X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=util%2Fsiconv.c;h=73d7148c06942fe3629c78419fd5f1875cfb5128;hp=c89de49d8140675b978e6ffe7fa2213e7d1c80b1;hb=4b3d1e197d75fcb2607be8d86d244f20c4180000;hpb=d054cdd41417401fbba915e20203882159f32283

diff --git a/util/siconv.c b/util/siconv.c
index c89de49..73d7148 100644
--- a/util/siconv.c
+++ b/util/siconv.c
@@ -1,8 +1,8 @@
 /*
- * Copyright (c) 1997-2002, Index Data
+ * Copyright (c) 1997-2003, Index Data
  * See the file LICENSE for details.
  *
- * $Id: siconv.c,v 1.6 2002-09-25 12:37:08 adam Exp $
+ * $Id: siconv.c,v 1.9 2003-01-06 08:20:28 adam Exp $
  */
 
 /* mini iconv and wrapper for system iconv library (if present) */
@@ -14,6 +14,9 @@
 #include <errno.h>
 #include <string.h>
 #include <ctype.h>
+#if HAVE_WCHAR_H
+#include <wchar.h>
+#endif
 
 #if HAVE_ICONV_H
 #include <iconv.h>
@@ -21,6 +24,9 @@
 
 #include <yaz/yaz-util.h>
 
+unsigned long yaz_marc8_conv (unsigned char *inp, size_t inbytesleft,
+                              size_t *no_read);
+    
 struct yaz_iconv_struct {
     int my_errno;
     int init_flag;
@@ -185,6 +191,34 @@ static unsigned long yaz_read_UCS4LE (yaz_iconv_t cd, unsigned char *inp,
     return x;
 }
 
+#if HAVE_WCHAR_H
+static unsigned long yaz_read_wchar_t (yaz_iconv_t cd, unsigned char *inp,
+                                       size_t inbytesleft, size_t *no_read)
+{
+    unsigned long x = 0;
+    
+    if (inbytesleft < sizeof(wchar_t))
+    {
+        cd->my_errno = YAZ_ICONV_EINVAL; /* incomplete input */
+        *no_read = 0;
+    }
+    else
+    {
+        wchar_t wch;
+        memcpy (&wch, inp, sizeof(wch));
+        x = wch;
+        *no_read = sizeof(wch);
+    }
+    return x;
+}
+#endif
+
+static unsigned long yaz_read_marc8 (yaz_iconv_t cd, unsigned char *inp,
+                                     size_t inbytesleft, size_t *no_read)
+{
+    return yaz_marc8_conv(inp, inbytesleft, no_read);
+}
+
 static size_t yaz_write_UTF8 (yaz_iconv_t cd, unsigned long x,
                               char **outbuf, size_t *outbytesleft)
 {
@@ -309,6 +343,34 @@ static size_t yaz_write_UCS4LE (yaz_iconv_t cd, unsigned long x,
     return 0;
 }
 
+#if HAVE_WCHAR_H
+static size_t yaz_write_wchar_t (yaz_iconv_t cd, unsigned long x,
+                                 char **outbuf, size_t *outbytesleft)
+{
+    unsigned char *outp = (unsigned char *) *outbuf;
+
+    if (*outbytesleft >= sizeof(wchar_t))
+    {
+        wchar_t wch = x;
+        memcpy(outp, &wch, sizeof(wch));
+        outp += sizeof(wch);
+        (*outbytesleft) -= sizeof(wch);
+    }
+    else
+    {
+        cd->my_errno = YAZ_ICONV_E2BIG;
+        return (size_t)(-1);
+    }
+    *outbuf = (char *) outp;
+    return 0;
+}
+#endif
+
+int yaz_iconv_isbuiltin(yaz_iconv_t cd)
+{
+    return cd->read_handle && cd->write_handle;
+}
+
 yaz_iconv_t yaz_iconv_open (const char *tocode, const char *fromcode)
 {
     yaz_iconv_t cd = (yaz_iconv_t) xmalloc (sizeof(*cd));
@@ -318,27 +380,43 @@ yaz_iconv_t yaz_iconv_open (const char *tocode, const char *fromcode)
     cd->init_handle = 0;
     cd->my_errno = YAZ_ICONV_UNKNOWN;
 
-    if (!yaz_matchstr(fromcode, "UTF8"))
+    /* a useful hack: if fromcode has leading @,
+       the library not use YAZ's own conversions .. */
+    if (fromcode[0] == '@')
+        fromcode++;
+    else
     {
-        cd->read_handle = yaz_read_UTF8;
-        cd->init_handle = yaz_init_UTF8;
+        if (!yaz_matchstr(fromcode, "UTF8"))
+        {
+            cd->read_handle = yaz_read_UTF8;
+            cd->init_handle = yaz_init_UTF8;
+        }
+        else if (!yaz_matchstr(fromcode, "ISO88591"))
+            cd->read_handle = yaz_read_ISO8859_1;
+        else if (!yaz_matchstr(fromcode, "UCS4"))
+            cd->read_handle = yaz_read_UCS4;
+        else if (!yaz_matchstr(fromcode, "UCS4LE"))
+            cd->read_handle = yaz_read_UCS4LE;
+        else if (!yaz_matchstr(fromcode, "MARC8"))
+            cd->read_handle = yaz_read_marc8;
+#if HAVE_WCHAR_H
+        else if (!yaz_matchstr(fromcode, "WCHAR_T"))
+            cd->read_handle = yaz_read_wchar_t;
+#endif
+        
+        if (!yaz_matchstr(tocode, "UTF8"))
+            cd->write_handle = yaz_write_UTF8;
+        else if (!yaz_matchstr(tocode, "ISO88591"))
+            cd->write_handle = yaz_write_ISO8859_1;
+        else if (!yaz_matchstr (tocode, "UCS4"))
+            cd->write_handle = yaz_write_UCS4;
+        else if (!yaz_matchstr(tocode, "UCS4LE"))
+            cd->write_handle = yaz_write_UCS4LE;
+#if HAVE_WCHAR_H
+        else if (!yaz_matchstr(tocode, "WCHAR_T"))
+            cd->write_handle = yaz_write_wchar_t;
+#endif
     }
-    else if (!yaz_matchstr(fromcode, "ISO88591"))
-        cd->read_handle = yaz_read_ISO8859_1;
-    else if (!yaz_matchstr(fromcode, "UCS4"))
-        cd->read_handle = yaz_read_UCS4;
-    else if (!yaz_matchstr(fromcode, "UCS4LE"))
-        cd->read_handle = yaz_read_UCS4LE;
-    
-    if (!yaz_matchstr(tocode, "UTF8"))
-        cd->write_handle = yaz_write_UTF8;
-    else if (!yaz_matchstr(tocode, "ISO88591"))
-        cd->write_handle = yaz_write_ISO8859_1;
-    else if (!yaz_matchstr (tocode, "UCS4"))
-        cd->write_handle = yaz_write_UCS4;
-    else if (!yaz_matchstr(tocode, "UCS4LE"))
-        cd->write_handle = yaz_write_UCS4LE;
-
 #if HAVE_ICONV_H
     cd->iconv_cd = 0;
     if (!cd->read_handle || !cd->write_handle)