PostgreSQL Source Code  git master
mbutils.c File Reference
#include "postgres.h"
#include "access/xact.h"
#include "catalog/namespace.h"
#include "mb/pg_wchar.h"
#include "utils/fmgrprotos.h"
#include "utils/memutils.h"
#include "varatt.h"
Include dependency graph for mbutils.c:

Go to the source code of this file.

Data Structures

struct  ConvProcInfo
 

Typedefs

typedef struct ConvProcInfo ConvProcInfo
 

Functions

static char * perform_default_encoding_conversion (const char *src, int len, bool is_client_to_server)
 
static int cliplen (const char *str, int len, int limit)
 
int PrepareClientEncoding (int encoding)
 
int SetClientEncoding (int encoding)
 
void InitializeClientEncoding (void)
 
int pg_get_client_encoding (void)
 
const char * pg_get_client_encoding_name (void)
 
unsigned char * pg_do_encoding_conversion (unsigned char *src, int len, int src_encoding, int dest_encoding)
 
int pg_do_encoding_conversion_buf (Oid proc, int src_encoding, int dest_encoding, unsigned char *src, int srclen, unsigned char *dest, int destlen, bool noError)
 
Datum pg_convert_to (PG_FUNCTION_ARGS)
 
Datum pg_convert_from (PG_FUNCTION_ARGS)
 
Datum pg_convert (PG_FUNCTION_ARGS)
 
Datum length_in_encoding (PG_FUNCTION_ARGS)
 
Datum pg_encoding_max_length_sql (PG_FUNCTION_ARGS)
 
char * pg_client_to_server (const char *s, int len)
 
char * pg_any_to_server (const char *s, int len, int encoding)
 
char * pg_server_to_client (const char *s, int len)
 
char * pg_server_to_any (const char *s, int len, int encoding)
 
void pg_unicode_to_server (pg_wchar c, unsigned char *s)
 
bool pg_unicode_to_server_noerror (pg_wchar c, unsigned char *s)
 
int pg_mb2wchar (const char *from, pg_wchar *to)
 
int pg_mb2wchar_with_len (const char *from, pg_wchar *to, int len)
 
int pg_encoding_mb2wchar_with_len (int encoding, const char *from, pg_wchar *to, int len)
 
int pg_wchar2mb (const pg_wchar *from, char *to)
 
int pg_wchar2mb_with_len (const pg_wchar *from, char *to, int len)
 
int pg_encoding_wchar2mb_with_len (int encoding, const pg_wchar *from, char *to, int len)
 
int pg_mblen (const char *mbstr)
 
int pg_dsplen (const char *mbstr)
 
int pg_mbstrlen (const char *mbstr)
 
int pg_mbstrlen_with_len (const char *mbstr, int limit)
 
int pg_mbcliplen (const char *mbstr, int len, int limit)
 
int pg_encoding_mbcliplen (int encoding, const char *mbstr, int len, int limit)
 
int pg_mbcharcliplen (const char *mbstr, int len, int limit)
 
void SetDatabaseEncoding (int encoding)
 
void SetMessageEncoding (int encoding)
 
int GetDatabaseEncoding (void)
 
const char * GetDatabaseEncodingName (void)
 
Datum getdatabaseencoding (PG_FUNCTION_ARGS)
 
Datum pg_client_encoding (PG_FUNCTION_ARGS)
 
Datum PG_char_to_encoding (PG_FUNCTION_ARGS)
 
Datum PG_encoding_to_char (PG_FUNCTION_ARGS)
 
int GetMessageEncoding (void)
 
static bool pg_generic_charinc (unsigned char *charptr, int len)
 
static bool pg_utf8_increment (unsigned char *charptr, int length)
 
static bool pg_eucjp_increment (unsigned char *charptr, int length)
 
mbcharacter_incrementer pg_database_encoding_character_incrementer (void)
 
int pg_database_encoding_max_length (void)
 
bool pg_verifymbstr (const char *mbstr, int len, bool noError)
 
bool pg_verify_mbstr (int encoding, const char *mbstr, int len, bool noError)
 
int pg_verify_mbstr_len (int encoding, const char *mbstr, int len, bool noError)
 
void check_encoding_conversion_args (int src_encoding, int dest_encoding, int len, int expected_src_encoding, int expected_dest_encoding)
 
void report_invalid_encoding (int encoding, const char *mbstr, int len)
 
void report_untranslatable_char (int src_encoding, int dest_encoding, const char *mbstr, int len)
 

Variables

static ListConvProcList = NIL
 
static FmgrInfoToServerConvProc = NULL
 
static FmgrInfoToClientConvProc = NULL
 
static FmgrInfoUtf8ToServerConvProc = NULL
 
static const pg_enc2nameClientEncoding = &pg_enc2name_tbl[PG_SQL_ASCII]
 
static const pg_enc2nameDatabaseEncoding = &pg_enc2name_tbl[PG_SQL_ASCII]
 
static const pg_enc2nameMessageEncoding = &pg_enc2name_tbl[PG_SQL_ASCII]
 
static bool backend_startup_complete = false
 
static int pending_client_encoding = PG_SQL_ASCII
 

Typedef Documentation

◆ ConvProcInfo

typedef struct ConvProcInfo ConvProcInfo

Function Documentation

◆ check_encoding_conversion_args()

void check_encoding_conversion_args ( int  src_encoding,
int  dest_encoding,
int  len,
int  expected_src_encoding,
int  expected_dest_encoding 
)

Definition at line 1669 of file mbutils.c.

1674 {
1675  if (!PG_VALID_ENCODING(src_encoding))
1676  elog(ERROR, "invalid source encoding ID: %d", src_encoding);
1677  if (src_encoding != expected_src_encoding && expected_src_encoding >= 0)
1678  elog(ERROR, "expected source encoding \"%s\", but got \"%s\"",
1679  pg_enc2name_tbl[expected_src_encoding].name,
1680  pg_enc2name_tbl[src_encoding].name);
1681  if (!PG_VALID_ENCODING(dest_encoding))
1682  elog(ERROR, "invalid destination encoding ID: %d", dest_encoding);
1683  if (dest_encoding != expected_dest_encoding && expected_dest_encoding >= 0)
1684  elog(ERROR, "expected destination encoding \"%s\", but got \"%s\"",
1685  pg_enc2name_tbl[expected_dest_encoding].name,
1686  pg_enc2name_tbl[dest_encoding].name);
1687  if (len < 0)
1688  elog(ERROR, "encoding conversion length must not be negative");
1689 }
#define ERROR
Definition: elog.h:39
#define elog(elevel,...)
Definition: elog.h:225
const pg_enc2name pg_enc2name_tbl[]
Definition: encnames.c:308
const void size_t len
#define PG_VALID_ENCODING(_enc)
Definition: pg_wchar.h:287
const char * name

References elog, ERROR, len, name, pg_enc2name_tbl, and PG_VALID_ENCODING.

◆ cliplen()

static int cliplen ( const char *  str,
int  len,
int  limit 
)
static

Definition at line 1150 of file mbutils.c.

1151 {
1152  int l = 0;
1153 
1154  len = Min(len, limit);
1155  while (l < len && str[l])
1156  l++;
1157  return l;
1158 }
#define Min(x, y)
Definition: c.h:958
const char * str

References len, Min, and str.

Referenced by pg_encoding_mbcliplen(), pg_mbcharcliplen(), and pgstat_clip_activity().

◆ getdatabaseencoding()

Datum getdatabaseencoding ( PG_FUNCTION_ARGS  )

Definition at line 1273 of file mbutils.c.

1274 {
1276 }
#define DirectFunctionCall1(func, arg1)
Definition: fmgr.h:641
static const pg_enc2name * DatabaseEncoding
Definition: mbutils.c:81
Datum namein(PG_FUNCTION_ARGS)
Definition: name.c:48
static Datum CStringGetDatum(const char *X)
Definition: postgres.h:350
const char * name
Definition: pg_wchar.h:341

References CStringGetDatum(), DatabaseEncoding, DirectFunctionCall1, pg_enc2name::name, and namein().

◆ GetDatabaseEncoding()

◆ GetDatabaseEncodingName()

◆ GetMessageEncoding()

int GetMessageEncoding ( void  )

Definition at line 1308 of file mbutils.c.

1309 {
1310  return MessageEncoding->encoding;
1311 }
static const pg_enc2name * MessageEncoding
Definition: mbutils.c:82

References pg_enc2name::encoding, and MessageEncoding.

◆ InitializeClientEncoding()

void InitializeClientEncoding ( void  )

Definition at line 281 of file mbutils.c.

