Added data1_chop_text which removes whitespace in cdata nodes

[idzebra-moved-to-github.git] / data1 / d1_read.c

diff --git a/data1/d1_read.c b/data1/d1_read.c
index 56fdffd..30e9732 100644 (file)
--- a/data1/d1_read.c
+++ b/data1/d1_read.c
@@ -1,4 +1,4 @@
-/* $Id: d1_read.c,v 1.22 2007-01-15 15:10:14 adam Exp $
+/* $Id: d1_read.c,v 1.27 2007-06-27 22:04:45 adam Exp $
    Copyright (C) 1995-2007
    Index Data ApS
 
@@ -174,6 +174,7 @@ data1_node *data1_mk_root (data1_handle dh, NMEM nmem, const char *name)
 {
     data1_absyn *absyn = data1_get_absyn(dh, name, 1);
     data1_node *res;
+
     if (!absyn)
     {
         yaz_log(YLOG_WARN, "Unable to acquire abstract syntax " "for '%s'",
@@ -592,7 +593,6 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
     for (;;)
     {
        data1_xattr *p;
-       int len;
        while (*amp || (c && d1_isspace(c)))
            c = ampr (get_byte, fh, amp);
        if (*amp == 0 && (c == 0 || c == '>' || c == '/'))
@@ -609,10 +609,7 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
            wrbuf_putc (wrbuf, c);
            c = ampr (get_byte, fh, amp);
        }
-       wrbuf_putc (wrbuf, '\0');
-       len = wrbuf_len(wrbuf);
-       p->name = (char*) nmem_malloc (m, len);
-       strcpy (p->name, wrbuf_buf(wrbuf));
+       p->name = nmem_strdup (m, wrbuf_cstr(wrbuf));
        if (c == '=')
        {
            c = ampr (get_byte, fh, amp);
@@ -649,10 +646,7 @@ data1_xattr *data1_read_xattr (data1_handle dh, NMEM m,
                    c = ampr (get_byte, fh, amp);
                }
             }
-           wrbuf_putc (wrbuf, '\0');
-           len = wrbuf_len(wrbuf);
-           p->value = (char*) nmem_malloc (m, len);
-           strcpy (p->value, wrbuf_buf(wrbuf));
+           p->value = nmem_strdup(m, wrbuf_cstr(wrbuf));
        }
     }
     *ch = c;
@@ -965,7 +959,7 @@ data1_node *data1_read_node (data1_handle dh, const char **buf, NMEM m)
     data1_node *node;
 
     node = data1_read_nodex(dh, m, getc_mem, (void *) (buf), wrbuf);
-    wrbuf_free (wrbuf, 1);
+    wrbuf_destroy(wrbuf);
     return node;
 }
 
@@ -1010,34 +1004,12 @@ data1_node *data1_read_sgml (data1_handle dh, NMEM m, const char *buf)
 }
 
 
-static int conv_item (NMEM m, yaz_iconv_t t, 
-                      WRBUF wrbuf, char *inbuf, size_t inlen)
+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);
-        }
-    }
+    wrbuf_rewind(wrbuf);
+    wrbuf_iconv_write(wrbuf, t, inbuf, inlen);
+    wrbuf_iconv_reset(wrbuf, t);
     return 0;
 }
 
@@ -1075,8 +1047,7 @@ static void data1_iconv_s (data1_handle dh, NMEM m, data1_node *n,
                         conv_item(m, t, wrbuf, p->value, strlen(p->value))
                         == 0)
                     {
-                        wrbuf_puts (wrbuf, "");
-                        p->value = nmem_strdup (m, wrbuf->buf);
+                        p->value = nmem_strdup(m, wrbuf_cstr(wrbuf));
                     }
                 }
             }
@@ -1115,8 +1086,8 @@ const char *data1_get_encoding (data1_handle dh, data1_node *n)
 }
 
 int data1_iconv (data1_handle dh, NMEM m, data1_node *n,
-                  const char *tocode, 
-                  const char *fromcode)
+                 const char *tocode, 
+                 const char *fromcode)
 {
     if (yaz_matchstr (tocode, fromcode))
     {
@@ -1124,16 +1095,45 @@ int data1_iconv (data1_handle dh, NMEM m, data1_node *n,
         yaz_iconv_t t = yaz_iconv_open(tocode, fromcode);
         if (!t)
        {
-            wrbuf_free(wrbuf, 1);
+            wrbuf_destroy(wrbuf);
             return -1;
        }
         data1_iconv_s(dh, m, n, t, wrbuf, tocode);
         yaz_iconv_close(t);
-        wrbuf_free(wrbuf, 1);
+        wrbuf_destroy(wrbuf);
     }
     return 0;
 }
 
+void data1_chop_text(data1_handle dh, NMEM m, data1_node *n)
+{
+    for (; n; n = n->next)
+    {
+        if (n->which == DATA1N_data)
+        {
+            
+            int sz = n->u.data.len;
+            const char *ndata = n->u.data.data;
+            int off = 0;
+            
+            for (off = 0; off < sz; off++)
+                if (!d1_isspace(ndata[off]))
+                    break;
+            sz = sz - off;
+            ndata += off;
+            
+            while (sz && d1_isspace(ndata[sz - 1]))
+                sz--;
+
+            n->u.data.data = nmem_malloc(m, sz);
+            n->u.data.len = sz;
+            memcpy(n->u.data.data, ndata, sz);
+               
+        }
+        data1_chop_text(dh, m, n->child);
+    }
+}
+
 void data1_concat_text(data1_handle dh, NMEM m, data1_node *n)
 {
     for (; n; n = n->next)
@@ -1163,6 +1163,7 @@ void data1_concat_text(data1_handle dh, NMEM m, data1_node *n)
         data1_concat_text(dh, m, n->child);
     }
 }
+
 /*
  * Local variables:
  * c-basic-offset: 4