From 65f1b9fce2d1d86df7d75645fb46fbba32dd01e3 Mon Sep 17 00:00:00 2001 From: Adam Dickmeiss Date: Tue, 1 Apr 2008 13:37:51 +0200 Subject: [PATCH] Factored character encoders out to separate files (iconv system). --- src/Makefile.am | 3 +- src/advancegreek.c | 20 +- src/iconv-p.h | 49 ++- src/iconv_encode_iso_8859_1.c | 255 +++++++++++++++ src/iconv_encode_marc8.c | 441 ++++++++++++++++++++++++++ src/iconv_encode_wchar.c | 78 +++++ src/iso5428.c | 20 +- src/siconv.c | 685 ++++------------------------------------- src/ucs4.c | 27 +- src/utf8.c | 20 +- win/makefile | 5 +- 11 files changed, 939 insertions(+), 664 deletions(-) create mode 100644 src/iconv_encode_iso_8859_1.c create mode 100644 src/iconv_encode_marc8.c create mode 100644 src/iconv_encode_wchar.c diff --git a/src/Makefile.am b/src/Makefile.am index b1dd090..23a5bfa 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -95,7 +95,8 @@ libyaz_la_SOURCES=version.c options.c log.c \ xmlquery.c xmlerror.c http.c \ mime.c mime.h oid_util.c tokenizer.c \ record_conv.c retrieval.c elementset.c snprintf.c query-charset.c \ - copy_types.c match_glob.c poll.c daemon.c + copy_types.c match_glob.c poll.c daemon.c \ + iconv_encode_marc8.c iconv_encode_iso_8859_1.c iconv_encode_wchar.c libyaz_la_LDFLAGS=-version-info $(YAZ_VERSION_INFO) diff --git a/src/advancegreek.c b/src/advancegreek.c index 6530f2d..56f3dc0 100644 --- a/src/advancegreek.c +++ b/src/advancegreek.c @@ -1,12 +1,10 @@ /* * Copyright (C) 1995-2008, Index Data ApS * See the file LICENSE for details. - * - * $Id: siconv.c,v 1.50 2008-03-12 08:53:28 adam Exp $ */ /** * \file - * \brief ISO-5428 character mapping (iconv) + * \brief Advance Greek encoding and decoding */ #if HAVE_CONFIG_H @@ -279,8 +277,9 @@ unsigned long yaz_read_advancegreek(yaz_iconv_t cd, unsigned char *inp, return x; } -size_t yaz_write_advancegreek(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static size_t write_advancegreek(yaz_iconv_t cd, yaz_iconv_encoder_t w, + unsigned long x, + char **outbuf, size_t *outbytesleft) { size_t k = 0; unsigned char *out = (unsigned char*) *outbuf; @@ -374,6 +373,17 @@ size_t yaz_write_advancegreek(yaz_iconv_t cd, unsigned long x, return 0; } +yaz_iconv_encoder_t yaz_advancegreek_encoder(const char *name, + yaz_iconv_encoder_t e) +{ + if (!yaz_matchstr(name, "advancegreek")) + { + e->write_handle = write_advancegreek; + return e; + } + return 0; +} + /* * Local variables: * c-basic-offset: 4 diff --git a/src/iconv-p.h b/src/iconv-p.h index 04a51b0..f9cc91e 100644 --- a/src/iconv-p.h +++ b/src/iconv-p.h @@ -2,7 +2,6 @@ * Copyright (C) 2005-2008, Index Data ApS * See the file LICENSE for details. * - * $Id: zoom-p.h,v 1.25 2007-09-11 08:40:28 adam Exp $ */ /** * \file @@ -21,30 +20,54 @@ void yaz_iconv_set_errno(yaz_iconv_t cd, int no); unsigned long yaz_read_iso5428_1984(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); -size_t yaz_write_iso5428_1984(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); - size_t yaz_init_UTF8(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); unsigned long yaz_read_UTF8(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); -size_t yaz_write_UTF8(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); - unsigned long yaz_read_UCS4(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); unsigned long yaz_read_UCS4LE(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); -size_t yaz_write_UCS4(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); -size_t yaz_write_UCS4LE(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); unsigned long yaz_read_advancegreek(yaz_iconv_t cd, unsigned char *inp, size_t inbytesleft, size_t *no_read); -size_t yaz_write_advancegreek(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); + +typedef struct yaz_iconv_encoder_s *yaz_iconv_encoder_t; +struct yaz_iconv_encoder_s { + void *data; + size_t (*write_handle)(yaz_iconv_t cd, yaz_iconv_encoder_t e, + unsigned long x, + char **outbuf, size_t *outbytesleft); + size_t (*flush_handle)(yaz_iconv_t cd, yaz_iconv_encoder_t e, + char **outbuf, size_t *outbytesleft); + void (*init_handle)(yaz_iconv_encoder_t e); + void (*destroy_handle)(yaz_iconv_encoder_t e); +}; + +yaz_iconv_encoder_t yaz_marc8_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_utf8_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_ucs4_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_iso_8859_1_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_iso_5428_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_advancegreek_encoder(const char *name, + yaz_iconv_encoder_t e); +yaz_iconv_encoder_t yaz_wchar_encoder(const char *name, + yaz_iconv_encoder_t e); +typedef unsigned long yaz_conv_func_t(unsigned char *inp, size_t inbytesleft, + size_t *no_read, int *combining, + unsigned mask, int boffset); + +int yaz_iso_8859_1_lookup_y(unsigned long v, + unsigned long *x1, unsigned long *x2); + +int yaz_iso_8859_1_lookup_x12(unsigned long x1, unsigned long x2, + unsigned long *y); #endif /* diff --git a/src/iconv_encode_iso_8859_1.c b/src/iconv_encode_iso_8859_1.c new file mode 100644 index 0000000..f847186 --- /dev/null +++ b/src/iconv_encode_iso_8859_1.c @@ -0,0 +1,255 @@ +/* + * Copyright (C) 1995-2008, Index Data ApS + * See the file LICENSE for details. + * + */ +/** + * \file + * \brief ISO-8859-1 encoding / decoding + * + */ + +#if HAVE_CONFIG_H +#include +#endif + +#include +#include +#include +#include + +#if HAVE_ICONV_H +#include +#endif + +#include +#include +#include "iconv-p.h" + +struct encoder_data +{ + unsigned long compose_char; +}; + + + +static struct { + unsigned long x1, x2; + unsigned y; +} latin1_comb[] = { + { 'A', 0x0300, 0xc0}, /* LATIN CAPITAL LETTER A WITH GRAVE */ + { 'A', 0x0301, 0xc1}, /* LATIN CAPITAL LETTER A WITH ACUTE */ + { 'A', 0x0302, 0xc2}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ + { 'A', 0x0303, 0xc3}, /* LATIN CAPITAL LETTER A WITH TILDE */ + { 'A', 0x0308, 0xc4}, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ + { 'A', 0x030a, 0xc5}, /* LATIN CAPITAL LETTER A WITH RING ABOVE */ + /* no need for 0xc6 LATIN CAPITAL LETTER AE */ + { 'C', 0x0327, 0xc7}, /* LATIN CAPITAL LETTER C WITH CEDILLA */ + { 'E', 0x0300, 0xc8}, /* LATIN CAPITAL LETTER E WITH GRAVE */ + { 'E', 0x0301, 0xc9}, /* LATIN CAPITAL LETTER E WITH ACUTE */ + { 'E', 0x0302, 0xca}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX */ + { 'E', 0x0308, 0xcb}, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ + { 'I', 0x0300, 0xcc}, /* LATIN CAPITAL LETTER I WITH GRAVE */ + { 'I', 0x0301, 0xcd}, /* LATIN CAPITAL LETTER I WITH ACUTE */ + { 'I', 0x0302, 0xce}, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ + { 'I', 0x0308, 0xcf}, /* LATIN CAPITAL LETTER I WITH DIAERESIS */ + { 'N', 0x0303, 0xd1}, /* LATIN CAPITAL LETTER N WITH TILDE */ + { 'O', 0x0300, 0xd2}, /* LATIN CAPITAL LETTER O WITH GRAVE */ + { 'O', 0x0301, 0xd3}, /* LATIN CAPITAL LETTER O WITH ACUTE */ + { 'O', 0x0302, 0xd4}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ + { 'O', 0x0303, 0xd5}, /* LATIN CAPITAL LETTER O WITH TILDE */ + { 'O', 0x0308, 0xd6}, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ + /* omitted: 0xd7 MULTIPLICATION SIGN */ + /* omitted: 0xd8 LATIN CAPITAL LETTER O WITH STROKE */ + { 'U', 0x0300, 0xd9}, /* LATIN CAPITAL LETTER U WITH GRAVE */ + { 'U', 0x0301, 0xda}, /* LATIN CAPITAL LETTER U WITH ACUTE */ + { 'U', 0x0302, 0xdb}, /* LATIN CAPITAL LETTER U WITH CIRCUMFLEX */ + { 'U', 0x0308, 0xdc}, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ + { 'Y', 0x0301, 0xdd}, /* LATIN CAPITAL LETTER Y WITH ACUTE */ + /* omitted: 0xde LATIN CAPITAL LETTER THORN */ + /* omitted: 0xdf LATIN SMALL LETTER SHARP S */ + { 'a', 0x0300, 0xe0}, /* LATIN SMALL LETTER A WITH GRAVE */ + { 'a', 0x0301, 0xe1}, /* LATIN SMALL LETTER A WITH ACUTE */ + { 'a', 0x0302, 0xe2}, /* LATIN SMALL LETTER A WITH CIRCUMFLEX */ + { 'a', 0x0303, 0xe3}, /* LATIN SMALL LETTER A WITH TILDE */ + { 'a', 0x0308, 0xe4}, /* LATIN SMALL LETTER A WITH DIAERESIS */ + { 'a', 0x030a, 0xe5}, /* LATIN SMALL LETTER A WITH RING ABOVE */ + /* omitted: 0xe6 LATIN SMALL LETTER AE */ + { 'c', 0x0327, 0xe7}, /* LATIN SMALL LETTER C WITH CEDILLA */ + { 'e', 0x0300, 0xe8}, /* LATIN SMALL LETTER E WITH GRAVE */ + { 'e', 0x0301, 0xe9}, /* LATIN SMALL LETTER E WITH ACUTE */ + { 'e', 0x0302, 0xea}, /* LATIN SMALL LETTER E WITH CIRCUMFLEX */ + { 'e', 0x0308, 0xeb}, /* LATIN SMALL LETTER E WITH DIAERESIS */ + { 'i', 0x0300, 0xec}, /* LATIN SMALL LETTER I WITH GRAVE */ + { 'i', 0x0301, 0xed}, /* LATIN SMALL LETTER I WITH ACUTE */ + { 'i', 0x0302, 0xee}, /* LATIN SMALL LETTER I WITH CIRCUMFLEX */ + { 'i', 0x0308, 0xef}, /* LATIN SMALL LETTER I WITH DIAERESIS */ + /* omitted: 0xf0 LATIN SMALL LETTER ETH */ + { 'n', 0x0303, 0xf1}, /* LATIN SMALL LETTER N WITH TILDE */ + { 'o', 0x0300, 0xf2}, /* LATIN SMALL LETTER O WITH GRAVE */ + { 'o', 0x0301, 0xf3}, /* LATIN SMALL LETTER O WITH ACUTE */ + { 'o', 0x0302, 0xf4}, /* LATIN SMALL LETTER O WITH CIRCUMFLEX */ + { 'o', 0x0303, 0xf5}, /* LATIN SMALL LETTER O WITH TILDE */ + { 'o', 0x0308, 0xf6}, /* LATIN SMALL LETTER O WITH DIAERESIS */ + /* omitted: 0xf7 DIVISION SIGN */ + /* omitted: 0xf8 LATIN SMALL LETTER O WITH STROKE */ + { 'u', 0x0300, 0xf9}, /* LATIN SMALL LETTER U WITH GRAVE */ + { 'u', 0x0301, 0xfa}, /* LATIN SMALL LETTER U WITH ACUTE */ + { 'u', 0x0302, 0xfb}, /* LATIN SMALL LETTER U WITH CIRCUMFLEX */ + { 'u', 0x0308, 0xfc}, /* LATIN SMALL LETTER U WITH DIAERESIS */ + { 'y', 0x0301, 0xfd}, /* LATIN SMALL LETTER Y WITH ACUTE */ + /* omitted: 0xfe LATIN SMALL LETTER THORN */ + { 'y', 0x0308, 0xff}, /* LATIN SMALL LETTER Y WITH DIAERESIS */ + + { 0, 0, 0} +}; + +int yaz_iso_8859_1_lookup_y(unsigned long v, + unsigned long *x1, unsigned long *x2) +{ + if (v >= 0xc0 && v <= 0xff) /* optimization. min and max .y values */ + { + int i; + for (i = 0; latin1_comb[i].x1; i++) + { + if (v == latin1_comb[i].y) + { + *x1 = latin1_comb[i].x1; + *x2 = latin1_comb[i].x2; + return 1; + } + } + } + return 0; +} + +int yaz_iso_8859_1_lookup_x12(unsigned long x1, unsigned long x2, + unsigned long *y) +{ + /* For MARC8s we try to get a Latin-1 page code out of it */ + int i; + for (i = 0; latin1_comb[i].x1; i++) + if (x2 == latin1_comb[i].x2 && x1 == latin1_comb[i].x1) + { + *y = latin1_comb[i].y; + return 1; + } + return 0; +} + +static size_t write_iso_8859_1(yaz_iconv_t cd, yaz_iconv_encoder_t e, + unsigned long x, + char **outbuf, size_t *outbytesleft) +{ + struct encoder_data *w = e->data; + /* list of two char unicode sequence that, when combined, are + equivalent to single unicode chars that can be represented in + ISO-8859-1/Latin-1. + Regular iconv on Linux at least does not seem to convert these, + but since MARC-8 to UTF-8 generates these composed sequence + we get a better chance of a successful MARC-8 -> ISO-8859-1 + conversion */ + unsigned char *outp = (unsigned char *) *outbuf; + + if (w->compose_char) + { + int i; + for (i = 0; latin1_comb[i].x1; i++) + if (w->compose_char == latin1_comb[i].x1 && x == latin1_comb[i].x2) + { + x = latin1_comb[i].y; + break; + } + if (*outbytesleft < 1) + { /* no room. Retain compose_char and bail out */ + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t)(-1); + } + if (!latin1_comb[i].x1) + { /* not found. Just write compose_char */ + *outp++ = (unsigned char) w->compose_char; + (*outbytesleft)--; + *outbuf = (char *) outp; + } + /* compose_char used so reset it. x now holds current char */ + w->compose_char = 0; + } + + if (x > 32 && x < 127 && w->compose_char == 0) + { + w->compose_char = x; + return 0; + } + else if (x > 255 || x < 1) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_EILSEQ); + return (size_t) -1; + } + else if (*outbytesleft < 1) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t)(-1); + } + *outp++ = (unsigned char) x; + (*outbytesleft)--; + *outbuf = (char *) outp; + return 0; +} + +static size_t flush_iso_8859_1(yaz_iconv_t cd, yaz_iconv_encoder_t e, + char **outbuf, size_t *outbytesleft) +{ + struct encoder_data *w = e->data; + if (w->compose_char) + { + unsigned char *outp = (unsigned char *) *outbuf; + if (*outbytesleft < 1) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t)(-1); + } + *outp++ = (unsigned char) w->compose_char; + (*outbytesleft)--; + *outbuf = (char *) outp; + w->compose_char = 0; + } + return 0; +} + + +void init_iso_8859_1(yaz_iconv_encoder_t e) +{ + struct encoder_data *w = e->data; + w->compose_char = 0; +} + +void destroy_iso_8859_1(yaz_iconv_encoder_t e) +{ + xfree(e->data); +} + +yaz_iconv_encoder_t yaz_iso_8859_1_encoder(const char *tocode, + yaz_iconv_encoder_t e) + +{ + if (!yaz_matchstr(tocode, "iso88591")) + { + struct encoder_data *data = xmalloc(sizeof(*data)); + e->data = data; + e->write_handle = write_iso_8859_1; + e->flush_handle = flush_iso_8859_1; + e->init_handle = init_iso_8859_1; + e->destroy_handle = destroy_iso_8859_1; + return e; + } + return 0; +} + + +/* + * Local variables: + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ diff --git a/src/iconv_encode_marc8.c b/src/iconv_encode_marc8.c new file mode 100644 index 0000000..c4e95a4 --- /dev/null +++ b/src/iconv_encode_marc8.c @@ -0,0 +1,441 @@ +/* + * Copyright (C) 1995-2008, Index Data ApS + * See the file LICENSE for details. + * + */ +/** + * \file + * \brief MARC-8 encoding + * + * MARC-8 reference: + * http://www.loc.gov/marc/specifications/speccharmarc8.html + */ + +#if HAVE_CONFIG_H +#include +#endif + +#include +#include +#include +#include + +#include +#include +#include +#include "iconv-p.h" + +yaz_conv_func_t yaz_marc8r_42_conv; +yaz_conv_func_t yaz_marc8r_45_conv; +yaz_conv_func_t yaz_marc8r_67_conv; +yaz_conv_func_t yaz_marc8r_62_conv; +yaz_conv_func_t yaz_marc8r_70_conv; +yaz_conv_func_t yaz_marc8r_32_conv; +yaz_conv_func_t yaz_marc8r_4E_conv; +yaz_conv_func_t yaz_marc8r_51_conv; +yaz_conv_func_t yaz_marc8r_33_conv; +yaz_conv_func_t yaz_marc8r_34_conv; +yaz_conv_func_t yaz_marc8r_53_conv; +yaz_conv_func_t yaz_marc8r_31_conv; + +#define ESC "\033" + +struct encoder_data +{ + unsigned write_marc8_second_half_char; + unsigned long write_marc8_last; + int write_marc8_ncr; + const char *write_marc8_lpage; + const char *write_marc8_g0; + const char *write_marc8_g1; +}; + +static void init_marc8(yaz_iconv_encoder_t w) +{ + struct encoder_data *data = w->data; + data->write_marc8_second_half_char = 0; + data->write_marc8_last = 0; + data->write_marc8_ncr = 0; + data->write_marc8_lpage = 0; + data->write_marc8_g0 = ESC "(B"; + data->write_marc8_g1 = 0; +} + +static size_t yaz_write_marc8_page_chr(yaz_iconv_t cd, + struct encoder_data *w, + char **outbuf, size_t *outbytesleft, + const char *page_chr); + +static unsigned long lookup_marc8(yaz_iconv_t cd, + unsigned long x, int *comb, + const char **page_chr) +{ + char utf8_buf[7]; + char *utf8_outbuf = utf8_buf; + size_t utf8_outbytesleft = sizeof(utf8_buf)-1, r; + int error_code; + + r = yaz_write_UTF8_char(x, &utf8_outbuf, &utf8_outbytesleft, &error_code); + if (r == (size_t)(-1)) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_EILSEQ); + return 0; + } + else + { + unsigned char *inp; + size_t inbytesleft, no_read_sub = 0; + unsigned long x; + + *utf8_outbuf = '\0'; + inp = (unsigned char *) utf8_buf; + inbytesleft = strlen(utf8_buf); + + x = yaz_marc8r_42_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(B"; + return x; + } + x = yaz_marc8r_45_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(B"; + return x; + } + x = yaz_marc8r_62_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "b"; + return x; + } + x = yaz_marc8r_70_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "p"; + return x; + } + x = yaz_marc8r_32_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(2"; + return x; + } + x = yaz_marc8r_4E_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(N"; + return x; + } + x = yaz_marc8r_51_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(Q"; + return x; + } + x = yaz_marc8r_33_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(3"; + return x; + } + x = yaz_marc8r_34_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(4"; + return x; + } + x = yaz_marc8r_53_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "(S"; + return x; + } + x = yaz_marc8r_31_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); + if (x) + { + *page_chr = ESC "$1"; + return x; + } + yaz_iconv_set_errno(cd, YAZ_ICONV_EILSEQ); + return x; + } +} + +static size_t flush_combos(yaz_iconv_t cd, + struct encoder_data *w, + char **outbuf, size_t *outbytesleft) +{ + unsigned long y = w->write_marc8_last; + + if (!y) + return 0; + + assert(w->write_marc8_lpage); + if (w->write_marc8_lpage) + { + size_t r = yaz_write_marc8_page_chr(cd, w, outbuf, outbytesleft, + w->write_marc8_lpage); + if (r) + return r; + } + + if (9 >= *outbytesleft) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t) (-1); + } + if (w->write_marc8_ncr) + { + yaz_snprintf(*outbuf, 9, "&#x%04x;", y); + (*outbytesleft) -= 8; + (*outbuf) += 8; + } + else + { + size_t out_no = 0; + unsigned char byte; + + byte = (unsigned char )((y>>16) & 0xff); + if (byte) + (*outbuf)[out_no++] = byte; + byte = (unsigned char)((y>>8) & 0xff); + if (byte) + (*outbuf)[out_no++] = byte; + byte = (unsigned char )(y & 0xff); + if (byte) + (*outbuf)[out_no++] = byte; + *outbuf += out_no; + (*outbytesleft) -= out_no; + } + + if (w->write_marc8_second_half_char) + { + *(*outbuf)++ = w->write_marc8_second_half_char; + (*outbytesleft)--; + } + + w->write_marc8_last = 0; + w->write_marc8_ncr = 0; + w->write_marc8_lpage = 0; + w->write_marc8_second_half_char = 0; + return 0; +} + +static size_t yaz_write_marc8_page_chr(yaz_iconv_t cd, + struct encoder_data *w, + char **outbuf, size_t *outbytesleft, + const char *page_chr) +{ + const char **old_page_chr = &w->write_marc8_g0; + + /* are we going to a G1-set (such as such as ESC ")!E") */ + if (page_chr && page_chr[1] == ')') + old_page_chr = &w->write_marc8_g1; + + if (!*old_page_chr || strcmp(page_chr, *old_page_chr)) + { + size_t plen = 0; + const char *page_out = page_chr; + + if (*outbytesleft < 8) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + + return (size_t) (-1); + } + + if (*old_page_chr) + { + if (!strcmp(*old_page_chr, ESC "p") + || !strcmp(*old_page_chr, ESC "g") + || !strcmp(*old_page_chr, ESC "b")) + { + page_out = ESC "s"; + /* Technique 1 leave */ + if (strcmp(page_chr, ESC "(B")) /* Not going ASCII page? */ + { + /* Must leave script + enter new page */ + plen = strlen(page_out); + memcpy(*outbuf, page_out, plen); + (*outbuf) += plen; + (*outbytesleft) -= plen; + page_out = ESC "(B"; + } + } + } + *old_page_chr = page_chr; + plen = strlen(page_out); + memcpy(*outbuf, page_out, plen); + (*outbuf) += plen; + (*outbytesleft) -= plen; + } + return 0; +} + + +static size_t yaz_write_marc8_2(yaz_iconv_t cd, struct encoder_data *w, + unsigned long x, + char **outbuf, size_t *outbytesleft, + int loss_mode) +{ + int comb = 0; + int enable_ncr = 0; + const char *page_chr = 0; + unsigned long y = lookup_marc8(cd, x, &comb, &page_chr); + + if (!y) + { + if (loss_mode == 0) + return (size_t) (-1); + page_chr = ESC "(B"; + if (loss_mode == 1) + y = '|'; + else + { + y = x; + enable_ncr = 1; + } + } + + if (comb) + { + if (page_chr) + { + size_t r = yaz_write_marc8_page_chr(cd, w, outbuf, outbytesleft, + page_chr); + if (r) + return r; + } + if (x == 0x0361) + w->write_marc8_second_half_char = 0xEC; + else if (x == 0x0360) + w->write_marc8_second_half_char = 0xFB; + + if (*outbytesleft <= 1) + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t) (-1); + } + *(*outbuf)++ = y; + (*outbytesleft)--; + } + else + { + size_t r = flush_combos(cd, w, outbuf, outbytesleft); + if (r) + return r; + + w->write_marc8_last = y; + w->write_marc8_lpage = page_chr; + w->write_marc8_ncr = enable_ncr; + } + return 0; +} + +static size_t flush_marc8(yaz_iconv_t cd, yaz_iconv_encoder_t en, + char **outbuf, size_t *outbytesleft) +{ + struct encoder_data *w = en->data; + size_t r = flush_combos(cd, w, outbuf, outbytesleft); + if (r) + return r; + w->write_marc8_g1 = 0; + return yaz_write_marc8_page_chr(cd, w, outbuf, outbytesleft, ESC "(B"); +} + +static size_t yaz_write_marc8_generic(yaz_iconv_t cd, struct encoder_data *w, + unsigned long x, + char **outbuf, size_t *outbytesleft, + int loss_mode) +{ + unsigned long x1, x2; + if (yaz_iso_8859_1_lookup_y(x, &x1, &x2)) + { + /* save the output pointers .. */ + char *outbuf0 = *outbuf; + size_t outbytesleft0 = *outbytesleft; + int last_ch = w->write_marc8_last; + int ncr = w->write_marc8_ncr; + const char *lpage = w->write_marc8_lpage; + size_t r; + + r = yaz_write_marc8_2(cd, w, x1, + outbuf, outbytesleft, loss_mode); + if (r) + return r; + r = yaz_write_marc8_2(cd, w, x2, + outbuf, outbytesleft, loss_mode); + if (r && yaz_iconv_error(cd) == YAZ_ICONV_E2BIG) + { + /* not enough room. reset output to original values */ + *outbuf = outbuf0; + *outbytesleft = outbytesleft0; + w->write_marc8_last = last_ch; + w->write_marc8_ncr = ncr; + w->write_marc8_lpage = lpage; + } + return r; + } + return yaz_write_marc8_2(cd, w, x, outbuf, outbytesleft, loss_mode); +} + +static size_t write_marc8_normal(yaz_iconv_t cd, yaz_iconv_encoder_t e, + unsigned long x, + char **outbuf, size_t *outbytesleft) +{ + return yaz_write_marc8_generic(cd, e->data, x, outbuf, outbytesleft, 0); +} + +static size_t write_marc8_lossy(yaz_iconv_t cd, yaz_iconv_encoder_t e, + unsigned long x, + char **outbuf, size_t *outbytesleft) +{ + return yaz_write_marc8_generic(cd, e->data, x, outbuf, outbytesleft, 1); +} + +static size_t write_marc8_lossless(yaz_iconv_t cd, yaz_iconv_encoder_t e, + unsigned long x, + char **outbuf, size_t *outbytesleft) +{ + return yaz_write_marc8_generic(cd, e->data, x, outbuf, outbytesleft, 2); +} + +static void destroy_marc8(yaz_iconv_encoder_t e) +{ + xfree(e->data); +} + +yaz_iconv_encoder_t yaz_marc8_encoder(const char *tocode, + yaz_iconv_encoder_t e) + +{ + if (!yaz_matchstr(tocode, "MARC8")) + e->write_handle = write_marc8_normal; + else if (!yaz_matchstr(tocode, "MARC8s")) + e->write_handle = write_marc8_normal; + else if (!yaz_matchstr(tocode, "MARC8lossy")) + e->write_handle = write_marc8_lossy; + else if (!yaz_matchstr(tocode, "MARC8lossless")) + e->write_handle = write_marc8_lossless; + else + return 0; + + { + struct encoder_data *data = xmalloc(sizeof(*data)); + e->data = data; + e->destroy_handle = destroy_marc8; + e->flush_handle = flush_marc8; + e->init_handle = init_marc8; + } + return e; +} + + +/* + * Local variables: + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ diff --git a/src/iconv_encode_wchar.c b/src/iconv_encode_wchar.c new file mode 100644 index 0000000..764988f --- /dev/null +++ b/src/iconv_encode_wchar.c @@ -0,0 +1,78 @@ +/* + * Copyright (C) 1995-2008, Index Data ApS + * See the file LICENSE for details. + * + */ +/** + * \file + * \brief WCHAR_T iconv encoding / decoding + */ + +#if HAVE_CONFIG_H +#include +#endif + +#include +#include +#include +#include +#if HAVE_WCHAR_H +#include +#endif + +#include +#include +#include +#include "iconv-p.h" + +struct encoder_data +{ + unsigned long compose_char; +}; + +#if HAVE_WCHAR_H +static size_t write_wchar_t(yaz_iconv_t cd, yaz_iconv_encoder_t en, + 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 + { + yaz_iconv_set_errno(cd, YAZ_ICONV_E2BIG); + return (size_t)(-1); + } + *outbuf = (char *) outp; + return 0; +} +#endif + +yaz_iconv_encoder_t yaz_wchar_encoder(const char *tocode, + yaz_iconv_encoder_t e) + +{ +#if HAVE_WCHAR_H + if (!yaz_matchstr(tocode, "wchar_t")) + { + e->write_handle = write_wchar_t; + return e; + } +#endif + return 0; +} + + +/* + * Local variables: + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ diff --git a/src/iso5428.c b/src/iso5428.c index a78b00d..de8044b 100644 --- a/src/iso5428.c +++ b/src/iso5428.c @@ -2,11 +2,10 @@ * Copyright (C) 1995-2008, Index Data ApS * See the file LICENSE for details. * - * $Id: siconv.c,v 1.50 2008-03-12 08:53:28 adam Exp $ */ /** * \file - * \brief ISO-5428 character mapping (iconv) + * \brief ISO-5428:1984 encoding and decoding */ #if HAVE_CONFIG_H @@ -270,8 +269,9 @@ unsigned long yaz_read_iso5428_1984(yaz_iconv_t cd, unsigned char *inp, return x; } -size_t yaz_write_iso5428_1984(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static size_t write_iso_5428_1984(yaz_iconv_t cd, yaz_iconv_encoder_t en, + unsigned long x, + char **outbuf, size_t *outbytesleft) { size_t k = 0; unsigned char *out = (unsigned char*) *outbuf; @@ -365,6 +365,18 @@ size_t yaz_write_iso5428_1984(yaz_iconv_t cd, unsigned long x, return 0; } +yaz_iconv_encoder_t yaz_iso_5428_encoder(const char *name, + yaz_iconv_encoder_t e) +{ + if (!yaz_matchstr(name, "iso54281984") + || !yaz_matchstr(name, "iso5428:1984")) + { + e->write_handle = write_iso_5428_1984; + return e; + } + return 0; +} + /* * Local variables: diff --git a/src/siconv.c b/src/siconv.c index de3c54e..e0f3514 100644 --- a/src/siconv.c +++ b/src/siconv.c @@ -1,11 +1,9 @@ /* * Copyright (C) 1995-2008, Index Data ApS * See the file LICENSE for details. - * - * $Id: siconv.c,v 1.50 2008-03-12 08:53:28 adam Exp $ */ /** - * \file siconv.c + * \file * \brief Implements simple ICONV * * This implements an interface similar to that of iconv and @@ -13,8 +11,6 @@ * For systems where iconv is not present, this layer * provides a few important conversions: UTF-8, MARC-8, Latin-1. * - * MARC-8 reference: - * http://www.loc.gov/marc/specifications/speccharmarc8.html */ #if HAVE_CONFIG_H @@ -35,14 +31,8 @@ #include #include -#include #include "iconv-p.h" -typedef unsigned long yaz_conv_func_t(unsigned char *inp, size_t inbytesleft, - size_t *no_read, int *combining, - unsigned mask, int boffset); - - yaz_conv_func_t yaz_marc8_42_conv; yaz_conv_func_t yaz_marc8_45_conv; yaz_conv_func_t yaz_marc8_67_conv; @@ -56,30 +46,13 @@ yaz_conv_func_t yaz_marc8_34_conv; yaz_conv_func_t yaz_marc8_53_conv; yaz_conv_func_t yaz_marc8_31_conv; -yaz_conv_func_t yaz_marc8r_42_conv; -yaz_conv_func_t yaz_marc8r_45_conv; -yaz_conv_func_t yaz_marc8r_67_conv; -yaz_conv_func_t yaz_marc8r_62_conv; -yaz_conv_func_t yaz_marc8r_70_conv; -yaz_conv_func_t yaz_marc8r_32_conv; -yaz_conv_func_t yaz_marc8r_4E_conv; -yaz_conv_func_t yaz_marc8r_51_conv; -yaz_conv_func_t yaz_marc8r_33_conv; -yaz_conv_func_t yaz_marc8r_34_conv; -yaz_conv_func_t yaz_marc8r_53_conv; -yaz_conv_func_t yaz_marc8r_31_conv; - struct yaz_iconv_struct { int my_errno; int init_flag; size_t (*init_handle)(yaz_iconv_t cd, unsigned char *inbuf, - size_t inbytesleft, size_t *no_read); + size_t inbytesleft, size_t *no_read); unsigned long (*read_handle)(yaz_iconv_t cd, unsigned char *inbuf, size_t inbytesleft, size_t *no_read); - size_t (*write_handle)(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft); - size_t (*flush_handle)(yaz_iconv_t cd, - char **outbuf, size_t *outbytesleft); int g0_mode; int g1_mode; @@ -92,94 +65,10 @@ struct yaz_iconv_struct { #if HAVE_ICONV_H iconv_t iconv_cd; #endif - unsigned long compose_char; - - unsigned write_marc8_second_half_char; - unsigned long write_marc8_last; - int write_marc8_ncr; - const char *write_marc8_lpage; - const char *write_marc8_g0; - const char *write_marc8_g1; + struct yaz_iconv_encoder_s encoder; }; -static struct { - unsigned long x1, x2; - unsigned y; -} latin1_comb[] = { - { 'A', 0x0300, 0xc0}, /* LATIN CAPITAL LETTER A WITH GRAVE */ - { 'A', 0x0301, 0xc1}, /* LATIN CAPITAL LETTER A WITH ACUTE */ - { 'A', 0x0302, 0xc2}, /* LATIN CAPITAL LETTER A WITH CIRCUMFLEX */ - { 'A', 0x0303, 0xc3}, /* LATIN CAPITAL LETTER A WITH TILDE */ - { 'A', 0x0308, 0xc4}, /* LATIN CAPITAL LETTER A WITH DIAERESIS */ - { 'A', 0x030a, 0xc5}, /* LATIN CAPITAL LETTER A WITH RING ABOVE */ - /* no need for 0xc6 LATIN CAPITAL LETTER AE */ - { 'C', 0x0327, 0xc7}, /* LATIN CAPITAL LETTER C WITH CEDILLA */ - { 'E', 0x0300, 0xc8}, /* LATIN CAPITAL LETTER E WITH GRAVE */ - { 'E', 0x0301, 0xc9}, /* LATIN CAPITAL LETTER E WITH ACUTE */ - { 'E', 0x0302, 0xca}, /* LATIN CAPITAL LETTER E WITH CIRCUMFLEX */ - { 'E', 0x0308, 0xcb}, /* LATIN CAPITAL LETTER E WITH DIAERESIS */ - { 'I', 0x0300, 0xcc}, /* LATIN CAPITAL LETTER I WITH GRAVE */ - { 'I', 0x0301, 0xcd}, /* LATIN CAPITAL LETTER I WITH ACUTE */ - { 'I', 0x0302, 0xce}, /* LATIN CAPITAL LETTER I WITH CIRCUMFLEX */ - { 'I', 0x0308, 0xcf}, /* LATIN CAPITAL LETTER I WITH DIAERESIS */ - { 'N', 0x0303, 0xd1}, /* LATIN CAPITAL LETTER N WITH TILDE */ - { 'O', 0x0300, 0xd2}, /* LATIN CAPITAL LETTER O WITH GRAVE */ - { 'O', 0x0301, 0xd3}, /* LATIN CAPITAL LETTER O WITH ACUTE */ - { 'O', 0x0302, 0xd4}, /* LATIN CAPITAL LETTER O WITH CIRCUMFLEX */ - { 'O', 0x0303, 0xd5}, /* LATIN CAPITAL LETTER O WITH TILDE */ - { 'O', 0x0308, 0xd6}, /* LATIN CAPITAL LETTER O WITH DIAERESIS */ - /* omitted: 0xd7 MULTIPLICATION SIGN */ - /* omitted: 0xd8 LATIN CAPITAL LETTER O WITH STROKE */ - { 'U', 0x0300, 0xd9}, /* LATIN CAPITAL LETTER U WITH GRAVE */ - { 'U', 0x0301, 0xda}, /* LATIN CAPITAL LETTER U WITH ACUTE */ - { 'U', 0x0302, 0xdb}, /* LATIN CAPITAL LETTER U WITH CIRCUMFLEX */ - { 'U', 0x0308, 0xdc}, /* LATIN CAPITAL LETTER U WITH DIAERESIS */ - { 'Y', 0x0301, 0xdd}, /* LATIN CAPITAL LETTER Y WITH ACUTE */ - /* omitted: 0xde LATIN CAPITAL LETTER THORN */ - /* omitted: 0xdf LATIN SMALL LETTER SHARP S */ - { 'a', 0x0300, 0xe0}, /* LATIN SMALL LETTER A WITH GRAVE */ - { 'a', 0x0301, 0xe1}, /* LATIN SMALL LETTER A WITH ACUTE */ - { 'a', 0x0302, 0xe2}, /* LATIN SMALL LETTER A WITH CIRCUMFLEX */ - { 'a', 0x0303, 0xe3}, /* LATIN SMALL LETTER A WITH TILDE */ - { 'a', 0x0308, 0xe4}, /* LATIN SMALL LETTER A WITH DIAERESIS */ - { 'a', 0x030a, 0xe5}, /* LATIN SMALL LETTER A WITH RING ABOVE */ - /* omitted: 0xe6 LATIN SMALL LETTER AE */ - { 'c', 0x0327, 0xe7}, /* LATIN SMALL LETTER C WITH CEDILLA */ - { 'e', 0x0300, 0xe8}, /* LATIN SMALL LETTER E WITH GRAVE */ - { 'e', 0x0301, 0xe9}, /* LATIN SMALL LETTER E WITH ACUTE */ - { 'e', 0x0302, 0xea}, /* LATIN SMALL LETTER E WITH CIRCUMFLEX */ - { 'e', 0x0308, 0xeb}, /* LATIN SMALL LETTER E WITH DIAERESIS */ - { 'i', 0x0300, 0xec}, /* LATIN SMALL LETTER I WITH GRAVE */ - { 'i', 