;
-
- if (parent->which == DATA1N_variant)
- return 0;
- if (partag)
- if (!(e = partag->u.tag.element))
- localtag = 1; /* our parent is a local tag */
-
- elem = data1_getelementbytagname(dh, absyn, e, tag);
- res = data1_mk_node_type (dh, m, DATA1N_tag);
- res->u.tag.tag = data1_insert_string (dh, res, m, tag);
- res->u.tag.element = elem;
-#if DATA1_USING_XATTR
- res->u.tag.attributes = xattr;
-#endif
- }
- if (parent)
- {
- parent->last_child = res;
- res->root = parent->root;
- }
- res->parent = parent;
- if (d1_stack[level])
- d1_stack[level]->next = res;
- else if (parent)
- parent->child = res;
+ else
+ {
+
+ /* tag .. acquire our element in the abstract syntax */
+ if (level == 0)
+ {
+ parent = data1_mk_root (dh, m, tag);
+ res = d1_stack[level] = parent;
+
+ if (data1_is_xmlmode(dh))
+ {
+ level++;
+ res = data1_mk_tag (dh, m, tag, 0 /* attr */, parent);
+ res->u.tag.attributes = xattr;
+ }
+ }
+ else
+ {
+ res = data1_mk_tag (dh, m, tag, 0 /* attr */, parent);
+ res->u.tag.attributes = xattr;
+ }
+ }
d1_stack[level] = res;
d1_stack[level+1] = 0;
- if (!null_tag)
+ if (level < 250 && !null_tag)
++level;
}
else /* != '<'... this is a body of text */
{
- const char *src;
- char *dst;
- int len, prev_char = 0;
+ int len;
if (level == 0)
{
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
continue;
}
- res = data1_mk_node_type (dh, m, DATA1N_data);
- res->parent = parent;
+ res = data1_mk_node2 (dh, m, DATA1N_data, parent);
res->u.data.what = DATA1I_text;
res->u.data.formatted_text = 0;
- res->root = parent->root;
- parent->last_child = res;
- if (d1_stack[level])
- d1_stack[level]->next = res;
- else
- parent->child = res;
d1_stack[level] = res;
wrbuf_rewind(wrbuf);
- while (c && c != '<')
+ while (amp || (c && c != '<'))
{
wrbuf_putc (wrbuf, c);
- c = (*get_byte)(fh);
+ c = ampr (get_byte, fh, &);
}
len = wrbuf_len(wrbuf);
@@ -622,26 +796,12 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
res->u.data.data = (char*) nmem_malloc (m, len);
else
res->u.data.data = res->lbuf;
-
- /* read "data" and transfer while removing white space */
- dst = res->u.data.data;
- for (src = wrbuf_buf(wrbuf); --len >= 0; src++)
- {
- if (*src == '\n')
- line++;
- if (d1_isspace (*src))
- prev_char = ' ';
- else
- {
- if (prev_char)
- {
- *dst++ = prev_char;
- prev_char = 0;
- }
- *dst++ = *src;
- }
- }
- res->u.data.len = dst - res->u.data.data;
+
+ if (len)
+ memcpy (res->u.data.data, wrbuf_buf(wrbuf), len);
+ else
+ res->u.data.data = 0;
+ res->u.data.len = len;
}
}
return 0;
@@ -705,3 +865,124 @@ data1_node *data1_read_sgml (data1_handle dh, NMEM m, const char *buf)
return data1_read_node (dh, &bp, m);
}
+
+static int conv_item (NMEM m, yaz_iconv_t t,
+ WRBUF wrbuf, char *inbuf, size_t inlen)
+{
+ wrbuf_rewind (wrbuf);
+ if (wrbuf->size < 10)
