From e0c6872cf40896c7be36b11dcc744620f10adf1d Mon Sep 17 00:00:00 2001 From: Norbert Preining Date: Mon, 2 Sep 2019 13:46:59 +0900 Subject: Initial commit --- support/dktools/dk4rec12.ctr | 194 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 194 insertions(+) create mode 100644 support/dktools/dk4rec12.ctr (limited to 'support/dktools/dk4rec12.ctr') diff --git a/support/dktools/dk4rec12.ctr b/support/dktools/dk4rec12.ctr new file mode 100644 index 0000000000..2259b75a7e --- /dev/null +++ b/support/dktools/dk4rec12.ctr @@ -0,0 +1,194 @@ +%% options + +copyright owner = Dirk Krause +copyright year = 2015-xxxx +license = bsd + + +%% header + +/** @file + String recoding from UTF-16 to UTF-8. + + CRT on Windows: Not used. +*/ + +#ifndef DK4CONF_H_INCLUDED +#include "dk4conf.h" +#endif + +#ifndef DK4TYPES_H_INCLUDED +#include "dk4types.h" +#endif + +#ifndef DK4ERROR_H_INCLUDED +#include "dk4error.h" +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +/** Recode string from UTF-16 to UTF-8. + @param dstb Destination buffer. + @param szdstb Size of destination buffer (number of bytes). + @param src Source string to convert. + @param erp Error report, may be NULL. + @return 1 on success, 0 on error. + + Error codes: + - DK4_E_INVALID_ARGUMENTS
+ if src or dstb is NULL or szdstb is 0, + - DK4_E_BUFFER_TOO_SMALL
+ if dstb is too small, + - DK4_E_SYNTAX
+ with the number of successfully recoded characters in nelem if a + non-recodable character was found. +*/ +int +dk4recode_utf16_to_utf8( + char *dstb, size_t szdstb, const dk4_c16_t *src, dk4_er_t *erp +); + +/** Calculate required buffer size (including final 0x00 byte) + for UTF-16 to UTF-8 conversion. + @param src UTF-16 encoded source string. + @param erp Error report, may be NULL. + @return Positive buffer size on success, 0 on error. + + Error codes: + - DK4_E_INVALID_ARGUMENTS
+ if src or dstb is NULL or szdstb is 0, + - DK4_E_SYNTAX
+ with the number of successfully recoded characters in nelem if a + non-recodable character was found. +*/ +size_t +dk4recode_size_utf16_to_utf8(const dk4_c16_t *src, dk4_er_t *erp); + +#ifdef __cplusplus +} +#endif + + +%% module + +#include "dk4rec12.h" +#include "dk4mem.h" +#include "dk4utf8.h" +#include "dk4utf16.h" + + + +size_t +dk4recode_size_utf16_to_utf8(const dk4_c16_t *src, dk4_er_t *erp) +{ + dk4_er_t er; + unsigned char buf[8]; + dk4_utf16_decoder_t dec; + size_t sz; + size_t rd = 0; + size_t back = 0; + dk4_c32_t c32 = dkC32(0); + if (NULL != src) { + dk4error_init(&er); + dk4utf16_init(&dec); + while(((dk4_c16_t)0U != *src) && (DK4_E_NONE == er.ec)) { + switch(dk4utf16_add(&dec, *(src++))) { + case DK4_EDSTM_FINISHED: { + c32 = dk4utf16_get(&dec); + dk4utf16_init(&dec); + sz = sizeof(buf); + if (0 != dk4utf8_encode(buf, &sz, c32, &er)) { + if ((SIZE_MAX - sz) >= back) { + back += sz; + rd++; + } else { + dk4error_set_simple_error_code(&er, DK4_E_MATH_OVERFLOW); + } + } + } break; + case DK4_EDSTM_ERROR: { + dk4error_set_elsize_nelem(&er, DK4_E_SYNTAX, 2, rd); + } break; + } + } + if (0 == dk4utf16_is_empty(&dec)) { + dk4error_set_elsize_nelem(&er, DK4_E_SYNTAX, 2, rd); + } + if (SIZE_MAX > back) { + back++; + } else { + dk4error_set_simple_error_code(&er, DK4_E_MATH_OVERFLOW); + } + if (DK4_E_NONE != er.ec) { + dk4error_copy(erp, &er); + back = 0; + } + } else { + dk4error_set_simple_error_code(erp, DK4_E_INVALID_ARGUMENTS); + } + return back; +} + + + +int +dk4recode_utf16_to_utf8( + char *dstb, size_t szdstb, const dk4_c16_t *src, dk4_er_t *erp +) +{ + unsigned char buf[8]; + dk4_utf16_decoder_t dec; + size_t sz; + size_t used = 0; + size_t rd = 0; + dk4_c32_t c32 = dkC32(0); + int back = 0; + + if ((NULL != dstb) && (NULL != src) && (0 < szdstb)) { + back = 1; + dk4utf16_init(&dec); + while(((dk4_c16_t)0U != *src) && (1 == back) && (used < szdstb)) { + switch(dk4utf16_add(&dec, *(src++))) { + case DK4_EDSTM_FINISHED: { + c32 = dk4utf16_get(&dec); + dk4utf16_init(&dec); + sz = sizeof(buf); + if (0 != dk4utf8_encode(buf, &sz, c32, erp)) { + if ((szdstb - used) > sz) { + DK4_MEMCPY(&(dstb[used]),buf,sz); + used += sz; + } else { + back = 0; + dk4error_set_simple_error_code(erp, DK4_E_BUFFER_TOO_SMALL); + } + } else { + back = 0; + dk4error_set_elsize_nelem(erp, DK4_E_SYNTAX, 2, rd); + } + } break; + case DK4_EDSTM_ERROR: { + back = 0; + dk4error_set_elsize_nelem(erp, DK4_E_SYNTAX, 2, rd); + } break; + } + if (0 != back) { rd++; } + } + if (used < szdstb) { + dstb[used] = '\0'; + } else { + dstb[szdstb - 1] = '\0'; + back = 0; + dk4error_set_simple_error_code(erp, DK4_E_BUFFER_TOO_SMALL); + } + if (0 == dk4utf16_is_empty(&dec)) { + back = 0; + dk4error_set_elsize_nelem(erp, DK4_E_SYNTAX, 2, rd); + } + } else { + dk4error_set_simple_error_code(erp, DK4_E_INVALID_ARGUMENTS); + } + return back; +} + -- cgit v1.2.3