0x0301, 0xed}, /* LATIN SMALL LETTER I WITH ACUTE */ - { 'i', 0x0302, 0xee}, /* LATIN SMALL LETTER I WITH CIRCUMFLEX */ - { 'i', 0x0308, 0xef}, /* LATIN SMALL LETTER I WITH DIAERESIS */ - /* omitted: 0xf0 LATIN SMALL LETTER ETH */ - { 'n', 0x0303, 0xf1}, /* LATIN SMALL LETTER N WITH TILDE */ - { 'o', 0x0300, 0xf2}, /* LATIN SMALL LETTER O WITH GRAVE */ - { 'o', 0x0301, 0xf3}, /* LATIN SMALL LETTER O WITH ACUTE */ - { 'o', 0x0302, 0xf4}, /* LATIN SMALL LETTER O WITH CIRCUMFLEX */ - { 'o', 0x0303, 0xf5}, /* LATIN SMALL LETTER O WITH TILDE */ - { 'o', 0x0308, 0xf6}, /* LATIN SMALL LETTER O WITH DIAERESIS */ - /* omitted: 0xf7 DIVISION SIGN */ - /* omitted: 0xf8 LATIN SMALL LETTER O WITH STROKE */ - { 'u', 0x0300, 0xf9}, /* LATIN SMALL LETTER U WITH GRAVE */ - { 'u', 0x0301, 0xfa}, /* LATIN SMALL LETTER U WITH ACUTE */ - { 'u', 0x0302, 0xfb}, /* LATIN SMALL LETTER U WITH CIRCUMFLEX */ - { 'u', 0x0308, 0xfc}, /* LATIN SMALL LETTER U WITH DIAERESIS */ - { 'y', 0x0301, 0xfd}, /* LATIN SMALL LETTER Y WITH ACUTE */ - /* omitted: 0xfe LATIN SMALL LETTER THORN */ - { 'y', 0x0308, 0xff}, /* LATIN SMALL LETTER Y WITH DIAERESIS */ - - { 0, 0, 0} -}; - -#define ESC "\033" - -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) { @@ -188,8 +77,6 @@ static unsigned long yaz_read_ISO8859_1(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) @@ -275,16 +162,11 @@ static unsigned long yaz_read_marc8s(yaz_iconv_t cd, unsigned char *inp, unsigned long x = yaz_read_marc8(cd, inp, inbytesleft, no_read); if (x && cd->comb_size == 1) { - /* For MARC8s we try to get a Latin-1 page code out of it */ - int i; - for (i = 0; latin1_comb[i].x1; i++) - if (cd->comb_x[0] == latin1_comb[i].x2 && x == latin1_comb[i].x1) - { - *no_read += cd->comb_no_read[0]; - cd->comb_size = 0; - x = latin1_comb[i].y; - break; - } + if (yaz_iso_8859_1_lookup_x12(x, cd->comb_x[0], &x)) + { + *no_read += cd->comb_no_read[0]; + cd->comb_size = 0; + } } return x; } @@ -402,455 +284,45 @@ incomplete: return 0; } -static size_t yaz_write_ISO8859_1(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) -{ - /* list of two char unicode sequence that, when combined, are - equivalent to single unicode chars that can be represented in - ISO-8859-1/Latin-1. - Regular iconv on Linux at least does not seem to convert these, - but since MARC-8 to UTF-8 generates these composed sequence - we get a better chance of a successful MARC-8 -> ISO-8859-1 - conversion */ - unsigned char *outp = (unsigned char *) *outbuf; - - if (cd->compose_char) - { - int i; - for (i = 0; latin1_comb[i].x1; i++) - if (cd->compose_char == latin1_comb[i].x1 && x == latin1_comb[i].x2) - { - x = latin1_comb[i].y; - break; - } - if (*outbytesleft < 1) - { /* no room. Retain compose_char and bail out */ - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t)(-1); - } - if (!latin1_comb[i].x1) - { /* not found. Just write compose_char */ - *outp++ = (unsigned char) cd->compose_char; - (*outbytesleft)--; - *outbuf = (char *) outp; - } - /* compose_char used so reset it. x now holds current char */ - cd->compose_char = 0; - } - if (x > 32 && x < 127 && cd->compose_char == 0) - { - cd->compose_char = x; - return 0; - } - else if (x > 255 || x < 1) - { - cd->my_errno = YAZ_ICONV_EILSEQ; - return (size_t) -1; - } - else if (*outbytesleft < 1) - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t)(-1); - } - *outp++ = (unsigned char) x; - (*outbytesleft)--; - *outbuf = (char *) outp; - return 0; -} - -static size_t yaz_flush_ISO8859_1(yaz_iconv_t cd, - char **outbuf, size_t *outbytesleft) -{ - if (cd->compose_char) - { - unsigned char *outp = (unsigned char *) *outbuf; - if (*outbytesleft < 1) - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t)(-1); - } - *outp++ = (unsigned char) cd->compose_char; - (*outbytesleft)--; - *outbuf = (char *) outp; - cd->compose_char = 0; - } - return 0; -} - -static unsigned long lookup_marc8(yaz_iconv_t cd, - unsigned long x, int *comb, - const char **page_chr) -{ - char utf8_buf[7]; - char *utf8_outbuf = utf8_buf; - size_t utf8_outbytesleft = sizeof(utf8_buf)-1, r; - - r = yaz_write_UTF8(cd, x, &utf8_outbuf, &utf8_outbytesleft); - if (r == (size_t)(-1)) - { - cd->my_errno = YAZ_ICONV_EILSEQ; - return 0; - } - else - { - unsigned char *inp; - size_t inbytesleft, no_read_sub = 0; - unsigned long x; - - *utf8_outbuf = '\0'; - inp = (unsigned char *) utf8_buf; - inbytesleft = strlen(utf8_buf); - - x = yaz_marc8r_42_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(B"; - return x; - } - x = yaz_marc8r_45_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(B"; - return x; - } - x = yaz_marc8r_62_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "b"; - return x; - } - x = yaz_marc8r_70_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "p"; - return x; - } - x = yaz_marc8r_32_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(2"; - return x; - } - x = yaz_marc8r_4E_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(N"; - return x; - } - x = yaz_marc8r_51_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(Q"; - return x; - } - x = yaz_marc8r_33_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(3"; - return x; - } - x = yaz_marc8r_34_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(4"; - return x; - } - x = yaz_marc8r_53_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "(S"; - return x; - } - x = yaz_marc8r_31_conv(inp, inbytesleft, &no_read_sub, comb, 255, 0); - if (x) - { - *page_chr = ESC "$1"; - return x; - } - cd->my_errno = YAZ_ICONV_EILSEQ; - return x; - } -} - -static size_t flush_combos(yaz_iconv_t cd, - char **outbuf, size_t *outbytesleft) -{ - unsigned long y = cd->write_marc8_last; - - 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; - } - - if (9 >= *outbytesleft) - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t) (-1); - } - if (cd->write_marc8_ncr) - { - yaz_snprintf(*outbuf, 9, "&#x%04x;", y); - (*outbytesleft) -= 8; - (*outbuf) += 8; - } - else - { - size_t out_no = 0; - unsigned char byte; - - byte = (unsigned char )((y>>16) & 0xff); - if (byte) - (*outbuf)[out_no++] = byte; - byte = (unsigned char)((y>>8) & 0xff); - if (byte) - (*outbuf)[out_no++] = byte; - byte = (unsigned char )(y & 0xff); - if (byte) - (*outbuf)[out_no++] = byte; - *outbuf += out_no; - (*outbytesleft) -= out_no; - } - - if (cd->write_marc8_second_half_char) - { - *(*outbuf)++ = cd->write_marc8_second_half_char; - (*outbytesleft)--; - } - - cd->write_marc8_last = 0; - cd->write_marc8_ncr = 0; - cd->write_marc8_lpage = 0; - cd->write_marc8_second_half_char = 0; - return 0; -} - -static size_t yaz_write_marc8_page_chr(yaz_iconv_t cd, - char **outbuf, size_t *outbytesleft, - const char *page_chr) -{ - const char **old_page_chr = &cd->write_marc8_g0; - /* are we going to a G1-set (such as such as ESC ")!E") */ - if (page_chr && page_chr[1] == ')') - old_page_chr = &cd->write_marc8_g1; - - if (!*old_page_chr || strcmp(page_chr, *old_page_chr)) - { - size_t plen = 0; - const char *page_out = page_chr; - - if (*outbytesleft < 8) - { - cd->my_errno = YAZ_ICONV_E2BIG; - - return (size_t) (-1); - } - - if (*old_page_chr) - { - if (!strcmp(*old_page_chr, ESC "p") - || !strcmp(*old_page_chr, ESC "g") - || !strcmp(*old_page_chr, ESC "b")) - { - page_out = ESC "s"; - /* Technique 1 leave */ - if (strcmp(page_chr, ESC "(B")) /* Not going ASCII page? */ - { - /* Must leave script + enter new page */ - plen = strlen(page_out); - memcpy(*outbuf, page_out, plen); - (*outbuf) += plen; - (*outbytesleft) -= plen; - page_out = ESC "(B"; - } - } - } - *old_page_chr = page_chr; - plen = strlen(page_out); - memcpy(*outbuf, page_out, plen); - (*outbuf) += plen; - (*outbytesleft) -= plen; - } - return 0; -} - - -static size_t yaz_write_marc8_2(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft, - int loss_mode) -{ - int comb = 0; - int enable_ncr = 0; - const char *page_chr = 0; - unsigned long y = lookup_marc8(cd, x, &comb, &page_chr); - - if (!y) - { - if (loss_mode == 0 || cd->my_errno != YAZ_ICONV_EILSEQ) - return (size_t) (-1); - page_chr = ESC "(B"; - if (loss_mode == 1) - y = '|'; - else - { - y = x; - enable_ncr = 1; - } - } - - if (comb) - { - if (page_chr) - { - size_t r = yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, - page_chr); - if (r) - return r; - } - if (x == 0x0361) - cd->write_marc8_second_half_char = 0xEC; - else if (x == 0x0360) - cd->write_marc8_second_half_char = 0xFB; - - if (*outbytesleft <= 1) - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t) (-1); - } - *(*outbuf)++ = y; - (*outbytesleft)--; - } - else - { - size_t r = flush_combos(cd, outbuf, outbytesleft); - if (r) - return r; - - cd->write_marc8_last = y; - cd->write_marc8_lpage = page_chr; - cd->write_marc8_ncr = enable_ncr; - } - return 0; -} - -static size_t yaz_flush_marc8(yaz_iconv_t cd, - char **outbuf, size_t *outbytesleft) -{ - size_t r = flush_combos(cd, outbuf, outbytesleft); - if (r) - return r; - cd->write_marc8_g1 = 0; - return yaz_write_marc8_page_chr(cd, outbuf, outbytesleft, ESC "(B"); -} - -static size_t yaz_write_marc8_generic(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft, - int loss_mode); - -static size_t yaz_write_marc8_normal(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) -{ - return yaz_write_marc8_generic(cd, x, outbuf, outbytesleft, 0); -} - -static size_t yaz_write_marc8_lossy(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) -{ - return yaz_write_marc8_generic(cd, x, outbuf, outbytesleft, 1); -} - -static size_t yaz_write_marc8_lossless(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) -{ - return yaz_write_marc8_generic(cd, x, outbuf, outbytesleft, 2); -} - -static size_t yaz_write_marc8_generic(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft, - int loss_mode) +int yaz_iconv_isbuiltin(yaz_iconv_t cd) { - if (x >= 0xc0 && x <= 0xff) /* optimization. min and max .y values */ - { - int i; - for (i = 0; latin1_comb[i].x1; i++) - { - if (x == latin1_comb[i].y) - { - size_t r ; - /* save the output pointers .. */ - char *outbuf0 = *outbuf; - size_t outbytesleft0 = *outbytesleft; - int last_ch = cd->write_marc8_last; - int ncr = cd->write_marc8_ncr; - const char *lpage = cd->write_marc8_lpage; - - r = yaz_write_marc8_2(cd, latin1_comb[i].x1, - outbuf, outbytesleft, loss_mode); - if (r) - return r; - r = yaz_write_marc8_2(cd, latin1_comb[i].x2, - outbuf, outbytesleft, loss_mode); - if (r && cd->my_errno == YAZ_ICONV_E2BIG) - { - /* not enough room. reset output to original values */ - *outbuf = outbuf0; - *outbytesleft = outbytesleft0; - cd->write_marc8_last = last_ch; - cd->write_marc8_ncr = ncr; - cd->write_marc8_lpage = lpage; - } - return r; - } - } - } - return yaz_write_marc8_2(cd, x, outbuf, outbytesleft, loss_mode); + return cd->read_handle && cd->encoder.write_handle; } -#if HAVE_WCHAR_H -static size_t yaz_write_wchar_t(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static int prepare_encoders(yaz_iconv_t cd, const char *tocode) { - 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; + if (yaz_marc8_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_utf8_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_ucs4_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_iso_8859_1_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_iso_5428_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_advancegreek_encoder(tocode, &cd->encoder)) + return 1; + if (yaz_wchar_encoder(tocode, &cd->encoder)) + return 1; 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)); - cd->write_handle = 0; + cd->encoder.data = 0; + cd->encoder.write_handle = 0; + cd->encoder.flush_handle = 0; + cd->encoder.init_handle = 0; + cd->encoder.destroy_handle = 0; + cd->read_handle = 0; cd->init_handle = 0; - cd->flush_handle = 0; cd->my_errno = YAZ_ICONV_UNKNOWN; /* a useful hack: if fromcode has leading @, @@ -884,73 +356,29 @@ yaz_iconv_t yaz_iconv_open(const char *tocode, const char *fromcode) 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; - cd->flush_handle = yaz_flush_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; - else if (!yaz_matchstr(tocode, "MARC8")) - { - cd->write_handle = yaz_write_marc8_normal; - cd->flush_handle = yaz_flush_marc8; - } - else if (!yaz_matchstr(tocode, "MARC8s")) - { - cd->write_handle = yaz_write_marc8_normal; - cd->flush_handle = yaz_flush_marc8; - } - else if (!yaz_matchstr(tocode, "MARC8lossy")) - { - cd->write_handle = yaz_write_marc8_lossy; - cd->flush_handle = yaz_flush_marc8; - } - else if (!yaz_matchstr(tocode, "MARC8lossless")) - { - cd->write_handle = yaz_write_marc8_lossless; - cd->flush_handle = yaz_flush_marc8; - } - else if (!yaz_matchstr(tocode, "advancegreek")) - { - cd->write_handle = yaz_write_advancegreek; - } - else if (!yaz_matchstr(tocode, "iso54281984")) - { - cd->write_handle = yaz_write_iso5428_1984; - } - else if (!yaz_matchstr(tocode, "iso5428:1984")) - { - cd->write_handle = yaz_write_iso5428_1984; - } -#if HAVE_WCHAR_H - else if (!yaz_matchstr(tocode, "WCHAR_T")) - cd->write_handle = yaz_write_wchar_t; -#endif + prepare_encoders(cd, tocode); } + if (cd->read_handle && cd->encoder.write_handle) + { #if HAVE_ICONV_H - cd->iconv_cd = 0; - if (!cd->read_handle || !cd->write_handle) + cd->iconv_cd = 0; +#endif + ; + } + else { +#if HAVE_ICONV_H cd->iconv_cd = iconv_open(tocode, fromcode); if (cd->iconv_cd == (iconv_t) (-1)) { - xfree(cd); + yaz_iconv_close(cd); return 0; } - } #else - if (!cd->read_handle || !cd->write_handle) - { - xfree(cd); + yaz_iconv_close(cd); return 0; - } #endif + } cd->init_flag = 1; return cd; } @@ -997,21 +425,13 @@ size_t yaz_iconv(yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, cd->g1_mode = 'E'; cd->comb_offset = cd->comb_size = 0; - cd->compose_char = 0; - - cd->write_marc8_second_half_char = 0; - cd->write_marc8_last = 0; - cd->write_marc8_ncr = 0; - cd->write_marc8_lpage = 0; - cd->write_marc8_g0 = ESC "(B"; - cd->write_marc8_g1 = 0; + + if (cd->encoder.init_handle) + (*cd->encoder.init_handle)(&cd->encoder); cd->unget_x = 0; cd->no_read_x = 0; - } - if (cd->init_flag) - { if (cd->init_handle && inbuf && *inbuf) { size_t no_read = 0; @@ -1035,9 +455,11 @@ size_t yaz_iconv(yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, if (outbuf && *outbuf) { if (cd->unget_x) - r = (*cd->write_handle)(cd, cd->unget_x, outbuf, outbytesleft); - if (cd->flush_handle) - r = (*cd->flush_handle)(cd, outbuf, outbytesleft); + r = (*cd->encoder.write_handle)(cd, &cd->encoder, + cd->unget_x, outbuf, outbytesleft); + if (cd->encoder.flush_handle) + r = (*cd->encoder.flush_handle)(cd, &cd->encoder, + outbuf, outbytesleft); } if (r == 0) cd->init_flag = 1; @@ -1071,7 +493,8 @@ size_t yaz_iconv(yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, } if (x) { - r = (*cd->write_handle)(cd, x, outbuf, outbytesleft); + r = (*cd->encoder.write_handle)(cd, &cd->encoder, + x, outbuf, outbytesleft); if (r) { /* unable to write it. save it because read_handle cannot @@ -1102,6 +525,8 @@ int yaz_iconv_close(yaz_iconv_t cd) if (cd->iconv_cd) iconv_close(cd->iconv_cd); #endif + if (cd->encoder.destroy_handle) + (*cd->encoder.destroy_handle)(&cd->encoder); xfree(cd); return 0; } diff --git a/src/ucs4.c b/src/ucs4.c index dedf4ec..43e9f84 100644 --- a/src/ucs4.c +++ b/src/ucs4.c @@ -2,11 +2,10 @@ * Copyright (C) 1995-2008, Index Data ApS * See the file LICENSE for details. * - * $Id: siconv.c,v 1.50 2008-03-12 08:53:28 adam Exp $ */ /** * \file - * \brief ISO-5428 character mapping (iconv) + * \brief UCS4 decoding and encoding */ #if HAVE_CONFIG_H @@ -56,8 +55,9 @@ unsigned long yaz_read_UCS4LE(yaz_iconv_t cd, unsigned char *inp, return x; } -size_t yaz_write_UCS4(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static size_t write_UCS4(yaz_iconv_t cd, yaz_iconv_encoder_t en, + unsigned long x, + char **outbuf, size_t *outbytesleft) { unsigned char *outp = (unsigned char *) *outbuf; if (*outbytesleft >= 4) @@ -77,8 +77,9 @@ size_t yaz_write_UCS4(yaz_iconv_t cd, unsigned long x, return 0; } -size_t yaz_write_UCS4LE(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static size_t write_UCS4LE(yaz_iconv_t cd, yaz_iconv_encoder_t en, + unsigned long x, + char **outbuf, size_t *outbytesleft) { unsigned char *outp = (unsigned char *) *outbuf; if (*outbytesleft >= 4) @@ -99,6 +100,20 @@ size_t yaz_write_UCS4LE(yaz_iconv_t cd, unsigned long x, } +yaz_iconv_encoder_t yaz_ucs4_encoder(const char *tocode, + yaz_iconv_encoder_t e) + +{ + if (!yaz_matchstr(tocode, "UCS4")) + e->write_handle = write_UCS4; + else if (!yaz_matchstr(tocode, "UCS4LE")) + e->write_handle = write_UCS4LE; + else + return 0; + return e; +} + + /* * Local variables: diff --git a/src/utf8.c b/src/utf8.c index b893e0d..37ee04c 100644 --- a/src/utf8.c +++ b/src/utf8.c @@ -2,11 +2,10 @@ * Copyright (C) 1995-2008, Index Data ApS * See the file LICENSE for details. * - * $Id: siconv.c,v 1.50 2008-03-12 08:53:28 adam Exp $ */ /** * \file - * \brief ISO-5428 character mapping (iconv) + * \brief UTF-8 encoding / decoding */ #if HAVE_CONFIG_H @@ -147,8 +146,9 @@ unsigned long yaz_read_UTF8(yaz_iconv_t cd, unsigned char *inp, } -size_t yaz_write_UTF8(yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static size_t write_UTF8(yaz_iconv_t cd, yaz_iconv_encoder_t en, + unsigned long x, + char **outbuf, size_t *outbytesleft) { int err = 0; int r = yaz_write_UTF8_char(x, outbuf, outbytesleft, &err); @@ -216,6 +216,18 @@ size_t yaz_write_UTF8_char(unsigned long x, return 0; } +yaz_iconv_encoder_t yaz_utf8_encoder(const char *tocode, + yaz_iconv_encoder_t e) + +{ + if (!yaz_matchstr(tocode, "UTF8")) + { + e->write_handle = write_UTF8; + return e; + } + return 0; +} + /* * Local variables: diff --git a/win/makefile b/win/makefile index 5671056..b8f9d8a 100644 --- a/win/makefile +++ b/win/makefile @@ -484,7 +484,10 @@ MISC_OBJS= \ $(OBJDIR)\http.obj \ $(OBJDIR)\match_glob.obj \ $(OBJDIR)\poll.obj \ - $(OBJDIR)\daemon.obj + $(OBJDIR)\daemon.obj \ + $(OBJDIR)\iconv_encode_iso_8859_1.obj \ + $(OBJDIR)\iconv_encode_marc8.obj \ + $(OBJDIR)\iconv_encode_wchar.obj Z3950_OBJS= \ $(OBJDIR)\z-date.obj\ -- 1.7.10.4