Fixed bug #416: First/second comb characters not handled in the

MARC-8 to UTF-8 conversion. Patch by Raj Patel.

diff --git a/NEWS b/NEWS
index e410645..4cb318b 100644 (file)
--- a/NEWS
+++ b/NEWS
@@ -1,9 +1,10 @@
+Fixed bug #416: First/second comb characters not handled in the
+MARC-8 to UTF-8 conversion. Patch by Raj Patel.
+
 Added option -m to GFS to control formatting of timestamps in log file.
 
 Added 'exit' as synonym for 'quit' in yaz-client
 
-Possible compatibility problems with earlier versions marked with '*'.
-
 Added support for specifying SRW resultSetId + resultSetIdleTime in a
 Generic Frontend Server's search handler. Refer to include/yaz/backend.h
 for the new members. Patch by Ko van der Sloot.

diff --git a/src/siconv.c b/src/siconv.c
index 378bd75..285d644 100644 (file)
--- a/src/siconv.c
+++ b/src/siconv.c
@@ -2,7 +2,7 @@
  * Copyright (C) 1995-2005, Index Data ApS
  * See the file LICENSE for details.
  *
- * $Id: siconv.c,v 1.13 2005-06-25 15:46:05 adam Exp $
+ * $Id: siconv.c,v 1.14 2005-10-28 18:36:58 adam Exp $
  */
 /**
  * \file siconv.c
@@ -266,6 +266,14 @@ static unsigned long yaz_read_marc8 (yaz_iconv_t cd, unsigned char *inp,
     {
         *no_read = cd->comb_no_read[cd->comb_offset];
         x = cd->comb_x[cd->comb_offset];
+
+        /* special case for double-diacritic combining characters, 
+           INVERTED BREVE and DOUBLE TILDE.
+           We'll increment the no_read counter by 1, since we want to skip over
+           the processing of the closing ligature character
+        */
+        if (x == 0x0361 || x == 0x0360)
+            *no_read += 1;
         cd->comb_offset++;
         return x;
     }

diff --git a/test/tsticonv.c b/test/tsticonv.c
index c4e75d7..61492b5 100644 (file)
--- a/test/tsticonv.c
+++ b/test/tsticonv.c
@@ -2,7 +2,7 @@
  * Copyright (C) 1995-2005, Index Data ApS
  * See the file LICENSE for details.
  *
- * $Id: tsticonv.c,v 1.11 2005-08-22 20:34:23 adam Exp $
+ * $Id: tsticonv.c,v 1.12 2005-10-28 18:36:59 adam Exp $
  */
 
 #if HAVE_CONFIG_H
@@ -159,6 +159,14 @@ static void tst_marc8_to_ucs4b()
         "\xe5\xe8\x41",
         12, "\x00\x00\x00\x41" "\x00\x00\x03\x04" "\x00\x00\x03\x08"
     }, 
+    { /* bug #416 */
+        "\xEB\x74\xEC\x73",
+        12, "\x00\x00\x00\x74" "\x00\x00\x03\x61" "\x00\x00\x00\x73"
+    },
+    { /* bug #416 */
+        "\xFA\x74\xFB\x73",
+        12, "\x00\x00\x00\x74" "\x00\x00\x03\x60" "\x00\x00\x00\x73"
+    },
     {
         0, 0, 0
     }