From: Adam Dickmeiss Date: Mon, 15 Oct 2007 20:45:05 +0000 (+0000) Subject: Fixed bug #1778: Extended ASCII not escaped always (UTF-8 to MARC-8). X-Git-Tag: YAZ.3.0.16~57 X-Git-Url: http://lists.indexdata.com/cgi-bin?a=commitdiff_plain;h=02ba601656652e2362aee6fd23d1b2ac5c35bcca;p=yaz-moved-to-github.git Fixed bug #1778: Extended ASCII not escaped always (UTF-8 to MARC-8). When in non-ASCII mode, YAZ now escapes combined character sequences. A test was added in tsticonv.c. --- diff --git a/src/siconv.c b/src/siconv.c index 945b6cc..663d03a 100644 --- a/src/siconv.c +++ b/src/siconv.c @@ -2,7 +2,7 @@ * Copyright (C) 1995-2007, Index Data ApS * See the file LICENSE for details. * - * $Id: siconv.c,v 1.47 2007-10-12 14:22:19 adam Exp $ + * $Id: siconv.c,v 1.48 2007-10-15 20:45:05 adam Exp $ */ /** * \file siconv.c @@ -113,10 +113,9 @@ struct yaz_iconv_struct { #endif unsigned long compose_char; - unsigned long write_marc8_comb_ch[8]; - size_t write_marc8_comb_no; unsigned write_marc8_second_half_char; unsigned long write_marc8_last; + const char *write_marc8_lpage; const char *write_marc8_g0; const char *write_marc8_g1; }; @@ -192,6 +191,10 @@ static struct { { 0, 0, 0} }; +static size_t yaz_write_marc8_page_chr(yaz_iconv_t cd, + char **outbuf, size_t *outbytesleft, + const char *page_chr); + static unsigned long yaz_read_ISO8859_1 (yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read) { @@ -1465,12 +1468,7 @@ static unsigned long lookup_marc8(yaz_iconv_t cd, x = yaz_marc8r_45_conv(inp, inbytesleft, &no_read_sub, comb); if (x) { -#if 1 *page_chr = ESC "(B"; -#else - /* this possibly solves bug #1778 */ - *page_chr = ESC ")!E"; -#endif return x; } x = yaz_marc8r_67_conv(inp, inbytesleft, &no_read_sub, comb); @@ -1543,12 +1541,21 @@ static size_t flush_combos(yaz_iconv_t cd, { unsigned long y = cd->write_marc8_last; unsigned char byte; - char out_buf[10]; - size_t i, out_no = 0; + char out_buf[4]; + size_t out_no = 0; if (!y) return 0; + assert(cd->write_marc8_lpage); + if (cd->write_marc8_lpage) + { + size_t r = yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, + cd->write_marc8_lpage); + if (r) + return r; + } + byte = (unsigned char )((y>>16) & 0xff); if (byte) out_buf[out_no++] = byte; @@ -1559,19 +1566,12 @@ static size_t flush_combos(yaz_iconv_t cd, if (byte) out_buf[out_no++] = byte; - if (out_no + cd->write_marc8_comb_no + 1 > *outbytesleft) + if (out_no + 2 >= *outbytesleft) { cd->my_errno = YAZ_ICONV_E2BIG; return (size_t) (-1); } - for (i = 0; i < cd->write_marc8_comb_no; i++) - { - /* all MARC-8 combined characters are simple bytes */ - byte = (unsigned char )(cd->write_marc8_comb_ch[i]); - *(*outbuf)++ = byte; - (*outbytesleft)--; - } memcpy(*outbuf, out_buf, out_no); *outbuf += out_no; (*outbytesleft) -= out_no; @@ -1582,7 +1582,7 @@ static size_t flush_combos(yaz_iconv_t cd, } cd->write_marc8_last = 0; - cd->write_marc8_comb_no = 0; + cd->write_marc8_lpage = 0; cd->write_marc8_second_half_char = 0; return 0; } @@ -1652,7 +1652,8 @@ static size_t yaz_write_marc8_2(yaz_iconv_t cd, unsigned long x, { if (page_chr) { - size_t r = yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, page_chr); + size_t r = yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, + page_chr); if (r) return r; } @@ -1661,8 +1662,13 @@ static size_t yaz_write_marc8_2(yaz_iconv_t cd, unsigned long x, else if (x == 0x0360) cd->write_marc8_second_half_char = 0xFB; - if (cd->write_marc8_comb_no < 6) - cd->write_marc8_comb_ch[cd->write_marc8_comb_no++] = y; + if (*outbytesleft <= 1) + { + cd->my_errno = YAZ_ICONV_E2BIG; + return (size_t) (-1); + } + *(*outbuf)++ = y; + (*outbytesleft)--; } else { @@ -1670,13 +1676,8 @@ static size_t yaz_write_marc8_2(yaz_iconv_t cd, unsigned long x, if (r) return r; - if (page_chr) - { - r = yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, page_chr); - if (r) - return r; - } cd->write_marc8_last = y; + cd->write_marc8_lpage = page_chr; } return 0; } @@ -1704,6 +1705,7 @@ static size_t yaz_write_marc8(yaz_iconv_t cd, unsigned long x, char *outbuf0 = *outbuf; size_t outbytesleft0 = *outbytesleft; int last_ch = cd->write_marc8_last; + const char *lpage = cd->write_marc8_lpage; r = yaz_write_marc8_2(cd, latin1_comb[i].x1, outbuf, outbytesleft); @@ -1717,6 +1719,7 @@ static size_t yaz_write_marc8(yaz_iconv_t cd, unsigned long x, *outbuf = outbuf0; *outbytesleft = outbytesleft0; cd->write_marc8_last = last_ch; + cd->write_marc8_lpage = lpage; } return r; } @@ -1898,9 +1901,9 @@ size_t yaz_iconv(yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, cd->comb_offset = cd->comb_size = 0; cd->compose_char = 0; - cd->write_marc8_comb_no = 0; cd->write_marc8_second_half_char = 0; cd->write_marc8_last = 0; + cd->write_marc8_lpage = 0; cd->write_marc8_g0 = ESC "(B"; cd->write_marc8_g1 = 0; diff --git a/test/tsticonv.c b/test/tsticonv.c index 609a6b9..f1883b9 100644 --- a/test/tsticonv.c +++ b/test/tsticonv.c @@ -2,7 +2,7 @@ * Copyright (C) 1995-2007, Index Data ApS * See the file LICENSE for details. * - * $Id: tsticonv.c,v 1.31 2007-09-22 18:56:22 adam Exp $ + * $Id: tsticonv.c,v 1.32 2007-10-15 20:45:05 adam Exp $ */ #if HAVE_CONFIG_H @@ -134,7 +134,7 @@ static int tst_convert(yaz_iconv_t cd, const char *buf, const char *cmpbuf) wrbuf_rewind(w); wrbuf_verbose_str(w, cmpbuf, strlen(cmpbuf)); - yaz_log(YLOG_LOG, "expected %s", wrbuf_cstr(w)); + yaz_log(YLOG_LOG, "exp %s", wrbuf_cstr(w)); wrbuf_destroy(w); } @@ -515,7 +515,15 @@ static void tst_utf8_to_marc8(void) "(\033p0\x1bs)")); - + /** bug #1778 */ + YAZ_CHECK(tst_convert(cd, + /* offset 0x530 in UTF-8 rec marccol4.u8.marc */ + "\xE3\x83\xB3" "\xE3\x82\xBF" + "\xCC\x84" "\xCC\x84" "\xE3\x83\xBC" /* UTF-8 */, + "\x1B\x24\x31" "\x69\x25\x73" + "\x1B\x28\x42" "\xE5\xE5" "\x1B\x24\x31" + "\x69\x25\x3F" + "\x69\x21\x3C" "\x1B\x28\x42")); { char *inbuf0 = "\xe2\x81\xb0";