+ wrbuf_grow (wrbuf, 10);
+ for (;;)
+ {
+ char *outbuf = wrbuf->buf + wrbuf->pos;
+ size_t outlen = wrbuf->size - wrbuf->pos;
+ if (yaz_iconv (t, &inbuf, &inlen, &outbuf, &outlen) ==
+ (size_t)(-1) && yaz_iconv_error(t) != YAZ_ICONV_E2BIG)
+ {
+ /* bad data. stop and skip conversion entirely */
+ return -1;
+ }
+ else if (inlen == 0)
+ { /* finished converting */
+ wrbuf->pos = wrbuf->size - outlen;
+ break;
+ }
+ else
+ {
+ /* buffer too small: make sure we expand buffer */
+ wrbuf->pos = wrbuf->size - outlen;
+ wrbuf_grow(wrbuf, 20);
+ }
+ }
+ return 0;
+}
+
+static void data1_iconv_s (data1_handle dh, NMEM m, data1_node *n,
+ yaz_iconv_t t, WRBUF wrbuf, const char *tocode)
+{
+ for (; n; n = n->next)
+ {
+ switch (n->which)
+ {
+ case DATA1N_data:
+ case DATA1N_comment:
+ if (conv_item (m, t, wrbuf, n->u.data.data, n->u.data.len) == 0)
+ {
+ n->u.data.data =
+ data1_insert_string_n (dh, n, m, wrbuf->buf,
+ wrbuf->pos);
+ n->u.data.len = wrbuf->pos;
+ }
+ break;
+ case DATA1N_tag:
+ if (conv_item (m, t, wrbuf, n->u.tag.tag, strlen(n->u.tag.tag))
+ == 0)
+ {
+ n->u.tag.tag =
+ data1_insert_string_n (dh, n, m,
+ wrbuf->buf, wrbuf->pos);
+ }
+ if (n->u.tag.attributes)
+ {
+ data1_xattr *p;
+ for (p = n->u.tag.attributes; p; p = p->next)
+ {
+ if (p->value &&
+ conv_item(m, t, wrbuf, p->value, strlen(p->value))
+ == 0)
+ {
+ wrbuf_puts (wrbuf, "");
+ p->value = nmem_strdup (m, wrbuf->buf);
+ }
+ }
+ }
+ break;
+ case DATA1N_preprocess:
+ if (strcmp(n->u.preprocess.target, "xml") == 0)
+ {
+ data1_xattr *p = n->u.preprocess.attributes;
+ for (; p; p = p->next)
+ if (strcmp (p->name, "encoding") == 0)
+ p->value = nmem_strdup (m, tocode);
+ }
+ break;
+ }
+ data1_iconv_s (dh, m, n->child, t, wrbuf, tocode);
+ }
+}
+
+const char *data1_get_encoding (data1_handle dh, data1_node *n)
+{
+ /* see if we have an xml header that specifies encoding */
+ if (n && n->child && n->child->which == DATA1N_preprocess &&
+ strcmp (n->child->u.preprocess.target, "xml") == 0)
+ {
+ data1_xattr *xp = n->child->u.preprocess.attributes;
+ for (; xp; xp = xp->next)
+ if (!strcmp (xp->name, "encoding") == 0)
+ return xp->value;
+ }
+ /* no encoding in header, so see if "encoding" was specified for abs */
+ if (n && n->which == DATA1N_root &&
+ n->u.root.absyn && n->u.root.absyn->encoding)
+ return n->u.root.absyn->encoding;
+ /* none of above, return a hard coded default */
+ return "ISO-8859-1";
+}
+
+int data1_iconv (data1_handle dh, NMEM m, data1_node *n,
+ const char *tocode,
+ const char *fromcode)
+{
+ if (strcmp (tocode, fromcode))
+ {
+ WRBUF wrbuf = wrbuf_alloc();
+ yaz_iconv_t t = yaz_iconv_open (tocode, fromcode);
+ if (!t)
+ return -1;
+ data1_iconv_s (dh, m, n, t, wrbuf, tocode);
+ yaz_iconv_close (t);
+ wrbuf_free (wrbuf, 1);
+ }
+ return 0;
+}