282 {
283  int current_server_encoding;
284 
287 
290  {
291  /*
292  * Oops, the requested conversion is not available. We couldn't fail
293  * before, but we can now.
294  */
295  ereport(FATAL,
296  (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
297  errmsg("conversion between %s and %s is not supported",
300  }
301 
302  /*
303  * Also look up the UTF8-to-server conversion function if needed. Since
304  * the server encoding is fixed within any one backend process, we don't
305  * have to do this more than once.
306  */
307  current_server_encoding = GetDatabaseEncoding();
308  if (current_server_encoding != PG_UTF8 &&
309  current_server_encoding != PG_SQL_ASCII)
310  {
311  Oid utf8_to_server_proc;
312 
314  utf8_to_server_proc =
316  current_server_encoding);
317  /* If there's no such conversion, just leave the pointer as NULL */
318  if (OidIsValid(utf8_to_server_proc))
319  {
320  FmgrInfo *finfo;
321 
323  sizeof(FmgrInfo));
324  fmgr_info_cxt(utf8_to_server_proc, finfo,
326  /* Set Utf8ToServerConvProc only after data is fully valid */
327  Utf8ToServerConvProc = finfo;
328  }
329  }
330 }
#define Assert(condition)
Definition: c.h:812
#define OidIsValid(objectId)
Definition: c.h:729
int errcode(int sqlerrcode)
Definition: elog.c:853
int errmsg(const char *fmt,...)
Definition: elog.c:1070
#define FATAL
Definition: elog.h:41
#define ereport(elevel,...)
Definition: elog.h:149
void fmgr_info_cxt(Oid functionId, FmgrInfo *finfo, MemoryContext mcxt)
Definition: fmgr.c:137
int GetDatabaseEncoding(void)
Definition: mbutils.c:1261
static FmgrInfo * Utf8ToServerConvProc
Definition: mbutils.c:75
int SetClientEncoding(int encoding)
Definition: mbutils.c:208
int PrepareClientEncoding(int encoding)
Definition: mbutils.c:110
const char * GetDatabaseEncodingName(void)
Definition: mbutils.c:1267
static bool backend_startup_complete
Definition: mbutils.c:90
static int pending_client_encoding
Definition: mbutils.c:91
MemoryContext TopMemoryContext
Definition: mcxt.c:149
void * MemoryContextAlloc(MemoryContext context, Size size)
Definition: mcxt.c:1181
Oid FindDefaultConversionProc(int32 for_encoding, int32 to_encoding)
Definition: namespace.c:4080
@ PG_SQL_ASCII
Definition: pg_wchar.h:226
@ PG_UTF8
Definition: pg_wchar.h:232
unsigned int Oid
Definition: postgres_ext.h:31
Definition: fmgr.h:57
bool IsTransactionState(void)
Definition: xact.c:386

References Assert, backend_startup_complete, ereport, errcode(), errmsg(), FATAL, FindDefaultConversionProc(), fmgr_info_cxt(), GetDatabaseEncoding(), GetDatabaseEncodingName(), IsTransactionState(), MemoryContextAlloc(), name, OidIsValid, pending_client_encoding, pg_enc2name_tbl, PG_SQL_ASCII, PG_UTF8, PrepareClientEncoding(), SetClientEncoding(), TopMemoryContext, and Utf8ToServerConvProc.

Referenced by InitPostgres().

◆ length_in_encoding()

Datum length_in_encoding ( PG_FUNCTION_ARGS  )

Definition at line 615 of file mbutils.c.

616 {
617  bytea *string = PG_GETARG_BYTEA_PP(0);
618  char *src_encoding_name = NameStr(*PG_GETARG_NAME(1));
619  int src_encoding = pg_char_to_encoding(src_encoding_name);
620  const char *src_str;
621  int len;
622  int retval;
623 
624  if (src_encoding < 0)
625  ereport(ERROR,
626  (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
627  errmsg("invalid encoding name \"%s\"",
628  src_encoding_name)));
629 
630  len = VARSIZE_ANY_EXHDR(string);
631  src_str = VARDATA_ANY(string);
632 
633  retval = pg_verify_mbstr_len(src_encoding, src_str, len, false);
634 
635  PG_RETURN_INT32(retval);
636 }
#define NameStr(name)
Definition: c.h:700
#define PG_GETARG_BYTEA_PP(n)
Definition: fmgr.h:308
#define PG_GETARG_NAME(n)
Definition: fmgr.h:278
#define PG_RETURN_INT32(x)
Definition: fmgr.h:354
int pg_verify_mbstr_len(int encoding, const char *mbstr, int len, bool noError)
Definition: mbutils.c:1597
#define pg_char_to_encoding
Definition: pg_wchar.h:629
Definition: c.h:641
#define VARDATA_ANY(PTR)
Definition: varatt.h:324
#define VARSIZE_ANY_EXHDR(PTR)
Definition: varatt.h:317

References ereport, errcode(), errmsg(), ERROR, len, NameStr, pg_char_to_encoding, PG_GETARG_BYTEA_PP, PG_GETARG_NAME, PG_RETURN_INT32, pg_verify_mbstr_len(), VARDATA_ANY, and VARSIZE_ANY_EXHDR.

◆ perform_default_encoding_conversion()

static char * perform_default_encoding_conversion ( const char *  src,
int  len,
bool  is_client_to_server 
)
static

Definition at line 783 of file mbutils.c.

785 {
786  char *result;
787  int src_encoding,
788  dest_encoding;
789  FmgrInfo *flinfo;
790 
791  if (is_client_to_server)
792  {
793  src_encoding = ClientEncoding->encoding;
794  dest_encoding = DatabaseEncoding->encoding;
795  flinfo = ToServerConvProc;
796  }
797  else
798  {
799  src_encoding = DatabaseEncoding->encoding;
800  dest_encoding = ClientEncoding->encoding;
801  flinfo = ToClientConvProc;
802  }
803 
804  if (flinfo == NULL)
805  return unconstify(char *, src);
806 
807  /*
808  * Allocate space for conversion result, being wary of integer overflow.
809  * See comments in pg_do_encoding_conversion.
810  */
812  ereport(ERROR,
813  (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
814  errmsg("out of memory"),
815  errdetail("String of %d bytes is too long for encoding conversion.",
816  len)));
817 
818  result = (char *)
820  (Size) len * MAX_CONVERSION_GROWTH + 1);
821 
822  FunctionCall6(flinfo,
823  Int32GetDatum(src_encoding),
824  Int32GetDatum(dest_encoding),
825  CStringGetDatum(src),
826  CStringGetDatum(result),
828  BoolGetDatum(false));
829 
830  /*
831  * Release extra space if there might be a lot --- see comments in
832  * pg_do_encoding_conversion.
833  */
834  if (len > 1000000)
835  {
836  Size resultlen = strlen(result);
837 
838  if (resultlen >= MaxAllocSize)
839  ereport(ERROR,
840  (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
841  errmsg("out of memory"),
842  errdetail("String of %d bytes is too long for encoding conversion.",
843  len)));
844 
845  result = (char *) repalloc(result, resultlen + 1);
846  }
847 
848  return result;
849 }
#define unconstify(underlying_type, expr)
Definition: c.h:1199
size_t Size
Definition: c.h:559
int errdetail(const char *fmt,...)
Definition: elog.c:1203
#define MaxAllocSize
Definition: fe_memutils.h:22
#define FunctionCall6(flinfo, arg1, arg2, arg3, arg4, arg5, arg6)
Definition: fmgr.h:669
static const pg_enc2name * ClientEncoding
Definition: mbutils.c:80
static FmgrInfo * ToServerConvProc
Definition: mbutils.c:67
static FmgrInfo * ToClientConvProc
Definition: mbutils.c:68
MemoryContext CurrentMemoryContext
Definition: mcxt.c:143
void * repalloc(void *pointer, Size size)
Definition: mcxt.c:1541
void * MemoryContextAllocHuge(MemoryContext context, Size size)
Definition: mcxt.c:1640
#define MaxAllocHugeSize
Definition: memutils.h:45
#define MAX_CONVERSION_GROWTH
Definition: pg_wchar.h:302
static Datum BoolGetDatum(bool X)
Definition: postgres.h:102
static Datum Int32GetDatum(int32 X)
Definition: postgres.h:212

References BoolGetDatum(), ClientEncoding, CStringGetDatum(), CurrentMemoryContext, DatabaseEncoding, pg_enc2name::encoding, ereport, errcode(), errdetail(), errmsg(), ERROR, FunctionCall6, Int32GetDatum(), len, MAX_CONVERSION_GROWTH, MaxAllocHugeSize, MaxAllocSize, MemoryContextAllocHuge(), repalloc(), ToClientConvProc, ToServerConvProc, and unconstify.

Referenced by pg_any_to_server(), and pg_server_to_any().

◆ pg_any_to_server()

char* pg_any_to_server ( const char *  s,
int  len,
int  encoding 
)

Definition at line 676 of file mbutils.c.

677 {
678  if (len <= 0)
679  return unconstify(char *, s); /* empty string is always valid */
680 
683  {
684  /*
685  * No conversion is needed, but we must still validate the data.
686  */
687  (void) pg_verify_mbstr(DatabaseEncoding->encoding, s, len, false);
688  return unconstify(char *, s);
689  }
690 
692  {
693  /*
694  * No conversion is possible, but we must still validate the data,
695  * because the client-side code might have done string escaping using
696  * the selected client_encoding. If the client encoding is ASCII-safe
697  * then we just do a straight validation under that encoding. For an
698  * ASCII-unsafe encoding we have a problem: we dare not pass such data
699  * to the parser but we have no way to convert it. We compromise by
700  * rejecting the data if it contains any non-ASCII characters.
701  */
703  (void) pg_verify_mbstr(encoding, s, len, false);
704  else
705  {
706  int i;
707 
708  for (i = 0; i < len; i++)
709  {
710  if (s[i] == '\0' || IS_HIGHBIT_SET(s[i]))
711  ereport(ERROR,
712  (errcode(ERRCODE_CHARACTER_NOT_IN_REPERTOIRE),
713  errmsg("invalid byte value for encoding \"%s\": 0x%02x",
715  (unsigned char) s[i])));
716  }
717  }
718  return unconstify(char *, s);
719  }
720 
721  /* Fast path if we can use cached conversion function */
723  return perform_default_encoding_conversion(s, len, true);
724 
725  /* General case ... will not work outside transactions */
726  return (char *) pg_do_encoding_conversion((unsigned char *) unconstify(char *, s),
727  len,
728  encoding,
730 }
#define IS_HIGHBIT_SET(ch)
Definition: c.h:1109
int i
Definition: isn.c:72
unsigned char * pg_do_encoding_conversion(unsigned char *src, int len, int src_encoding, int dest_encoding)
Definition: mbutils.c:356
bool pg_verify_mbstr(int encoding, const char *mbstr, int len, bool noError)
Definition: mbutils.c:1566
static char * perform_default_encoding_conversion(const char *src, int len, bool is_client_to_server)
Definition: mbutils.c:783
int32 encoding
Definition: pg_database.h:41
#define PG_VALID_BE_ENCODING(_enc)
Definition: pg_wchar.h:281

References ClientEncoding, DatabaseEncoding, encoding, pg_enc2name::encoding, ereport, errcode(), errmsg(), ERROR, i, IS_HIGHBIT_SET, len, name, perform_default_encoding_conversion(), pg_do_encoding_conversion(), pg_enc2name_tbl, PG_SQL_ASCII, PG_VALID_BE_ENCODING, pg_verify_mbstr(), and unconstify.

Referenced by ASN1_STRING_to_text(), cache_single_string(), db_encoding_convert(), dsnowball_lexize(), pg_client_to_server(), pg_stat_statements_internal(), pgp_armor_headers(), PLyUnicode_Bytes(), read_extension_script_file(), tsearch_readline(), utf_u2e(), X509_NAME_to_cstring(), and xml_recv().

◆ PG_char_to_encoding()

Datum PG_char_to_encoding ( PG_FUNCTION_ARGS  )

Definition at line 1285 of file mbutils.c.

1286 {
1287  Name s = PG_GETARG_NAME(0);
1288 
1290 }
Definition: c.h:695

References NameStr, pg_char_to_encoding, PG_GETARG_NAME, and PG_RETURN_INT32.

◆ pg_client_encoding()

Datum pg_client_encoding ( PG_FUNCTION_ARGS  )

◆ pg_client_to_server()

char* pg_client_to_server ( const char *  s,
int  len 
)

Definition at line 660 of file mbutils.c.

661 {
663 }
char * pg_any_to_server(const char *s, int len, int encoding)
Definition: mbutils.c:676

References ClientEncoding, pg_enc2name::encoding, len, and pg_any_to_server().

Referenced by exec_bind_message(), parse_fcall_arguments(), pq_getmsgstring(), and pq_getmsgtext().

◆ pg_convert()

Datum pg_convert ( PG_FUNCTION_ARGS  )

Definition at line 553 of file mbutils.c.

554 {
555  bytea *string = PG_GETARG_BYTEA_PP(0);
556  char *src_encoding_name = NameStr(*PG_GETARG_NAME(1));
557  int src_encoding = pg_char_to_encoding(src_encoding_name);
558  char *dest_encoding_name = NameStr(*PG_GETARG_NAME(2));
559  int dest_encoding = pg_char_to_encoding(dest_encoding_name);
560  const char *src_str;
561  char *dest_str;
562  bytea *retval;
563  int len;
564 
565  if (src_encoding < 0)
566  ereport(ERROR,
567  (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
568  errmsg("invalid source encoding name \"%s\"",
569  src_encoding_name)));
570  if (dest_encoding < 0)
571  ereport(ERROR,
572  (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
573  errmsg("invalid destination encoding name \"%s\"",
574  dest_encoding_name)));
575 
576  /* make sure that source string is valid */
577  len = VARSIZE_ANY_EXHDR(string);
578  src_str = VARDATA_ANY(string);
579  (void) pg_verify_mbstr(src_encoding, src_str, len, false);
580 
581  /* perform conversion */
582  dest_str = (char *) pg_do_encoding_conversion((unsigned char *) unconstify(char *, src_str),
583  len,
584  src_encoding,
585  dest_encoding);
586 
587 
588  /* return source string if no conversion happened */
589  if (dest_str == src_str)
590  PG_RETURN_BYTEA_P(string);
591 
592  /*
593  * build bytea data type structure.
594  */
595  len = strlen(dest_str);
596  retval = (bytea *) palloc(len + VARHDRSZ);
597  SET_VARSIZE(retval, len + VARHDRSZ);
598  memcpy(VARDATA(retval), dest_str, len);
599  pfree(dest_str);
600 
601  /* free memory if allocated by the toaster */
602  PG_FREE_IF_COPY(string, 0);
603 
604  PG_RETURN_BYTEA_P(retval);
605 }
#define VARHDRSZ
Definition: c.h:646
#define PG_FREE_IF_COPY(ptr, n)
Definition: fmgr.h:260
#define PG_RETURN_BYTEA_P(x)
Definition: fmgr.h:371
void pfree(void *pointer)
Definition: mcxt.c:1521
void * palloc(Size size)
Definition: mcxt.c:1317
#define VARDATA(PTR)
Definition: varatt.h:278
#define SET_VARSIZE(PTR, len)
Definition: varatt.h:305

References ereport, errcode(), errmsg(), ERROR, len, NameStr, palloc(), pfree(), pg_char_to_encoding, pg_do_encoding_conversion(), PG_FREE_IF_COPY, PG_GETARG_BYTEA_PP, PG_GETARG_NAME, PG_RETURN_BYTEA_P, pg_verify_mbstr(), SET_VARSIZE, unconstify, VARDATA, VARDATA_ANY, VARHDRSZ, and VARSIZE_ANY_EXHDR.

Referenced by pg_convert_from(), and pg_convert_to().

◆ pg_convert_from()

Datum pg_convert_from ( PG_FUNCTION_ARGS  )

Definition at line 526 of file mbutils.c.

527 {
528  Datum string = PG_GETARG_DATUM(0);
529  Datum src_encoding_name = PG_GETARG_DATUM(1);
530  Datum dest_encoding_name = DirectFunctionCall1(namein,
532  Datum result;
533 
534  result = DirectFunctionCall3(pg_convert, string,
535  src_encoding_name, dest_encoding_name);
536 
537  /*
538  * pg_convert returns a bytea, which we in turn return as text, relying on
539  * the fact that they are both in fact varlena types, and thus
540  * structurally identical. Although not all bytea values are valid text,
541  * in this case it will be because we've told pg_convert to return one
542  * that is valid as text in the current database encoding.
543  */
544  PG_RETURN_DATUM(result);
545 }
#define PG_GETARG_DATUM(n)
Definition: fmgr.h:268
#define PG_RETURN_DATUM(x)
Definition: fmgr.h:353
#define DirectFunctionCall3(func, arg1, arg2, arg3)
Definition: fmgr.h:645
Datum pg_convert(PG_FUNCTION_ARGS)
Definition: mbutils.c:553
uintptr_t Datum
Definition: postgres.h:64

References CStringGetDatum(), DatabaseEncoding, DirectFunctionCall1, DirectFunctionCall3, pg_enc2name::name, namein(), pg_convert(), PG_GETARG_DATUM, and PG_RETURN_DATUM.

◆ pg_convert_to()

Datum pg_convert_to ( PG_FUNCTION_ARGS  )

Definition at line 501 of file mbutils.c.

502 {
503  Datum string = PG_GETARG_DATUM(0);
504  Datum dest_encoding_name = PG_GETARG_DATUM(1);
505  Datum src_encoding_name = DirectFunctionCall1(namein,
507  Datum result;
508 
509  /*
510  * pg_convert expects a bytea as its first argument. We're passing it a
511  * text argument here, relying on the fact that they are both in fact
512  * varlena types, and thus structurally identical.
513  */
514  result = DirectFunctionCall3(pg_convert, string,
515  src_encoding_name, dest_encoding_name);
516 
517  PG_RETURN_DATUM(result);
518 }

References CStringGetDatum(), DatabaseEncoding, DirectFunctionCall1, DirectFunctionCall3, pg_enc2name::name, namein(), pg_convert(), PG_GETARG_DATUM, and PG_RETURN_DATUM.

◆ pg_database_encoding_character_incrementer()

mbcharacter_incrementer pg_database_encoding_character_incrementer ( void  )

Definition at line 1523 of file mbutils.c.

1524 {
1525  /*
1526  * Eventually it might be best to add a field to pg_wchar_table[], but for
1527  * now we just use a switch.
1528  */
1529  switch (GetDatabaseEncoding())
1530  {
1531  case PG_UTF8:
1532  return pg_utf8_increment;
1533 
1534  case PG_EUC_JP:
1535  return pg_eucjp_increment;
1536 
1537  default:
1538  return pg_generic_charinc;
1539  }
1540 }
static bool pg_generic_charinc(unsigned char *charptr, int len)
Definition: mbutils.c:1325
static bool pg_utf8_increment(unsigned char *charptr, int length)
Definition: mbutils.c:1359
static bool pg_eucjp_increment(unsigned char *charptr, int length)
Definition: mbutils.c:1437
@ PG_EUC_JP
Definition: pg_wchar.h:227

References GetDatabaseEncoding(), PG_EUC_JP, pg_eucjp_increment(), pg_generic_charinc(), PG_UTF8, and pg_utf8_increment().

Referenced by make_greater_string().

◆ pg_database_encoding_max_length()

◆ pg_do_encoding_conversion()

unsigned char* pg_do_encoding_conversion ( unsigned char *  src,
int  len,
int  src_encoding,
int  dest_encoding 
)

Definition at line 356 of file mbutils.c.

358 {
359  unsigned char *result;
360  Oid proc;
361 
362  if (len <= 0)
363  return src; /* empty string is always valid */
364 
365  if (src_encoding == dest_encoding)
366  return src; /* no conversion required, assume valid */
367 
368  if (dest_encoding == PG_SQL_ASCII)
369  return src; /* any string is valid in SQL_ASCII */
370 
371  if (src_encoding == PG_SQL_ASCII)
372  {
373  /* No conversion is possible, but we must validate the result */
374  (void) pg_verify_mbstr(dest_encoding, (const char *) src, len, false);
375  return src;
376  }
377 
378  if (!IsTransactionState()) /* shouldn't happen */
379  elog(ERROR, "cannot perform encoding conversion outside a transaction");
380 
381  proc = FindDefaultConversionProc(src_encoding, dest_encoding);
382  if (!OidIsValid(proc))
383  ereport(ERROR,
384  (errcode(ERRCODE_UNDEFINED_FUNCTION),
385  errmsg("default conversion function for encoding \"%s\" to \"%s\" does not exist",
386  pg_encoding_to_char(src_encoding),
387  pg_encoding_to_char(dest_encoding))));
388 
389  /*
390  * Allocate space for conversion result, being wary of integer overflow.
391  *
392  * len * MAX_CONVERSION_GROWTH is typically a vast overestimate of the
393  * required space, so it might exceed MaxAllocSize even though the result
394  * would actually fit. We do not want to hand back a result string that
395  * exceeds MaxAllocSize, because callers might not cope gracefully --- but
396  * if we just allocate more than that, and don't use it, that's fine.
397  */
399  ereport(ERROR,
400  (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
401  errmsg("out of memory"),
402  errdetail("String of %d bytes is too long for encoding conversion.",
403  len)));
404 
405  result = (unsigned char *)
407  (Size) len * MAX_CONVERSION_GROWTH + 1);
408 
409  (void) OidFunctionCall6(proc,
410  Int32GetDatum(src_encoding),
411  Int32GetDatum(dest_encoding),
412  CStringGetDatum((char *) src),
413  CStringGetDatum((char *) result),
415  BoolGetDatum(false));
416 
417  /*
418  * If the result is large, it's worth repalloc'ing to release any extra
419  * space we asked for. The cutoff here is somewhat arbitrary, but we
420  * *must* check when len * MAX_CONVERSION_GROWTH exceeds MaxAllocSize.
421  */
422  if (len > 1000000)
423  {
424  Size resultlen = strlen((char *) result);
425 
426  if (resultlen >= MaxAllocSize)
427  ereport(ERROR,
428  (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
429  errmsg("out of memory"),
430  errdetail("String of %d bytes is too long for encoding conversion.",
431  len)));
432 
433  result = (unsigned char *) repalloc(result, resultlen + 1);
434  }
435 
436  return result;
437 }
#define OidFunctionCall6(functionId, arg1, arg2, arg3, arg4, arg5, arg6)
Definition: fmgr.h:689
#define pg_encoding_to_char
Definition: pg_wchar.h:630

References BoolGetDatum(), CStringGetDatum(), CurrentMemoryContext, elog, ereport, errcode(), errdetail(), errmsg(), ERROR, FindDefaultConversionProc(), Int32GetDatum(), IsTransactionState(), len, MAX_CONVERSION_GROWTH, MaxAllocHugeSize, MaxAllocSize, MemoryContextAllocHuge(), OidFunctionCall6, OidIsValid, pg_encoding_to_char, PG_SQL_ASCII, pg_verify_mbstr(), and repalloc().

Referenced by convert_charset(), pg_any_to_server(), pg_convert(), and pg_server_to_any().

◆ pg_do_encoding_conversion_buf()

int pg_do_encoding_conversion_buf ( Oid  proc,
int  src_encoding,
int  dest_encoding,
unsigned char *  src,
int  srclen,
unsigned char *  dest,
int  destlen,
bool  noError 
)

Definition at line 469 of file mbutils.c.

475 {
476  Datum result;
477 
478  /*
479  * If the destination buffer is not large enough to hold the result in the
480  * worst case, limit the input size passed to the conversion function.
481  */
482  if ((Size) srclen >= ((destlen - 1) / (Size) MAX_CONVERSION_GROWTH))
483  srclen = ((destlen - 1) / (Size) MAX_CONVERSION_GROWTH);
484 
485  result = OidFunctionCall6(proc,
486  Int32GetDatum(src_encoding),
487  Int32GetDatum(dest_encoding),
488  CStringGetDatum((char *) src),
489  CStringGetDatum((char *) dest),
490  Int32GetDatum(srclen),
491  BoolGetDatum(noError));
492  return DatumGetInt32(result);
493 }
static int32 DatumGetInt32(Datum X)
Definition: postgres.h:202

References BoolGetDatum(), CStringGetDatum(), DatumGetInt32(), generate_unaccent_rules::dest, Int32GetDatum(), MAX_CONVERSION_GROWTH, and OidFunctionCall6.

Referenced by CopyConversionError(), CopyConvertBuf(), and test_enc_conversion().

◆ pg_dsplen()

int pg_dsplen ( const char *  mbstr)

Definition at line 1030 of file mbutils.c.

1031 {
1032  return pg_wchar_table[DatabaseEncoding->encoding].dsplen((const unsigned char *) mbstr);
1033 }
mbdisplaylen_converter dsplen
Definition: pg_wchar.h:383

References DatabaseEncoding, pg_wchar_tbl::dsplen, pg_enc2name::encoding, and pg_wchar_table.

Referenced by p_isspecial().

◆ pg_encoding_max_length_sql()

Datum pg_encoding_max_length_sql ( PG_FUNCTION_ARGS  )

Definition at line 644 of file mbutils.c.

645 {
646  int encoding = PG_GETARG_INT32(0);
647 
650  else
651  PG_RETURN_NULL();
652 }
#define PG_RETURN_NULL()
Definition: fmgr.h:345
#define PG_GETARG_INT32(n)
Definition: fmgr.h:269

References encoding, PG_GETARG_INT32, PG_RETURN_INT32, PG_RETURN_NULL, PG_VALID_ENCODING, and pg_wchar_table.

◆ pg_encoding_mb2wchar_with_len()

int pg_encoding_mb2wchar_with_len ( int  encoding,
const char *  from,
pg_wchar to,
int  len 
)

Definition at line 993 of file mbutils.c.

995 {
996  return pg_wchar_table[encoding].mb2wchar_with_len((const unsigned char *) from, to, len);
997 }
mb2wchar_with_len_converter mb2wchar_with_len
Definition: pg_wchar.h:378

References encoding, len, pg_wchar_tbl::mb2wchar_with_len, and pg_wchar_table.

◆ pg_encoding_mbcliplen()

int pg_encoding_mbcliplen ( int  encoding,
const char *  mbstr,
int  len,
int  limit 
)

Definition at line 1093 of file mbutils.c.

1095 {
1096  mblen_converter mblen_fn;
1097  int clen = 0;
1098  int l;
1099 
1100  /* optimization for single byte encoding */
1101  if (pg_encoding_max_length(encoding) == 1)
1102  return cliplen(mbstr, len, limit);
1103 
1104  mblen_fn = pg_wchar_table[encoding].mblen;
1105 
1106  while (len > 0 && *mbstr)
1107  {
1108  l = (*mblen_fn) ((const unsigned char *) mbstr);
1109  if ((clen + l) > limit)
1110  break;
1111  clen += l;
1112  if (clen == limit)
1113  break;
1114  len -= l;
1115  mbstr += l;
1116  }
1117  return clen;
1118 }
static int cliplen(const char *str, int len, int limit)
Definition: mbutils.c:1150
int(* mblen_converter)(const unsigned char *mbstr)
Definition: pg_wchar.h:366
mblen_converter mblen
Definition: pg_wchar.h:382
int pg_encoding_max_length(int encoding)
Definition: wchar.c:2127

References cliplen(), encoding, len, pg_wchar_tbl::mblen, pg_encoding_max_length(), and pg_wchar_table.

Referenced by pg_mbcliplen().

◆ PG_encoding_to_char()

Datum PG_encoding_to_char ( PG_FUNCTION_ARGS  )

Definition at line 1293 of file mbutils.c.

1294 {
1296  const char *encoding_name = pg_encoding_to_char(encoding);
1297 
1298  return DirectFunctionCall1(namein, CStringGetDatum(encoding_name));
1299 }
int32_t int32
Definition: c.h:481

References CStringGetDatum(), DirectFunctionCall1, encoding, namein(), pg_encoding_to_char, and PG_GETARG_INT32.

◆ pg_encoding_wchar2mb_with_len()

int pg_encoding_wchar2mb_with_len ( int  encoding,
const pg_wchar from,
char *  to,
int  len 
)

Definition at line 1015 of file mbutils.c.

1017 {
1018  return pg_wchar_table[encoding].wchar2mb_with_len(from, (unsigned char *) to, len);
1019 }
wchar2mb_with_len_converter wchar2mb_with_len
Definition: pg_wchar.h:380

References encoding, len, pg_wchar_table, and pg_wchar_tbl::wchar2mb_with_len.

◆ pg_eucjp_increment()

static bool pg_eucjp_increment ( unsigned char *  charptr,
int  length 
)
static

Definition at line 1437 of file mbutils.c.

1438 {
1439  unsigned char c1,
1440  c2;
1441  int i;
1442 
1443  c1 = *charptr;
1444 
1445  switch (c1)
1446  {
1447  case SS2: /* JIS X 0201 */
1448  if (length != 2)
1449  return false;
1450 
1451  c2 = charptr[1];
1452 
1453  if (c2 >= 0xdf)
1454  charptr[0] = charptr[1] = 0xa1;
1455  else if (c2 < 0xa1)
1456  charptr[1] = 0xa1;
1457  else
1458  charptr[1]++;
1459  break;
1460 
1461  case SS3: /* JIS X 0212 */
1462  if (length != 3)
1463  return false;
1464 
1465  for (i = 2; i > 0; i--)
1466  {
1467  c2 = charptr[i];
1468  if (c2 < 0xa1)
1469  {
1470  charptr[i] = 0xa1;
1471  return true;
1472  }
1473  else if (c2 < 0xfe)
1474  {
1475  charptr[i]++;
1476  return true;
1477  }
1478  }
1479 
1480  /* Out of 3-byte code region */
1481  return false;
1482 
1483  default:
1484  if (IS_HIGHBIT_SET(c1)) /* JIS X 0208? */
1485  {
1486  if (length != 2)
1487  return false;
1488 
1489  for (i = 1; i >= 0; i--)
1490  {
1491  c2 = charptr[i];
1492  if (c2 < 0xa1)
1493  {
1494  charptr[i] = 0xa1;
1495  return true;
1496  }
1497  else if (c2 < 0xfe)
1498  {
1499  charptr[i]++;
1500  return true;
1501  }
1502  }
1503 
1504  /* Out of 2 byte code region */
1505  return false;
1506  }
1507  else
1508  { /* ASCII, single byte */
1509  if (c1 > 0x7e)
1510  return false;
1511  (*charptr)++;
1512  }
1513  break;
1514  }
1515 
1516  return true;
1517 }
#define SS2
Definition: pg_wchar.h:38
#define SS3
Definition: pg_wchar.h:39

References i, IS_HIGHBIT_SET, SS2, and SS3.

Referenced by pg_database_encoding_character_incrementer().

◆ pg_generic_charinc()

static bool pg_generic_charinc ( unsigned char *  charptr,
int  len 
)
static

Definition at line 1325 of file mbutils.c.

1326 {
1327  unsigned char *lastbyte = charptr + len - 1;
1328  mbchar_verifier mbverify;
1329 
1330  /* We can just invoke the character verifier directly. */
1332 
1333  while (*lastbyte < (unsigned char) 255)
1334  {
1335  (*lastbyte)++;
1336  if ((*mbverify) (charptr, len) == len)
1337  return true;
1338  }
1339 
1340  return false;
1341 }
int(* mbchar_verifier)(const unsigned char *mbstr, int len)
Definition: pg_wchar.h:372
mbchar_verifier mbverifychar
Definition: pg_wchar.h:384

References GetDatabaseEncoding(), len, pg_wchar_tbl::mbverifychar, and pg_wchar_table.

Referenced by pg_database_encoding_character_incrementer().

◆ pg_get_client_encoding()

int pg_get_client_encoding ( void  )

Definition at line 336 of file mbutils.c.

337 {
338  return ClientEncoding->encoding;
339 }

References ClientEncoding, and pg_enc2name::encoding.

Referenced by BeginCopyFrom(), BeginCopyTo(), and xml_send().

◆ pg_get_client_encoding_name()

const char* pg_get_client_encoding_name ( void  )

Definition at line 345 of file mbutils.c.

346 {
347  return ClientEncoding->name;
348 }

References ClientEncoding, and pg_enc2name::name.

◆ pg_mb2wchar()

int pg_mb2wchar ( const char *  from,
pg_wchar to 
)

Definition at line 979 of file mbutils.c.

980 {
981  return pg_wchar_table[DatabaseEncoding->encoding].mb2wchar_with_len((const unsigned char *) from, to, strlen(from));
982 }

References DatabaseEncoding, pg_enc2name::encoding, pg_wchar_tbl::mb2wchar_with_len, and pg_wchar_table.

◆ pg_mb2wchar_with_len()

◆ pg_mbcharcliplen()

int pg_mbcharcliplen ( const char *  mbstr,
int  len,
int  limit 
)

Definition at line 1125 of file mbutils.c.

1126 {
1127  int clen = 0;
1128  int nch = 0;
1129  int l;
1130 
1131  /* optimization for single byte encoding */
1133  return cliplen(mbstr, len, limit);
1134 
1135  while (len > 0 && *mbstr)
1136  {
1137  l = pg_mblen(mbstr);
1138  nch++;
1139  if (nch > limit)
1140  break;
1141  clen += l;
1142  len -= l;
1143  mbstr += l;
1144  }
1145  return clen;
1146 }
int pg_database_encoding_max_length(void)
Definition: mbutils.c:1546
int pg_mblen(const char *mbstr)
Definition: mbutils.c:1023

References cliplen(), len, pg_database_encoding_max_length(), and pg_mblen().

Referenced by bpchar(), bpchar_input(), text_left(), text_right(), varchar(), and varchar_input().

◆ pg_mbcliplen()

◆ pg_mblen()

int pg_mblen ( const char *  mbstr)

Definition at line 1023 of file mbutils.c.

1024 {
1025  return pg_wchar_table[DatabaseEncoding->encoding].mblen((const unsigned char *) mbstr);
1026 }

References DatabaseEncoding, pg_enc2name::encoding, pg_wchar_tbl::mblen, and pg_wchar_table.

Referenced by addCompoundAffixFlagValue(), bit_in(), charlen_to_bytelen(), DCH_from_char(), dotrim(), find_word(), findchar(), findchar2(), findwrd(), gbt_var_node_cp_len(), get_modifiers(), get_nextfield(), get_wildcard_part(), getlexeme(), getNextFlagFromString(), gettoken_query(), gettoken_query_standard(), gettoken_query_websearch(), gettoken_tsvector(), hex_decode_safe(), infix(), initTrie(), lpad(), make_trigrams(), map_sql_identifier_to_xml_name(), map_xml_name_to_sql_identifier(), match_prosrc_to_literal(), mb_strchr(), NIImportAffixes(), NIImportDictionary(), NIImportOOAffixes(), NUM_eat_non_data_chars(), NUM_processor(), parse_affentry(), parse_format(), parse_lquery(), parse_ltree(), parse_or_operator(), parse_re_flags(), parse_test_flags(), pg_base64_decode(), pg_mbcharcliplen(), pg_mbstrlen(), pg_mbstrlen_with_len(), prssyntaxerror(), readstoplist(), report_json_context(), rpad(), RS_compile(), RS_execute(), RS_isRegis(), similar_escape_internal(), split_text(), t_isalnum(), t_isalpha(), t_isdigit(), t_isprint(), t_isspace(), text_format(), text_position_next(), text_reverse(), text_substring(), text_to_bits(), textregexreplace(), thesaurusRead(), TParserGet(), translate(), ts_stat_sql(), tsvectorout(), unaccent_lexize(), varbit_in(), varstr_levenshtein(), and wchareq().

◆ pg_mbstrlen()

int pg_mbstrlen ( const char *  mbstr)

Definition at line 1037 of file mbutils.c.

1038 {
1039  int len = 0;
1040 
1041  /* optimization for single byte encoding */
1043  return strlen(mbstr);
1044 
1045  while (*mbstr)
1046  {
1047  mbstr += pg_mblen(mbstr);
1048  len++;
1049  }
1050  return len;
1051 }

References len, pg_database_encoding_max_length(), and pg_mblen().

Referenced by NUM_processor(), and text_format_append_string().

◆ pg_mbstrlen_with_len()

int pg_mbstrlen_with_len ( const char *  mbstr,
int  limit 
)

Definition at line 1057 of file mbutils.c.

1058 {
1059  int len = 0;
1060 
1061  /* optimization for single byte encoding */
1063  return limit;
1064 
1065  while (limit > 0 && *mbstr)
1066  {
1067  int l = pg_mblen(mbstr);
1068 
1069  limit -= l;
1070  mbstr += l;
1071  len++;
1072  }
1073  return len;
1074 }

References len, pg_database_encoding_max_length(), and pg_mblen().

Referenced by bpchar(), bpchar_input(), bpcharlen(), executor_errposition(), lpad(), match_prosrc_to_query(), parser_errposition(), plpgsql_scanner_errposition(), rpad(), similar_escape_internal(), text_left(), text_length(), text_position_get_match_pos(), text_right(), text_substring(), unicode_assigned(), unicode_is_normalized(), unicode_normalize_func(), and varstr_levenshtein().

◆ pg_server_to_any()

char* pg_server_to_any ( const char *  s,
int  len,
int  encoding 
)

Definition at line 749 of file mbutils.c.

750 {
751  if (len <= 0)
752  return unconstify(char *, s); /* empty string is always valid */
753 
756  return unconstify(char *, s); /* assume data is valid */
757 
759  {
760  /* No conversion is possible, but we must validate the result */
761  (void) pg_verify_mbstr(encoding, s, len, false);
762  return unconstify(char *, s);
763  }
764 
765  /* Fast path if we can use cached conversion function */
767  return perform_default_encoding_conversion(s, len, false);
768 
769  /* General case ... will not work outside transactions */
770  return (char *) pg_do_encoding_conversion((unsigned char *) unconstify(char *, s),
771  len,
773  encoding);
774 }

References ClientEncoding, DatabaseEncoding, encoding, pg_enc2name::encoding, len, perform_default_encoding_conversion(), pg_do_encoding_conversion(), PG_SQL_ASCII, pg_verify_mbstr(), and unconstify.

Referenced by compareStrings(), CopyAttributeOutCSV(), CopyAttributeOutText(), daitch_mokotoff(), DoCopyTo(), dsnowball_lexize(), hv_fetch_string(), hv_store_string(), pg_server_to_client(), PLyUnicode_FromStringAndSize(), and utf_e2u().

◆ pg_server_to_client()

char* pg_server_to_client ( const char *  s,
int  len 
)

Definition at line 738 of file mbutils.c.

739 {
741 }
char * pg_server_to_any(const char *s, int len, int encoding)
Definition: mbutils.c:749

References ClientEncoding, pg_enc2name::encoding, len, and pg_server_to_any().

Referenced by pq_puttextmessage(), pq_sendcountedtext(), pq_sendstring(), pq_sendtext(), and pq_writestring().

◆ pg_unicode_to_server()

void pg_unicode_to_server ( pg_wchar  c,
unsigned char *  s 
)

Definition at line 864 of file mbutils.c.

865 {
866  unsigned char c_as_utf8[MAX_MULTIBYTE_CHAR_LEN + 1];
867  int c_as_utf8_len;
868  int server_encoding;
869 
870  /*
871  * Complain if invalid Unicode code point. The choice of errcode here is
872  * debatable, but really our caller should have checked this anyway.
873  */
875  ereport(ERROR,
876  (errcode(ERRCODE_SYNTAX_ERROR),
877  errmsg("invalid Unicode code point")));
878 
879  /* Otherwise, if it's in ASCII range, conversion is trivial */
880  if (c <= 0x7F)
881  {
882  s[0] = (unsigned char) c;
883  s[1] = '\0';
884  return;
885  }
886 
887  /* If the server encoding is UTF-8, we just need to reformat the code */
888  server_encoding = GetDatabaseEncoding();
889  if (server_encoding == PG_UTF8)
890  {
891  unicode_to_utf8(c, s);
892  s[pg_utf_mblen(s)] = '\0';
893  return;
894  }
895 
896  /* For all other cases, we must have a conversion function available */
897  if (Utf8ToServerConvProc == NULL)
898  ereport(ERROR,
899  (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
900  errmsg("conversion between %s and %s is not supported",
903 
904  /* Construct UTF-8 source string */
905  unicode_to_utf8(c, c_as_utf8);
906  c_as_utf8_len = pg_utf_mblen(c_as_utf8);
907  c_as_utf8[c_as_utf8_len] = '\0';
908 
909  /* Convert, or throw error if we can't */
912  Int32GetDatum(server_encoding),
913  CStringGetDatum((char *) c_as_utf8),
914  CStringGetDatum((char *) s),
915  Int32GetDatum(c_as_utf8_len),
916  BoolGetDatum(false));
917 }
#define MAX_MULTIBYTE_CHAR_LEN
Definition: pg_wchar.h:33
#define pg_utf_mblen
Definition: pg_wchar.h:633
static unsigned char * unicode_to_utf8(pg_wchar c, unsigned char *utf8string)
Definition: pg_wchar.h:575
static bool is_valid_unicode_codepoint(pg_wchar c)
Definition: pg_wchar.h:519
char * c

References BoolGetDatum(), CStringGetDatum(), ereport, errcode(), errmsg(), ERROR, FunctionCall6, GetDatabaseEncoding(), GetDatabaseEncodingName(), Int32GetDatum(), is_valid_unicode_codepoint(), MAX_MULTIBYTE_CHAR_LEN, name, pg_enc2name_tbl, PG_UTF8, pg_utf_mblen, unicode_to_utf8(), and Utf8ToServerConvProc.

Referenced by map_xml_name_to_sql_identifier(), str_udeescape(), and unistr().

◆ pg_unicode_to_server_noerror()

bool pg_unicode_to_server_noerror ( pg_wchar  c,
unsigned char *  s 
)

Definition at line 926 of file mbutils.c.

927 {
928  unsigned char c_as_utf8[MAX_MULTIBYTE_CHAR_LEN + 1];
929  int c_as_utf8_len;
930  int converted_len;
931  int server_encoding;
932 
933  /* Fail if invalid Unicode code point */
935  return false;
936 
937  /* Otherwise, if it's in ASCII range, conversion is trivial */
938  if (c <= 0x7F)
939  {
940  s[0] = (unsigned char) c;
941  s[1] = '\0';
942  return true;
943  }
944 
945  /* If the server encoding is UTF-8, we just need to reformat the code */
946  server_encoding = GetDatabaseEncoding();
947  if (server_encoding == PG_UTF8)
948  {
949  unicode_to_utf8(c, s);
950  s[pg_utf_mblen(s)] = '\0';
951  return true;
952  }
953 
954  /* For all other cases, we must have a conversion function available */
955  if (Utf8ToServerConvProc == NULL)
956  return false;
957 
958  /* Construct UTF-8 source string */
959  unicode_to_utf8(c, c_as_utf8);
960  c_as_utf8_len = pg_utf_mblen(c_as_utf8);
961  c_as_utf8[c_as_utf8_len] = '\0';
962 
963  /* Convert, but without throwing error if we can't */
966  Int32GetDatum(server_encoding),
967  CStringGetDatum((char *) c_as_utf8),
968  CStringGetDatum((char *) s),
969  Int32GetDatum(c_as_utf8_len),
970  BoolGetDatum(true)));
971 
972  /* Conversion was successful iff it consumed the whole input */
973  return (converted_len == c_as_utf8_len);
974 }

References BoolGetDatum(), CStringGetDatum(), DatumGetInt32(), FunctionCall6, GetDatabaseEncoding(), Int32GetDatum(), is_valid_unicode_codepoint(), MAX_MULTIBYTE_CHAR_LEN, PG_UTF8, pg_utf_mblen, unicode_to_utf8(), and Utf8ToServerConvProc.

Referenced by json_lex_string().

◆ pg_utf8_increment()

static bool pg_utf8_increment ( unsigned char *  charptr,
int  length 
)
static

Definition at line 1359 of file mbutils.c.

1360 {
1361  unsigned char a;
1362  unsigned char limit;
1363 
1364  switch (length)
1365  {
1366  default:
1367  /* reject lengths 5 and 6 for now */
1368  return false;
1369  case 4:
1370  a = charptr[3];
1371  if (a < 0xBF)
1372  {
1373  charptr[3]++;
1374  break;
1375  }
1376  /* FALL THRU */
1377  case 3:
1378  a = charptr[2];
1379  if (a < 0xBF)
1380  {
1381  charptr[2]++;
1382  break;
1383  }
1384  /* FALL THRU */
1385  case 2:
1386  a = charptr[1];
1387  switch (*charptr)
1388  {
1389  case 0xED:
1390  limit = 0x9F;
1391  break;
1392  case 0xF4:
1393  limit = 0x8F;
1394  break;
1395  default:
1396  limit = 0xBF;
1397  break;
1398  }
1399  if (a < limit)
1400  {
1401  charptr[1]++;
1402  break;
1403  }
1404  /* FALL THRU */
1405  case 1:
1406  a = *charptr;
1407  if (a == 0x7F || a == 0xDF || a == 0xEF || a == 0xF4)
1408  return false;
1409  charptr[0]++;
1410  break;
1411  }
1412 
1413  return true;
1414 }
int a
Definition: isn.c:68

References a.

Referenced by pg_database_encoding_character_incrementer().

◆ pg_verify_mbstr()

bool pg_verify_mbstr ( int  encoding,
const char *  mbstr,
int  len,
bool  noError 
)

Definition at line 1566 of file mbutils.c.

1567 {
1568  int oklen;
1569 
1571 
1572  oklen = pg_wchar_table[encoding].mbverifystr((const unsigned char *) mbstr, len);
1573  if (oklen != len)
1574  {
1575  if (noError)
1576  return false;
1577  report_invalid_encoding(encoding, mbstr + oklen, len - oklen);
1578  }
1579  return true;
1580 }
void report_invalid_encoding(int encoding, const char *mbstr, int len)
Definition: mbutils.c:1698
mbstr_verifier mbverifystr
Definition: pg_wchar.h:385

References Assert, encoding, len, pg_wchar_tbl::mbverifystr, PG_VALID_ENCODING, pg_wchar_table, and report_invalid_encoding().

Referenced by AddFileToBackupManifest(), LogicalOutputWrite(), pg_any_to_server(), pg_convert(), pg_do_encoding_conversion(), pg_server_to_any(), pg_verifymbstr(), and read_extension_script_file().

◆ pg_verify_mbstr_len()

int pg_verify_mbstr_len ( int  encoding,
const char *  mbstr,
int  len,
bool  noError 
)

Definition at line 1597 of file mbutils.c.

1598 {
1599  mbchar_verifier mbverifychar;
1600  int mb_len;
1601 
1603 
1604  /*
1605  * In single-byte encodings, we need only reject nulls (\0).
1606  */
1607  if (pg_encoding_max_length(encoding) <= 1)
1608  {
1609  const char *nullpos = memchr(mbstr, 0, len);
1610 
1611  if (nullpos == NULL)
1612  return len;
1613  if (noError)
1614  return -1;
1615  report_invalid_encoding(encoding, nullpos, 1);
1616  }
1617 
1618  /* fetch function pointer just once */
1619  mbverifychar = pg_wchar_table[encoding].mbverifychar;
1620 
1621  mb_len = 0;
1622 
1623  while (len > 0)
1624  {
1625  int l;
1626 
1627  /* fast path for ASCII-subset characters */
1628  if (!IS_HIGHBIT_SET(*mbstr))
1629  {
1630  if (*mbstr != '\0')
1631  {
1632  mb_len++;
1633  mbstr++;
1634  len--;
1635  continue;
1636  }
1637  if (noError)
1638  return -1;
1640  }
1641 
1642  l = (*mbverifychar) ((const unsigned char *) mbstr, len);
1643 
1644  if (l < 0)
1645  {
1646  if (noError)
1647  return -1;
1649  }
1650 
1651  mbstr += l;
1652  len -= l;
1653  mb_len++;
1654  }
1655  return mb_len;
1656 }

References Assert, encoding, IS_HIGHBIT_SET, len, pg_wchar_tbl::mbverifychar, pg_encoding_max_length(), PG_VALID_ENCODING, pg_wchar_table, and report_invalid_encoding().

Referenced by length_in_encoding().

◆ pg_verifymbstr()

bool pg_verifymbstr ( const char *  mbstr,
int  len,
bool  noError 
)

◆ pg_wchar2mb()

int pg_wchar2mb ( const pg_wchar from,
char *  to 
)

Definition at line 1001 of file mbutils.c.

1002 {
1003  return pg_wchar_table[DatabaseEncoding->encoding].wchar2mb_with_len(from, (unsigned char *) to, pg_wchar_strlen(from));
1004 }
size_t pg_wchar_strlen(const pg_wchar *str)
Definition: wstrncmp.c:70

References DatabaseEncoding, pg_enc2name::encoding, pg_wchar_strlen(), pg_wchar_table, and pg_wchar_tbl::wchar2mb_with_len.

◆ pg_wchar2mb_with_len()

int pg_wchar2mb_with_len ( const pg_wchar from,
char *  to,
int  len 
)

◆ PrepareClientEncoding()

int PrepareClientEncoding ( int  encoding)

Definition at line 110 of file mbutils.c.

111 {
112  int current_server_encoding;
113  ListCell *lc;
114 
116  return -1;
117 
118  /* Can't do anything during startup, per notes above */
120  return 0;
121 
122  current_server_encoding = GetDatabaseEncoding();
123 
124  /*
125  * Check for cases that require no conversion function.
126  */
127  if (current_server_encoding == encoding ||
128  current_server_encoding == PG_SQL_ASCII ||
130  return 0;
131 
132  if (IsTransactionState())
133  {
134  /*
135  * If we're in a live transaction, it's safe to access the catalogs,
136  * so look up the functions. We repeat the lookup even if the info is
137  * already cached, so that we can react to changes in the contents of
138  * pg_conversion.
139  */
140  Oid to_server_proc,
141  to_client_proc;
142  ConvProcInfo *convinfo;
143  MemoryContext oldcontext;
144 
145  to_server_proc = FindDefaultConversionProc(encoding,
146  current_server_encoding);
147  if (!OidIsValid(to_server_proc))
148  return -1;
149  to_client_proc = FindDefaultConversionProc(current_server_encoding,
150  encoding);
151  if (!OidIsValid(to_client_proc))
152  return -1;
153 
154  /*
155  * Load the fmgr info into TopMemoryContext (could still fail here)
156  */
158  sizeof(ConvProcInfo));
159  convinfo->s_encoding = current_server_encoding;
160  convinfo->c_encoding = encoding;
161  fmgr_info_cxt(to_server_proc, &convinfo->to_server_info,
163  fmgr_info_cxt(to_client_proc, &convinfo->to_client_info,
165 
166  /* Attach new info to head of list */
168  ConvProcList = lcons(convinfo, ConvProcList);
169  MemoryContextSwitchTo(oldcontext);
170 
171  /*
172  * We cannot yet remove any older entry for the same encoding pair,
173  * since it could still be in use. SetClientEncoding will clean up.
174  */
175 
176  return 0; /* success */
177  }
178  else
179  {
180  /*
181  * If we're not in a live transaction, the only thing we can do is
182  * restore a previous setting using the cache. This covers all
183  * transaction-rollback cases. The only case it might not work for is
184  * trying to change client_encoding on the fly by editing
185  * postgresql.conf and SIGHUP'ing. Which would probably be a stupid
186  * thing to do anyway.
187  */
188  foreach(lc, ConvProcList)
189  {
190  ConvProcInfo *oldinfo = (ConvProcInfo *) lfirst(lc);
191 
192  if (oldinfo->s_encoding == current_server_encoding &&
193  oldinfo->c_encoding == encoding)
194  return 0;
195  }
196 
197  return -1; /* it's not cached, so fail */
198  }
199 }
List * lcons(void *datum, List *list)
Definition: list.c:495
static List * ConvProcList
Definition: mbutils.c:61
#define lfirst(lc)
Definition: pg_list.h:172
#define PG_VALID_FE_ENCODING(_enc)
Definition: pg_wchar.h:291
MemoryContextSwitchTo(old_ctx)
int s_encoding
Definition: mbutils.c:55
FmgrInfo to_client_info
Definition: mbutils.c:58
int c_encoding
Definition: mbutils.c:56
FmgrInfo to_server_info
Definition: mbutils.c:57

References backend_startup_complete, ConvProcInfo::c_encoding, ConvProcList, encoding, FindDefaultConversionProc(), fmgr_info_cxt(), GetDatabaseEncoding(), IsTransactionState(), lcons(), lfirst, MemoryContextAlloc(), MemoryContextSwitchTo(), OidIsValid, PG_SQL_ASCII, PG_VALID_FE_ENCODING, ConvProcInfo::s_encoding, ConvProcInfo::to_client_info, ConvProcInfo::to_server_info, and TopMemoryContext.

Referenced by check_client_encoding(), and InitializeClientEncoding().

◆ report_invalid_encoding()

void report_invalid_encoding ( int  encoding,
const char *  mbstr,
int  len 
)

Definition at line 1698 of file mbutils.c.

1699 {
1700  int l = pg_encoding_mblen(encoding, mbstr);
1701  char buf[8 * 5 + 1];
1702  char *p = buf;
1703  int j,
1704  jlimit;
1705 
1706  jlimit = Min(l, len);
1707  jlimit = Min(jlimit, 8); /* prevent buffer overrun */
1708 
1709  for (j = 0; j < jlimit; j++)
1710  {
1711  p += sprintf(p, "0x%02x", (unsigned char) mbstr[j]);
1712  if (j < jlimit - 1)
1713  p += sprintf(p, " ");
1714  }
1715 
1716  ereport(ERROR,
1717  (errcode(ERRCODE_CHARACTER_NOT_IN_REPERTOIRE),
1718  errmsg("invalid byte sequence for encoding \"%s\": %s",
1720  buf)));
1721 }
int j
Definition: isn.c:73
static char * buf
Definition: pg_test_fsync.c:72
#define sprintf
Definition: port.h:240
int pg_encoding_mblen(int encoding, const char *mbstr)
Definition: wchar.c:2069

References buf, encoding, ereport, errcode(), errmsg(), ERROR, j, len, Min, name, pg_enc2name_tbl, pg_encoding_mblen(), and sprintf.

Referenced by big52euc_tw(), big52mic(), CopyConversionError(), euc_cn2mic(), euc_jis_20042shift_jis_2004(), euc_jp2mic(), euc_jp2sjis(), euc_kr2mic(), euc_tw2big5(), euc_tw2mic(), iso8859_1_to_utf8(), latin2mic(), latin2mic_with_table(), local2local(), LocalToUtf(), mic2big5(), mic2euc_cn(), mic2euc_jp(), mic2euc_kr(), mic2euc_tw(), mic2latin(), mic2latin_with_table(), mic2sjis(), pg_verify_mbstr(), pg_verify_mbstr_len(), shift_jis_20042euc_jis_2004(), sjis2euc_jp(), sjis2mic(), test_enc_conversion(), utf8_to_iso8859_1(), and UtfToLocal().

◆ report_untranslatable_char()

void report_untranslatable_char ( int  src_encoding,
int  dest_encoding,
const char *  mbstr,
int  len 
)

Definition at line 1730 of file mbutils.c.

1732 {
1733  int l = pg_encoding_mblen(src_encoding, mbstr);
1734  char buf[8 * 5 + 1];
1735  char *p = buf;
1736  int j,
1737  jlimit;
1738 
1739  jlimit = Min(l, len);
1740  jlimit = Min(jlimit, 8); /* prevent buffer overrun */
1741 
1742  for (j = 0; j < jlimit; j++)
1743  {
1744  p += sprintf(p, "0x%02x", (unsigned char) mbstr[j]);
1745  if (j < jlimit - 1)
1746  p += sprintf(p, " ");
1747  }
1748 
1749  ereport(ERROR,
1750  (errcode(ERRCODE_UNTRANSLATABLE_CHARACTER),
1751  errmsg("character with byte sequence %s in encoding \"%s\" has no equivalent in encoding \"%s\"",
1752  buf,
1753  pg_enc2name_tbl[src_encoding].name,
1754  pg_enc2name_tbl[dest_encoding].name)));
1755 }

References buf, ereport, errcode(), errmsg(), ERROR, j, len, Min, name, pg_enc2name_tbl, pg_encoding_mblen(), and sprintf.

Referenced by big52euc_tw(), big52mic(), euc_tw2big5(), latin2mic_with_table(), local2local(), LocalToUtf(), mic2big5(), mic2euc_cn(), mic2euc_jp(), mic2euc_kr(), mic2euc_tw(), mic2latin(), mic2latin_with_table(), mic2sjis(), utf8_to_iso8859_1(), and UtfToLocal().

◆ SetClientEncoding()

int SetClientEncoding ( int  encoding)

Definition at line 208 of file mbutils.c.

209 {
210  int current_server_encoding;
211  bool found;
212  ListCell *lc;
213 
215  return -1;
216 
217  /* Can't do anything during startup, per notes above */
219  {
221  return 0;
222  }
223 
224  current_server_encoding = GetDatabaseEncoding();
225 
226  /*
227  * Check for cases that require no conversion function.
228  */
229  if (current_server_encoding == encoding ||
230  current_server_encoding == PG_SQL_ASCII ||
232  {
234  ToServerConvProc = NULL;
235  ToClientConvProc = NULL;
236  return 0;
237  }
238 
239  /*
240  * Search the cache for the entry previously prepared by
241  * PrepareClientEncoding; if there isn't one, we lose. While at it,
242  * release any duplicate entries so that repeated Prepare/Set cycles don't
243  * leak memory.
244  */
245  found = false;
246  foreach(lc, ConvProcList)
247  {
248  ConvProcInfo *convinfo = (ConvProcInfo *) lfirst(lc);
249 
250  if (convinfo->s_encoding == current_server_encoding &&
251  convinfo->c_encoding == encoding)
252  {
253  if (!found)
254  {
255  /* Found newest entry, so set up */
257  ToServerConvProc = &convinfo->to_server_info;
258  ToClientConvProc = &convinfo->to_client_info;
259  found = true;
260  }
261  else
262  {
263  /* Duplicate entry, release it */
265  pfree(convinfo);
266  }
267  }
268  }
269 
270  if (found)
271  return 0; /* success */
272  else
273  return -1; /* it's not cached, so fail */
274 }
#define foreach_delete_current(lst, var_or_cell)
Definition: pg_list.h:391

References backend_startup_complete, ConvProcInfo::c_encoding, ClientEncoding, ConvProcList, encoding, foreach_delete_current, GetDatabaseEncoding(), lfirst, pending_client_encoding, pfree(), pg_enc2name_tbl, PG_SQL_ASCII, PG_VALID_FE_ENCODING, ConvProcInfo::s_encoding, ConvProcInfo::to_client_info, ConvProcInfo::to_server_info, ToClientConvProc, and ToServerConvProc.

Referenced by assign_client_encoding(), InitializeClientEncoding(), and ParallelWorkerMain().

◆ SetDatabaseEncoding()

void SetDatabaseEncoding ( int  encoding)

Definition at line 1161 of file mbutils.c.

1162 {
1164  elog(ERROR, "invalid database encoding: %d", encoding);
1165 
1168 }

References Assert, DatabaseEncoding, elog, encoding, pg_enc2name::encoding, ERROR, pg_enc2name_tbl, and PG_VALID_BE_ENCODING.

Referenced by CheckMyDatabase().

◆ SetMessageEncoding()

void SetMessageEncoding ( int  encoding)

Definition at line 1171 of file mbutils.c.

1172 {
1173  /* Some calls happen before we can elog()! */
1175 
1178 }

References Assert, encoding, pg_enc2name::encoding, MessageEncoding, pg_enc2name_tbl, and PG_VALID_ENCODING.

Referenced by pg_perm_setlocale().

Variable Documentation

◆ backend_startup_complete

bool backend_startup_complete = false
static

Definition at line 90 of file mbutils.c.

Referenced by InitializeClientEncoding(), PrepareClientEncoding(), and SetClientEncoding().

◆ ClientEncoding

◆ ConvProcList

List* ConvProcList = NIL
static

Definition at line 61 of file mbutils.c.

Referenced by PrepareClientEncoding(), and SetClientEncoding().

◆ DatabaseEncoding

◆ MessageEncoding

const pg_enc2name* MessageEncoding = &pg_enc2name_tbl[PG_SQL_ASCII]
static

Definition at line 82 of file mbutils.c.

Referenced by GetMessageEncoding(), and SetMessageEncoding().

◆ pending_client_encoding

int pending_client_encoding = PG_SQL_ASCII
static

Definition at line 91 of file mbutils.c.

Referenced by InitializeClientEncoding(), and SetClientEncoding().

◆ ToClientConvProc

FmgrInfo* ToClientConvProc = NULL
static

Definition at line 68 of file mbutils.c.

Referenced by perform_default_encoding_conversion(), and SetClientEncoding().

◆ ToServerConvProc

FmgrInfo* ToServerConvProc = NULL
static

Definition at line 67 of file mbutils.c.

Referenced by perform_default_encoding_conversion(), and SetClientEncoding().

◆ Utf8ToServerConvProc

FmgrInfo* Utf8ToServerConvProc = NULL
static