Implemented yaz_iconv to support conversion to MARC-8

[yaz-moved-to-github.git] / test / tsticonv.c

diff --git a/test/tsticonv.c b/test/tsticonv.c
index cafe901..fd8f2c8 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 1.14 2006-02-22 19:48:53 adam Exp $
+ * $Id: tsticonv.c,v 1.17 2006-04-19 23:15:40 adam Exp $
  */
 
 #if HAVE_CONFIG_H
@@ -349,10 +349,147 @@ int utf8_check(unsigned c)
     return 1;
 }
         
+static int tst_convert(yaz_iconv_t cd, const char *buf, const char *cmpbuf)
+{
+    int ret = 0;
+    WRBUF b = wrbuf_alloc();
+    char outbuf[12];
+    size_t inbytesleft = strlen(buf);
+    const char *inp = buf;
+    while (inbytesleft)
+    {
+        size_t outbytesleft = sizeof(outbuf);
+        char *outp = outbuf;
+        size_t r = yaz_iconv(cd, (char**) &inp,  &inbytesleft,
+                             &outp, &outbytesleft);
+        if (r == (size_t) (-1))
+        {
+            int e = yaz_iconv_error(cd);
+            if (e != YAZ_ICONV_E2BIG)
+                break;
+        }
+        wrbuf_write(b, outbuf, outp - outbuf);
+    }
+    if (wrbuf_len(b) == strlen(cmpbuf) 
+        && !memcmp(cmpbuf, wrbuf_buf(b), wrbuf_len(b)))
+        ret = 1;
+    else
+        yaz_log(YLOG_LOG, "GOT (%.*s)", wrbuf_len(b), wrbuf_buf(b));
+    wrbuf_free(b, 1);
+    return ret;
+}
+
+static void tst_conversion_marc8_to_latin1()
+{
+    yaz_iconv_t cd = yaz_iconv_open("ISO-8859-1", "MARC8");
+
+    YAZ_CHECK(cd);
+    if (!cd)
+        return;
+
+    YAZ_CHECK(tst_convert(cd, "Cours de math", 
+                          "Cours de math"));
+    YAZ_CHECK(tst_convert(cd, "Cours de mathâe", 
+                          "Cours de mathé"));
+    YAZ_CHECK(tst_convert(cd, "12345678âe", 
+                          "12345678é"));
+    YAZ_CHECK(tst_convert(cd, "123456789âe", 
+                          "123456789é"));
+    YAZ_CHECK(tst_convert(cd, "1234567890âe", 
+                          "1234567890é"));
+    YAZ_CHECK(tst_convert(cd, "12345678901âe", 
+                          "12345678901é"));
+    YAZ_CHECK(tst_convert(cd, "Cours de mathâem", 
+                          "Cours de mathém"));
+    YAZ_CHECK(tst_convert(cd, "Cours de mathâematiques", 
+                          "Cours de mathématiques"));
+
+    yaz_iconv_close(cd);
+}
+
+static void tst_conversion_utf8_to_marc8()
+{
+    yaz_iconv_t cd = yaz_iconv_open("MARC8", "UTF-8");
+
+    YAZ_CHECK(cd);
+    if (!cd)
+        return;
+
+    YAZ_CHECK(tst_convert(cd, "Cours ", "Cours "));
+
+    /** Pure ASCII. 11 characters (sizeof(outbuf)-1) */
+    YAZ_CHECK(tst_convert(cd, "Cours de mat", "Cours de mat"));
+
+    /** Pure ASCII. 12 characters (sizeof(outbuf)) */
+    YAZ_CHECK(tst_convert(cd, "Cours de math", "Cours de math"));
+
+    /** Pure ASCII. 13 characters (sizeof(outbuf)) */
+    YAZ_CHECK(tst_convert(cd, "Cours de math.", "Cours de math."));
+
+    /** UPPERCASE SCANDINAVIAN O */
+    YAZ_CHECK(tst_convert(cd, "S\xc3\x98", "S\xa2"));
+
+    /** ARING */
+    YAZ_CHECK(tst_convert(cd, "A" "\xCC\x8A", "\xEA" "A"));
+
+    /** A MACRON + UMLAUT, DIAERESIS */
+    YAZ_CHECK(tst_convert(cd, "A" "\xCC\x84" "\xCC\x88",
+                          "\xE5\xE8\x41"));
+    
+    /* Ligature spanning two characters */
+    YAZ_CHECK(tst_convert(cd,
+                          "\x74" "\xCD\xA1" "\x73",  /* UTF-8 */
+                          "\xEB\x74\xEC\x73"));      /* MARC-8 */
+
+    /* Double title spanning two characters */
+    YAZ_CHECK(tst_convert(cd,
+                          "\x74" "\xCD\xA0" "\x73",  /* UTF-8 */
+                          "\xFA\x74\xFB\x73"));      /* MARC-8 */
+
+    /** Ideographic question mark (Unicode FF1F) */
+    YAZ_CHECK(tst_convert(cd,
+                          "\xEF\xBC\x9F" "o",        /* UTF-8 */
+                          "\033(1" "\x21\x2B\x3B" "\033(B" "o" ));
+
+    yaz_iconv_close(cd);
+}
+
+
+static void tst_conversion_latin1_to_marc8()
+{
+    yaz_iconv_t cd = yaz_iconv_open("MARC8", "ISO-8859-1");
+
+    YAZ_CHECK(cd);
+    if (!cd)
+        return;
+
+    YAZ_CHECK(tst_convert(cd, "Cours ", "Cours "));
+
+    /** Pure ASCII. 11 characters (sizeof(outbuf)-1) */
+    YAZ_CHECK(tst_convert(cd, "Cours de mat", "Cours de mat"));
+
+    /** Pure ASCII. 12 characters (sizeof(outbuf)) */
+    YAZ_CHECK(tst_convert(cd, "Cours de math", "Cours de math"));
+
+    /** Pure ASCII. 13 characters (sizeof(outbuf)) */
+    YAZ_CHECK(tst_convert(cd, "Cours de math.", "Cours de math."));
+
+    /** UPPERCASE SCANDINAVIAN O */
+    YAZ_CHECK(tst_convert(cd, "SØ", "S\xa2"));
+
+    yaz_iconv_close(cd);
+}
+
 int main (int argc, char **argv)
 {
     YAZ_CHECK_INIT(argc, argv);
 
+    tst_conversion_marc8_to_latin1();
+
+    tst_conversion_utf8_to_marc8();
+
+    tst_conversion_latin1_to_marc8();
+
     YAZ_CHECK(utf8_check(3));
     YAZ_CHECK(utf8_check(127));
     YAZ_CHECK(utf8_check(128));
@@ -362,7 +499,6 @@ int main (int argc, char **argv)
     YAZ_CHECK(utf8_check(1000));
     YAZ_CHECK(utf8_check(10000));
     YAZ_CHECK(utf8_check(100000));
-    YAZ_CHECK(utf8_check(100000));
     YAZ_CHECK(utf8_check(1000000));
     YAZ_CHECK(utf8_check(10000000));
     YAZ_CHECK(utf8_check(100000000));