X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=src%2Fsiconv.c;h=42220e2cb67b2324eee368bc3e217c448e5a7b59;hp=36f2c2d3cdee09f5faa316f0a4dbb0b8e4675c87;hb=a6b61b9f8b6576e8159a4e7e6aee88bba7e5d07d;hpb=c6e47cbbff56f39f6d81b079ebaeac41d793d4d9 diff --git a/src/siconv.c b/src/siconv.c index 36f2c2d..42220e2 100644 --- a/src/siconv.c +++ b/src/siconv.c @@ -1,383 +1,112 @@ -/* - * Copyright (c) 1997-2003, Index Data +/* This file is part of the YAZ toolkit. + * Copyright (C) Index Data * See the file LICENSE for details. + */ +/** + * \file + * \brief Implements simple ICONV + * + * This implements an interface similar to that of iconv and + * is used by YAZ to interface with iconv (if present). + * For systems where iconv is not present, this layer + * provides a few important conversions: UTF-8, MARC-8, Latin-1. * - * $Id: siconv.c,v 1.1 2003-10-27 12:21:35 adam Exp $ */ -/* mini iconv and wrapper for system iconv library (if present) */ - #if HAVE_CONFIG_H #include #endif +#include #include #include -#include -#if HAVE_WCHAR_H -#include -#endif #if HAVE_ICONV_H #include #endif -#include +#include +#include +#include "iconv-p.h" -unsigned long yaz_marc8_conv (unsigned char *inp, size_t inbytesleft, - size_t *no_read); - 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); - 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 no_read_x; + unsigned long unget_x; #if HAVE_ICONV_H iconv_t iconv_cd; #endif + struct yaz_iconv_encoder_s encoder; + struct yaz_iconv_decoder_s decoder; }; -static unsigned long yaz_read_ISO8859_1 (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - unsigned long x = inp[0]; - *no_read = 1; - return x; -} -static size_t yaz_init_UTF8 (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - if (inp[0] != 0xef) - { - *no_read = 0; - return 0; - } - if (inbytesleft < 3) - { - cd->my_errno = YAZ_ICONV_EINVAL; - return (size_t) -1; - } - if (inp[1] != 0xbb || inp[2] != 0xbf) - { - cd->my_errno = YAZ_ICONV_EILSEQ; - return (size_t) -1; - } - *no_read = 3; - return 0; -} - -static unsigned long yaz_read_UTF8 (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - unsigned long x = 0; - - if (inp[0] <= 0x7f) - { - x = inp[0]; - *no_read = 1; - } - else if (inp[0] <= 0xbf || inp[0] >= 0xfe) - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - else if (inp[0] <= 0xdf && inbytesleft >= 2) - { - x = ((inp[0] & 0x1f) << 6) | (inp[1] & 0x3f); - if (x >= 0x80) - *no_read = 2; - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - } - else if (inp[0] <= 0xef && inbytesleft >= 3) - { - x = ((inp[0] & 0x0f) << 12) | ((inp[1] & 0x3f) << 6) | - (inp[1] & 0x3f); - if (x >= 0x800) - *no_read = 3; - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - } - else if (inp[0] <= 0xf7 && inbytesleft >= 4) - { - x = ((inp[0] & 0x07) << 18) | ((inp[1] & 0x3f) << 12) | - ((inp[2] & 0x3f) << 6) | (inp[3] & 0x3f); - if (x >= 0x10000) - *no_read = 4; - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - } - else if (inp[0] <= 0xfb && inbytesleft >= 5) - { - x = ((inp[0] & 0x03) << 24) | ((inp[1] & 0x3f) << 18) | - ((inp[2] & 0x3f) << 12) | ((inp[3] & 0x3f) << 6) | - (inp[4] & 0x3f); - if (x >= 0x200000) - *no_read = 5; - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - } - else if (inp[0] <= 0xfd && inbytesleft >= 6) - { - x = ((inp[0] & 0x01) << 30) | ((inp[1] & 0x3f) << 24) | - ((inp[2] & 0x3f) << 18) | ((inp[3] & 0x3f) << 12) | - ((inp[4] & 0x3f) << 6) | (inp[5] & 0x3f); - if (x >= 0x4000000) - *no_read = 6; - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EILSEQ; - } - } - else - { - *no_read = 0; - cd->my_errno = YAZ_ICONV_EINVAL; - } - return x; -} - -static unsigned long yaz_read_UCS4 (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - unsigned long x = 0; - - if (inbytesleft < 4) - { - cd->my_errno = YAZ_ICONV_EINVAL; /* incomplete input */ - *no_read = 0; - } - else - { - x = (inp[0]<<24) | (inp[1]<<16) | (inp[2]<<8) | inp[3]; - *no_read = 4; - } - return x; -} - -static unsigned long yaz_read_UCS4LE (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - unsigned long x = 0; - - if (inbytesleft < 4) - { - cd->my_errno = YAZ_ICONV_EINVAL; /* incomplete input */ - *no_read = 0; - } - else - { - x = (inp[3]<<24) | (inp[2]<<16) | (inp[1]<<8) | inp[0]; - *no_read = 4; - } - 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) -{ - unsigned long x = 0; - - if (inbytesleft < sizeof(wchar_t)) - { - cd->my_errno = YAZ_ICONV_EINVAL; /* incomplete input */ - *no_read = 0; - } - else - { - wchar_t wch; - memcpy (&wch, inp, sizeof(wch)); - x = wch; - *no_read = sizeof(wch); - } - return x; -} -#endif - -static unsigned long yaz_read_marc8 (yaz_iconv_t cd, unsigned char *inp, - size_t inbytesleft, size_t *no_read) -{ - return yaz_marc8_conv(inp, inbytesleft, no_read); -} - -static size_t yaz_write_UTF8 (yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) -{ - unsigned char *outp = (unsigned char *) *outbuf; - if (x <= 0x7f && *outbytesleft >= 1) - { - *outp++ = (unsigned char) x; - (*outbytesleft)--; - } - else if (x <= 0x7ff && *outbytesleft >= 2) - { - *outp++ = (unsigned char) ((x >> 6) | 0xc0); - *outp++ = (unsigned char) ((x & 0x3f) | 0x80); - (*outbytesleft) -= 2; - } - else if (x <= 0xffff && *outbytesleft >= 3) - { - *outp++ = (unsigned char) ((x >> 12) | 0xe0); - *outp++ = (unsigned char) (((x >> 6) & 0x3f) | 0x80); - *outp++ = (unsigned char) ((x & 0x3f) | 0x80); - (*outbytesleft) -= 3; - } - else if (x <= 0x1fffff && *outbytesleft >= 4) - { - *outp++ = (unsigned char) ((x >> 18) | 0xf0); - *outp++ = (unsigned char) (((x >> 12) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 6) & 0x3f) | 0x80); - *outp++ = (unsigned char) ((x & 0x3f) | 0x80); - (*outbytesleft) -= 4; - } - else if (x <= 0x3ffffff && *outbytesleft >= 5) - { - *outp++ = (unsigned char) ((x >> 24) | 0xf8); - *outp++ = (unsigned char) (((x >> 18) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 12) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 6) & 0x3f) | 0x80); - *outp++ = (unsigned char) ((x & 0x3f) | 0x80); - (*outbytesleft) -= 5; - } - else if (*outbytesleft >= 6) - { - *outp++ = (unsigned char) ((x >> 30) | 0xfc); - *outp++ = (unsigned char) (((x >> 24) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 18) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 12) & 0x3f) | 0x80); - *outp++ = (unsigned char) (((x >> 6) & 0x3f) | 0x80); - *outp++ = (unsigned char) ((x & 0x3f) | 0x80); - (*outbytesleft) -= 6; - } - else - { - cd->my_errno = YAZ_ICONV_E2BIG; /* not room for output */ - return (size_t)(-1); - } - *outbuf = (char *) outp; - return 0; -} - -static size_t yaz_write_ISO8859_1 (yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +int yaz_iconv_isbuiltin(yaz_iconv_t cd) { - unsigned char *outp = (unsigned char *) *outbuf; - if (x > 255 || x < 1) - { - cd->my_errno = YAZ_ICONV_EILSEQ; - return (size_t) -1; - } - else if (*outbytesleft >= 1) - { - *outp++ = (unsigned char) x; - (*outbytesleft)--; - } - else - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t)(-1); - } - *outbuf = (char *) outp; - return 0; + return cd->decoder.read_handle && cd->encoder.write_handle; } -static size_t yaz_write_UCS4 (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 >= 4) - { - *outp++ = (unsigned char) (x<<24); - *outp++ = (unsigned char) (x<<16); - *outp++ = (unsigned char) (x<<8); - *outp++ = (unsigned char) x; - (*outbytesleft) -= 4; - } - 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; + if (yaz_danmarc_encoder(tocode, &cd->encoder)) + return 1; return 0; } -static size_t yaz_write_UCS4LE (yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +static int prepare_decoders(yaz_iconv_t cd, const char *tocode) { - unsigned char *outp = (unsigned char *) *outbuf; - if (*outbytesleft >= 4) - { - *outp++ = (unsigned char) x; - *outp++ = (unsigned char) (x<<8); - *outp++ = (unsigned char) (x<<16); - *outp++ = (unsigned char) (x<<24); - (*outbytesleft) -= 4; - } - else - { - cd->my_errno = YAZ_ICONV_E2BIG; - return (size_t)(-1); - } - *outbuf = (char *) outp; + if (yaz_marc8_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_iso5426_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_utf8_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_ucs4_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_iso_8859_1_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_iso_5428_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_advancegreek_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_wchar_decoder(tocode, &cd->decoder)) + return 1; + if (yaz_danmarc_decoder(tocode, &cd->decoder)) + return 1; return 0; } -#if HAVE_WCHAR_H -static size_t yaz_write_wchar_t (yaz_iconv_t cd, unsigned long x, - char **outbuf, size_t *outbytesleft) +yaz_iconv_t yaz_iconv_open(const char *tocode, const char *fromcode) { - 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; - return 0; -} -#endif + yaz_iconv_t cd = (yaz_iconv_t) xmalloc(sizeof(*cd)); -int yaz_iconv_isbuiltin(yaz_iconv_t cd) -{ - return cd->read_handle && cd->write_handle; -} + cd->encoder.data = 0; + cd->encoder.write_handle = 0; + cd->encoder.flush_handle = 0; + cd->encoder.init_handle = 0; + cd->encoder.destroy_handle = 0; -yaz_iconv_t yaz_iconv_open (const char *tocode, const char *fromcode) -{ - yaz_iconv_t cd = (yaz_iconv_t) xmalloc (sizeof(*cd)); + cd->decoder.data = 0; + cd->decoder.read_handle = 0; + cd->decoder.init_handle = 0; + cd->decoder.destroy_handle = 0; - cd->write_handle = 0; - cd->read_handle = 0; - cd->init_handle = 0; cd->my_errno = YAZ_ICONV_UNKNOWN; /* a useful hack: if fromcode has leading @, @@ -386,66 +115,42 @@ yaz_iconv_t yaz_iconv_open (const char *tocode, const char *fromcode) fromcode++; else { - if (!yaz_matchstr(fromcode, "UTF8")) - { - cd->read_handle = yaz_read_UTF8; - cd->init_handle = yaz_init_UTF8; - } - else if (!yaz_matchstr(fromcode, "ISO88591")) - cd->read_handle = yaz_read_ISO8859_1; - else if (!yaz_matchstr(fromcode, "UCS4")) - cd->read_handle = yaz_read_UCS4; - else if (!yaz_matchstr(fromcode, "UCS4LE")) - cd->read_handle = yaz_read_UCS4LE; - else if (!yaz_matchstr(fromcode, "MARC8")) - cd->read_handle = yaz_read_marc8; -#if HAVE_WCHAR_H - 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; - else if (!yaz_matchstr (tocode, "UCS4")) - cd->write_handle = yaz_write_UCS4; - else if (!yaz_matchstr(tocode, "UCS4LE")) - cd->write_handle = yaz_write_UCS4LE; -#if HAVE_WCHAR_H - else if (!yaz_matchstr(tocode, "WCHAR_T")) - cd->write_handle = yaz_write_wchar_t; -#endif + prepare_encoders(cd, tocode); + prepare_decoders(cd, fromcode); } + if (cd->decoder.read_handle && cd->encoder.write_handle) + { #if HAVE_ICONV_H - cd->iconv_cd = 0; - if (!cd->read_handle || !cd->write_handle) + cd->iconv_cd = (iconv_t) (-1); +#endif + ; + } + else { - cd->iconv_cd = iconv_open (tocode, fromcode); +#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; } -size_t yaz_iconv (yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, - char **outbuf, size_t *outbytesleft) +size_t yaz_iconv(yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, + char **outbuf, size_t *outbytesleft) { - char *inbuf0; + char *inbuf0 = 0; size_t r = 0; + #if HAVE_ICONV_H - if (cd->iconv_cd) + if (cd->iconv_cd != (iconv_t) (-1)) { size_t r = iconv(cd->iconv_cd, inbuf, inbytesleft, outbuf, outbytesleft); @@ -469,21 +174,28 @@ size_t yaz_iconv (yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, return r; } #endif - if (inbuf == 0 || *inbuf == 0) - { - cd->init_flag = 1; - cd->my_errno = YAZ_ICONV_UNKNOWN; - return 0; - } - inbuf0 = *inbuf; + + if (inbuf) + inbuf0 = *inbuf; if (cd->init_flag) { - if (cd->init_handle) + cd->my_errno = YAZ_ICONV_UNKNOWN; + + if (cd->encoder.init_handle) + (*cd->encoder.init_handle)(&cd->encoder); + + cd->unget_x = 0; + cd->no_read_x = 0; + + if (cd->decoder.init_handle) { - size_t no_read; - size_t r = (cd->init_handle)(cd, (unsigned char *) *inbuf, - *inbytesleft, &no_read); + size_t no_read = 0; + size_t r = (cd->decoder.init_handle)( + cd, &cd->decoder, + inbuf ? (unsigned char *) *inbuf : 0, + inbytesleft ? *inbytesleft : 0, + &no_read); if (r) { if (cd->my_errno == YAZ_ICONV_EINVAL) @@ -491,51 +203,109 @@ size_t yaz_iconv (yaz_iconv_t cd, char **inbuf, size_t *inbytesleft, cd->init_flag = 0; return r; } - *inbytesleft -= no_read; - *inbuf += no_read; + if (inbytesleft) + *inbytesleft -= no_read; + if (inbuf) + *inbuf += no_read; + } + } + cd->init_flag = 0; + + if (!inbuf || !*inbuf) + { + if (outbuf && *outbuf) + { + if (cd->unget_x) + 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); } - cd->init_flag = 0; + if (r == 0) + cd->init_flag = 1; + cd->unget_x = 0; + return r; } while (1) { unsigned long x; size_t no_read; - if (*inbytesleft == 0) + if (cd->unget_x) { - r = *inbuf - inbuf0; - break; + x = cd->unget_x; + no_read = cd->no_read_x; + } + else + { + if (*inbytesleft == 0) + { + r = *inbuf - inbuf0; + break; + } + x = (*cd->decoder.read_handle)( + cd, &cd->decoder, + (unsigned char *) *inbuf, *inbytesleft, &no_read); + if (no_read == 0) + { + r = (size_t)(-1); + break; + } } - - x = (cd->read_handle)(cd, (unsigned char *) *inbuf, *inbytesleft, - &no_read); - if (no_read == 0) + if (x) { - r = (size_t)(-1); - break; + r = (*cd->encoder.write_handle)(cd, &cd->encoder, + x, outbuf, outbytesleft); + if (r) + { + /* unable to write it. save it because read_handle cannot + rewind .. */ + if (cd->my_errno == YAZ_ICONV_E2BIG) + { + cd->unget_x = x; + cd->no_read_x = no_read; + break; + } + } + cd->unget_x = 0; } - r = (cd->write_handle)(cd, x, outbuf, outbytesleft); - if (r) - break; *inbytesleft -= no_read; (*inbuf) += no_read; } return r; } -int yaz_iconv_error (yaz_iconv_t cd) +int yaz_iconv_error(yaz_iconv_t cd) { return cd->my_errno; } -int yaz_iconv_close (yaz_iconv_t cd) +int yaz_iconv_close(yaz_iconv_t cd) { #if HAVE_ICONV_H - if (cd->iconv_cd) - iconv_close (cd->iconv_cd); + if (cd->iconv_cd != (iconv_t) (-1)) + iconv_close(cd->iconv_cd); #endif - xfree (cd); + if (cd->encoder.destroy_handle) + (*cd->encoder.destroy_handle)(&cd->encoder); + if (cd->decoder.destroy_handle) + (*cd->decoder.destroy_handle)(&cd->decoder); + xfree(cd); return 0; } - +void yaz_iconv_set_errno(yaz_iconv_t cd, int no) +{ + cd->my_errno = no; +} + +/* + * Local variables: + * c-basic-offset: 4 + * c-file-style: "Stroustrup" + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ +