jbe@0: /*
jbe@0:  *  Copyright (c) 2006, FlexiGuided GmbH, Berlin, Germany
jbe@0:  *  Author: Jan Behrens <jan.behrens@flexiguided.de>
jbe@0:  *  All rights reserved.
jbe@0:  *
jbe@0:  *  Redistribution and use in source and binary forms, with or without
jbe@0:  *  modification, are permitted provided that the following conditions are
jbe@0:  *  met:
jbe@0:  *
jbe@0:  *  1. Redistributions of source code must retain the above copyright
jbe@0:  *     notice, this list of conditions and the following disclaimer.
jbe@0:  *  2. Redistributions in binary form must reproduce the above copyright
jbe@0:  *     notice, this list of conditions and the following disclaimer in the
jbe@0:  *     documentation and/or other materials provided with the distribution.
jbe@0:  *  3. Neither the name of the FlexiGuided GmbH nor the names of its
jbe@0:  *     contributors may be used to endorse or promote products derived from
jbe@0:  *     this software without specific prior written permission.
jbe@0:  *
jbe@0:  *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
jbe@0:  *  "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
jbe@0:  *  LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
jbe@0:  *  PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
jbe@0:  *  OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
jbe@0:  *  EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
jbe@0:  *  PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
jbe@0:  *  PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
jbe@0:  *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
jbe@0:  *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
jbe@0:  *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
jbe@0:  *
jbe@0:  */
jbe@0:  
jbe@0: 
jbe@0: /*
jbe@0:  *  File name:    pgsql/utf8proc_pgsql.c
jbe@3:  *  Version:      1.0
jbe@3:  *  Last changed: 2006-09-17
jbe@0:  *
jbe@0:  *  Description:
jbe@0:  *  PostgreSQL extension to provide a function 'unifold', which can be used
jbe@0:  *  to case-fold and normalize index fields.
jbe@0:  */
jbe@0: 
jbe@0: 
jbe@0: #include "../utf8proc.c"
jbe@0: 
jbe@0: #include <postgres.h>
jbe@0: #include <utils/elog.h>
jbe@0: #include <fmgr.h>
jbe@1: #include <string.h>
jbe@0: #include <unistd.h>
jbe@0: #include <utils/builtins.h>
jbe@0: 
jbe@5: #ifdef PG_MODULE_MAGIC
jbe@5: PG_MODULE_MAGIC;
jbe@5: #endif
jbe@5: 
jbe@2: #define UTF8PROC_PGSQL_OPTS ( UTF8PROC_REJECTNA | UTF8PROC_COMPAT | \
jbe@2:   UTF8PROC_COMPOSE | UTF8PROC_STABLE | UTF8PROC_IGNORE | UTF8PROC_STRIPCC | \
jbe@3:   UTF8PROC_NLF2LF | UTF8PROC_CASEFOLD | UTF8PROC_LUMP )
jbe@1: 
jbe@0: PG_FUNCTION_INFO_V1(utf8proc_pgsql_unifold);
jbe@0: Datum utf8proc_pgsql_unifold(PG_FUNCTION_ARGS) {
jbe@1:   text *input_string;
jbe@1:   text *output_string = NULL;
jbe@0:   ssize_t result;
jbe@1:   input_string = PG_GETARG_TEXT_P(0);
jbe@1:   do {
jbe@1:     result = utf8proc_decompose(
jbe@1:       VARDATA(input_string), VARSIZE(input_string) - VARHDRSZ,
jbe@1:       NULL, 0, UTF8PROC_PGSQL_OPTS
jbe@1:     );
jbe@1:     if (result < 0) break;
jbe@1:     if (result > (SIZE_MAX-1-VARHDRSZ)/sizeof(int32_t)) {
jbe@1:       result = UTF8PROC_ERROR_OVERFLOW;
jbe@1:       break;
jbe@1:     }
jbe@1:     output_string = palloc(result * sizeof(int32_t) + 1 + VARHDRSZ);
jbe@1:     // reserve one extra byte for termination
jbe@1:     if (!output_string) {
jbe@1:       result = UTF8PROC_ERROR_NOMEM;
jbe@1:       break;
jbe@1:     }
jbe@1:     result = utf8proc_decompose(
jbe@1:       VARDATA(input_string), VARSIZE(input_string) - VARHDRSZ,
jbe@1:       (int32_t *)VARDATA(output_string), result, UTF8PROC_PGSQL_OPTS);
jbe@1:     if (result < 0) break;
jbe@1:     result = utf8proc_reencode((int32_t *)VARDATA(output_string), result,
jbe@1:       UTF8PROC_PGSQL_OPTS);
jbe@1:   } while (0);
jbe@1:   PG_FREE_IF_COPY(input_string, 0);
jbe@0:   if (result < 0) {
jbe@0:     int sqlerrcode;
jbe@1:     if (output_string) pfree(output_string);
jbe@0:     switch(result) {
jbe@0:       case UTF8PROC_ERROR_NOMEM:
jbe@0:       sqlerrcode = ERRCODE_OUT_OF_MEMORY; break;
jbe@0:       case UTF8PROC_ERROR_OVERFLOW:
jbe@0:       sqlerrcode = ERRCODE_PROGRAM_LIMIT_EXCEEDED; break;
jbe@0:       case UTF8PROC_ERROR_INVALIDUTF8:
jbe@0:       case UTF8PROC_ERROR_NOTASSIGNED:
jbe@1:       PG_RETURN_NULL();
jbe@0:       default:
jbe@0:       sqlerrcode = ERRCODE_INTERNAL_ERROR;
jbe@0:     }
jbe@0:     ereport(ERROR, (
jbe@0:       errcode(sqlerrcode),
jbe@0:       errmsg("%s", utf8proc_errmsg(result))
jbe@0:     ));
jbe@1:   } else {
jbe@1:     VARATT_SIZEP(output_string) = result + VARHDRSZ;
jbe@1:     PG_RETURN_TEXT_P(output_string);
jbe@0:   }
jbe@1:   PG_RETURN_NULL();  // prohibit compiler warning
jbe@0: }
jbe@0: 
jbe@0: