X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Fmarcdisp.c;h=edabacdd626ad9b7d69d93c103ddd99ec4f5a37b;hb=5265b9beb2ccb4208c503ad60c56daa97a9c194d;hp=8473f9f0d29715a7b4a0d9edf65f8e3b0cee5505;hpb=96a3a7bc95bf4014fb1224a7e3ad711e5e20e785;p=yaz-moved-to-github.git diff --git a/src/marcdisp.c b/src/marcdisp.c index 8473f9f..edabacd 100644 --- a/src/marcdisp.c +++ b/src/marcdisp.c @@ -2,7 +2,7 @@ * Copyright (C) 1995-2005, Index Data ApS * See the file LICENSE for details. * - * $Id: marcdisp.c,v 1.14 2005-02-02 23:07:56 adam Exp $ + * $Id: marcdisp.c,v 1.17 2005-02-25 09:37:53 adam Exp $ */ /** @@ -72,6 +72,14 @@ static void marc_cdata (yaz_marc_t mt, const char *buf, size_t len, WRBUF wr) wrbuf_iconv_write_cdata(wr, mt->iconv_cd, buf, len); } +static int atoi_n_check(const char *buf, int size, int *val) +{ + if (!isdigit(*(const unsigned char *) buf)) + return 0; + *val = atoi_n(buf, size); + return 1; +} + int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) { int entry_p; @@ -82,6 +90,14 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) int length_data_entry; int length_starting; int length_implementation; + char lead[24]; + int produce_warnings = 0; + + if (mt->debug) + produce_warnings = 1; + if (mt->xml == YAZ_MARC_SIMPLEXML || mt->xml == YAZ_MARC_OAIMARC + || mt->xml == YAZ_MARC_MARCXML || mt->xml == YAZ_MARC_XCHANGE) + produce_warnings = 1; wrbuf_rewind(wr); @@ -92,37 +108,57 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) { char str[40]; - sprintf (str, "Record length %d - aborting\n", record_length); - wrbuf_puts (wr, str); + wrbuf_printf(wr, "\n", + record_length); } return -1; } + memcpy(lead, buf, 24); /* se can modify the header for output */ + /* ballout if bsize is known and record_length is less than that */ if (bsize != -1 && record_length > bsize) return -1; - if (isdigit(((const unsigned char *) buf)[10])) - indicator_length = atoi_n (buf+10, 1); - else - indicator_length = 2; - if (isdigit(((const unsigned char *) buf)[11])) - identifier_length = atoi_n (buf+11, 1); - else + if (!atoi_n_check(buf+10, 1, &indicator_length)) + { + if (produce_warnings) + wrbuf_printf(wr, "\n"); + lead[10] = '2'; + indicator_length = 2; + } + if (!atoi_n_check(buf+11, 1, &identifier_length)) + { + if (produce_warnings) + wrbuf_printf(wr, "\n"); + lead[11] = '2'; identifier_length = 2; - base_address = atoi_n (buf+12, 5); - - length_data_entry = atoi_n (buf+20, 1); - if (buf[20] <= '0' || buf[20] >= '9') + } + if (!atoi_n_check(buf+12, 5, &base_address)) + { + if (produce_warnings) + wrbuf_printf(wr, "\n"); + base_address = 0; + } + if (!atoi_n_check(buf+20, 1, &length_data_entry)) + { + if (produce_warnings) + wrbuf_printf(wr, "\n"); + length_data_entry = 4; + lead[20] = '4'; + } + if (!atoi_n_check(buf+21, 1, &length_starting)) { - wrbuf_printf(wr, "\n"); - length_data_entry = 4; + if (produce_warnings) + wrbuf_printf(wr, "\n"); + length_starting = 5; + lead[21] = '5'; } - length_starting = atoi_n (buf+21, 1); - if (buf[21] <= '0' || buf[21] >= '9') + if (!atoi_n_check(buf+22, 1, &length_implementation)) { - wrbuf_printf(wr, "\n"); - length_starting = 5; + if (produce_warnings) + wrbuf_printf(wr, "\n"); + length_implementation = 0; + lead[22] = '0'; } - length_implementation = atoi_n (buf+22, 1); if (mt->xml != YAZ_MARC_LINE) { @@ -165,13 +201,16 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) wr, "\n" " "); -#if 1 - marc_cdata(mt, buf, 9, wr); - marc_cdata(mt, "a", 1, wr); /* set leader to signal unicode */ - marc_cdata(mt, buf+10, 14, wr); -#else - marc_cdata(mt, buf, 24, wr); /* leave header as is .. */ -#endif + lead[9] = 'a'; /* set leader to signal unicode */ + marc_cdata(mt, lead, 24, wr); + wrbuf_printf(wr, "\n"); + break; + case YAZ_MARC_XCHANGE: + wrbuf_printf( + wr, + "\n" + " "); + marc_cdata(mt, lead, 24, wr); wrbuf_printf(wr, "\n"); break; } @@ -207,10 +246,11 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) if (entry_p >= record_length) return -1; } - if (mt->debug && base_address != entry_p+1) + if (base_address != entry_p+1) { - wrbuf_printf (wr," \n", base_address, entry_p+1); + if (produce_warnings) + wrbuf_printf (wr," \n", base_address, entry_p+1); } base_address = entry_p+1; @@ -237,8 +277,11 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) i = data_offset + base_address; end_offset = i+data_length-1; - while (buf[i] != ISO2709_RS && buf[i] != ISO2709_FS && - i < end_offset) + if (data_length <= 0 || data_offset < 0 || end_offset >= record_length) + return -1; + + while (i < end_offset && + buf[i] != ISO2709_RS && buf[i] != ISO2709_FS) i++; sz1 = 1+i - (data_offset + base_address); if (mt->iconv_cd) @@ -255,9 +298,9 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) } wrbuf_putc(wr_dir, ISO2709_FS); wrbuf_printf(wr_head, "%05d", data_p+1 + base_address); - wrbuf_write(wr_head, buf+5, 7); + wrbuf_write(wr_head, lead+5, 7); wrbuf_printf(wr_head, "%05d", base_address); - wrbuf_write(wr_head, buf+17, 7); + wrbuf_write(wr_head, lead+17, 7); wrbuf_write(wr, wrbuf_buf(wr_head), 24); wrbuf_write(wr, wrbuf_buf(wr_dir), wrbuf_len(wr_dir)); @@ -286,6 +329,15 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) entry_p += length_starting; i = data_offset + base_address; end_offset = i+data_length-1; + + if (data_length <= 0 || data_offset < 0 || end_offset >= record_length) + { + if (produce_warnings) + wrbuf_printf (wr," \n", + data_length, data_offset); + break; + } if (mt->debug) { @@ -293,16 +345,23 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) entry_p0, data_length, data_offset); } - if (indicator_length < 4 && indicator_length > 0) + if (memcmp (tag, "00", 2)) + identifier_flag = 1; /* if not 00X assume subfields */ + else if (indicator_length < 4 && indicator_length > 0) { + /* Danmarc 00X have subfields */ if (buf[i + indicator_length] == ISO2709_IDFS) identifier_flag = 1; else if (buf[i + indicator_length + 1] == ISO2709_IDFS) identifier_flag = 2; } - else if (memcmp (tag, "00", 2)) - identifier_flag = 1; - + + if (mt->debug) + { + wrbuf_printf(wr, "\n", + identifier_flag); + } + switch(mt->xml) { case YAZ_MARC_LINE: @@ -325,6 +384,7 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) wrbuf_printf(wr, "\""); break; case YAZ_MARC_MARCXML: + case YAZ_MARC_XCHANGE: if (identifier_flag) wrbuf_printf (wr, " xml == YAZ_MARC_SIMPLEXML || mt->xml == YAZ_MARC_MARCXML - || mt->xml == YAZ_MARC_OAIMARC) + || mt->xml == YAZ_MARC_OAIMARC || mt->xml == YAZ_MARC_XCHANGE) { wrbuf_puts (wr, ">"); if (identifier_flag) @@ -379,7 +440,8 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) } if (identifier_flag) { - while (buf[i] != ISO2709_RS && buf[i] != ISO2709_FS && i < end_offset) + while (i < end_offset && + buf[i] != ISO2709_RS && buf[i] != ISO2709_FS) { int i0; i++; @@ -410,6 +472,7 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) wrbuf_puts (wr, "\">"); break; case YAZ_MARC_MARCXML: + case YAZ_MARC_XCHANGE: wrbuf_puts (wr, " xml == YAZ_MARC_SIMPLEXML || mt->xml == YAZ_MARC_MARCXML || + mt->xml == YAZ_MARC_XCHANGE || mt->xml == YAZ_MARC_OAIMARC) wrbuf_puts (wr, "\n"); } @@ -434,7 +499,8 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) else { int i0 = i; - while (buf[i] != ISO2709_RS && buf[i] != ISO2709_FS && i < end_offset) + while (i < end_offset && + buf[i] != ISO2709_RS && buf[i] != ISO2709_FS) i++; marc_cdata(mt, buf + i0, i - i0, wr); if (mt->xml == YAZ_MARC_ISO2709) @@ -458,6 +524,7 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) wrbuf_puts (wr, "\n"); break; case YAZ_MARC_MARCXML: + case YAZ_MARC_XCHANGE: if (identifier_flag) wrbuf_puts (wr, " \n"); else @@ -477,6 +544,7 @@ int yaz_marc_decode_wrbuf (yaz_marc_t mt, const char *buf, int bsize, WRBUF wr) wrbuf_puts (wr, "\n"); break; case YAZ_MARC_MARCXML: + case YAZ_MARC_XCHANGE: wrbuf_puts (wr, "\n"); break; case YAZ_MARC_ISO2709: