X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=retrieval%2Fd1_read.c;h=130c51853d7dfcf2acef1e66f045b50a3f11ff45;hb=ac5f8ad9508f5b3ada38117b0da7ae23c650ab1e;hp=66baf41cc3a20db3a2decf2dea20c57fcb0bd1b1;hpb=a774858896a9b56c966edbfb70d8eae1c91c4ad1;p=yaz-moved-to-github.git

diff --git a/retrieval/d1_read.c b/retrieval/d1_read.c
index 66baf41..130c518 100644
--- a/retrieval/d1_read.c
+++ b/retrieval/d1_read.c
@@ -3,7 +3,7 @@
  * See the file LICENSE for details.
  * Sebastian Hammer, Adam Dickmeiss
  *
- * $Id: d1_read.c,v 1.44 2002-07-03 10:04:04 adam Exp $
+ * $Id: d1_read.c,v 1.51 2002-08-28 07:54:11 adam Exp $
  */
 
 #include <assert.h>
@@ -11,6 +11,7 @@
 #include <stdlib.h>
 
 #include <yaz/xmalloc.h>
+#include <yaz/yaz-util.h>
 #include <yaz/log.h>
 #include <yaz/data1.h>
 
@@ -18,9 +19,12 @@ data1_node *data1_get_root_tag (data1_handle dh, data1_node *n)
 {
     if (!n)
         return 0;
-    n = n->child;
-    while (n && n->which != DATA1N_tag)
-        n = n->next;
+    if (data1_is_xmlmode(dh))
+    {
+        n = n->child;
+        while (n && n->which != DATA1N_tag)
+            n = n->next;
+    }
     return n;
 }
         
@@ -30,9 +34,19 @@ data1_node *data1_get_root_tag (data1_handle dh, data1_node *n)
  */
 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;
 }
 
@@ -176,16 +190,24 @@ data1_node *data1_mk_tag_n (data1_handle dh, NMEM nmem,
 {
     data1_node *partag = get_parent_tag(dh, at);
     data1_node *res = data1_mk_node2 (dh, nmem, DATA1N_tag, at);
-    data1_element *e = NULL;
     data1_xattr **p;
+    data1_element *e = 0;
     
     res->u.tag.tag = data1_insert_string_n (dh, res, nmem, tag, len);
     
-    if (partag)
-	e = partag->u.tag.element;
-    res->u.tag.element =
-	data1_getelementbytagname (dh, at->root->u.root.absyn,
-				   e, res->u.tag.tag);
+    if (!partag)  /* top tag? */
+        e  = data1_getelementbytagname (dh, at->root->u.root.absyn,
+                                        0 /* index as local */,
+                                        res->u.tag.tag);
+    else
+    {
+        /* 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)
     {
@@ -329,9 +351,14 @@ static data1_node *data1_add_insert_taggeddata(data1_handle dh,
     data1_node *datn = 0;
     data1_node *tagn = 0;
 
-    if (partag)
+    if (!partag)
+        e = data1_getelementbytagname (dh, root->u.root.absyn, 0, tagname);
+    else 
+    {
 	e = partag->u.tag.element;
-    e = data1_getelementbytagname (dh, root->u.root.absyn, e, tagname);
+        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);
@@ -451,10 +478,48 @@ data1_node *data1_mk_tag_data_text_uni (data1_handle dh, data1_node *at,
     }
 }
 
+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;
+}
 
 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;
@@ -463,9 +528,9 @@ 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 = (data1_xattr *) nmem_malloc (m, sizeof(*p));
 	p->next = 0;
@@ -476,7 +541,7 @@ 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);
@@ -484,26 +549,38 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
 	strcpy (p->name, wrbuf_buf(wrbuf));
 	if (c == '=')
 	{
-	    c = (*get_byte)(fh);
-	    if (c == '"')
+	    c = ampr (get_byte, fh, amp);
+	    if (*amp == 0 && c == '"')
 	    {
-		c = (*get_byte)(fh);	
+		c = ampr (get_byte, fh, amp);	
 		wrbuf_rewind(wrbuf);
-		while (c && c != '"')
+		while (*amp || (c && c != '"'))
 		{
 		    wrbuf_putc (wrbuf, c);
-		    c = (*get_byte)(fh);
+		    c = ampr (get_byte, fh, amp);
 	        }
 	        if (c)
-		    c = (*get_byte)(fh);	
+		    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 (c && c != '>' && c != '/')
+	        while (*amp || (c && c != '>' && c != '/'))
 	        {
 		    wrbuf_putc (wrbuf, c);
-		    c = (*get_byte)(fh);
+		    c = ampr (get_byte, fh, amp);
 	        }
             }
 	    wrbuf_putc (wrbuf, '\0');
@@ -525,25 +602,17 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
 {
     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, &amp);
+    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 */
 	{
 	    data1_xattr *xattr;
 
@@ -553,41 +622,42 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
 	    int end_tag = 0;
 	    size_t i = 0;
 
-	    c = (*get_byte)(fh);
-	    if (c == '/')
+	    c = ampr (get_byte, fh, &amp);
+	    if (amp == 0 && c == '/')
 	    {
 		end_tag = 1;
-		c = (*get_byte)(fh);
+		c = ampr (get_byte, fh, &amp);
 	    }
-	    else if (c == '!')  /* tags/comments that we don't deal with yet */
+	    else if (amp == 0 && c == '!')
+                /* tags/comments that we don't deal with yet */
 	    {
-		while (c && c != '>')
-		    c = (*get_byte)(fh);
+		while (amp || (c && c != '>'))
+		    c = ampr (get_byte, fh, &amp);
 		if (c)
-		    c = (*get_byte)(fh);
+		    c = ampr (get_byte, fh, &amp);
 		continue;
 	    }
-	    while (c && c != '>' && c != '/' && !d1_isspace(c))
+	    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, &amp);
 	    }
 	    tag[i] = '\0';
-	    xattr = data1_read_xattr (dh, m, get_byte, fh, wrbuf, &c);
+	    xattr = data1_read_xattr (dh, m, get_byte, fh, wrbuf, &c, &amp);
 	    args[0] = '\0';
-	    if (c == '/')
+	    if (amp == 0 && c == '/')
 	    {    /* <tag attrs/> or <tag/> */
 		null_tag = 1;
-		c = (*get_byte)(fh);
+		c = ampr (get_byte, fh, &amp);
 	    }
-	    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, &amp);
 
 	    /* End tag? */
 	    if (end_tag)       
@@ -616,8 +686,16 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
 			break;
 		    }
 		}
-		if (level <= 1)
-		    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;
 	    }	
 	    else if (!strcmp(tag, "var"))
@@ -670,10 +748,20 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
                 if (level == 0)
                 {
                     parent = data1_mk_root (dh, m, tag);
-                    d1_stack[level++] = parent;
+                    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;
                 }
-                res = data1_mk_tag (dh, m, tag, 0 /* attr */, parent);
-                res->u.tag.attributes = xattr;
             }
 	    d1_stack[level] = res;
 	    d1_stack[level+1] = 0;
@@ -682,13 +770,11 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
 	}
 	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, &amp);
 		continue;
 	    }
 	    res = data1_mk_node2 (dh, m, DATA1N_data, parent);
@@ -698,10 +784,10 @@ data1_node *data1_read_nodex (data1_handle dh, NMEM m,
 	    
 	    wrbuf_rewind(wrbuf);
 
-	    while (c && c != '<')
+	    while (amp || (c && c != '<'))
 	    {
 		wrbuf_putc (wrbuf, c);
-		c = (*get_byte)(fh);
+		c = ampr (get_byte, fh, &amp);
 	    }
 	    len = wrbuf_len(wrbuf);
 
@@ -710,26 +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;
@@ -793,3 +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;
+}