From b45242fd30ffa6e1e7f490cc400ecbd966880f41 Mon Sep 17 00:00:00 2001 From: Michael Paquier Date: Wed, 2 Jul 2025 09:52:21 +0900 Subject: [PATCH] Move code for the bytea data type from varlena.c to new bytea.c This commit moves all the routines related to the bytea data type into its own new file, called bytea.c, clearing some of the bloat in varlena.c. This includes the routines for: - Input, output, receive and send - Comparison - Casts to integer types - bytea-specific functions The internals of the routines moved here are unchanged, with one exception. This comes with a twist in bytea_string_agg_transfn(), where the call to makeStringAggState() is replaced by the internals of this routine, still located in varlena.c. This simplifies the move to the new file by not having to expose makeStringAggState(). Author: Aleksander Alekseev Reviewed-by: Peter Eisentraut Discussion: https://postgr.es/m/CAJ7c6TMPVPJ5DL447zDz5ydctB8OmuviURtSwd=PHCRFEPDEAQ@mail.gmail.com --- src/backend/utils/adt/Makefile | 1 + src/backend/utils/adt/bytea.c | 1143 +++++++++++++++++++++++++++ src/backend/utils/adt/meson.build | 1 + src/backend/utils/adt/varlena.c | 1214 ++--------------------------- 4 files changed, 1203 insertions(+), 1156 deletions(-) create mode 100644 src/backend/utils/adt/bytea.c diff --git a/src/backend/utils/adt/Makefile b/src/backend/utils/adt/Makefile index 4a233b63c32..ffeacf2b819 100644 --- a/src/backend/utils/adt/Makefile +++ b/src/backend/utils/adt/Makefile @@ -23,6 +23,7 @@ OBJS = \ arrayutils.o \ ascii.o \ bool.o \ + bytea.o \ cash.o \ char.o \ cryptohashfuncs.o \ diff --git a/src/backend/utils/adt/bytea.c b/src/backend/utils/adt/bytea.c new file mode 100644 index 00000000000..2e539c2504e --- /dev/null +++ b/src/backend/utils/adt/bytea.c @@ -0,0 +1,1143 @@ +/*------------------------------------------------------------------------- + * + * bytea.c + * Functions for the bytea type. + * + * Portions Copyright (c) 2025, PostgreSQL Global Development Group + * + * + * IDENTIFICATION + * src/backend/utils/adt/bytea.c + * + *------------------------------------------------------------------------- + */ + +#include "postgres.h" + +#include "access/detoast.h" +#include "catalog/pg_collation_d.h" +#include "catalog/pg_type_d.h" +#include "common/int.h" +#include "fmgr.h" +#include "libpq/pqformat.h" +#include "port/pg_bitutils.h" +#include "utils/builtins.h" +#include "utils/bytea.h" +#include "utils/fmgrprotos.h" +#include "utils/memutils.h" +#include "utils/sortsupport.h" +#include "utils/varlena.h" +#include "varatt.h" + +/* GUC variable */ +int bytea_output = BYTEA_OUTPUT_HEX; + +static bytea *bytea_catenate(bytea *t1, bytea *t2); +static bytea *bytea_substring(Datum str, int S, int L, + bool length_not_specified); +static bytea *bytea_overlay(bytea *t1, bytea *t2, int sp, int sl); + +/* + * bytea_catenate + * Guts of byteacat(), broken out so it can be used by other functions + * + * Arguments can be in short-header form, but not compressed or out-of-line + */ +static bytea * +bytea_catenate(bytea *t1, bytea *t2) +{ + bytea *result; + int len1, + len2, + len; + char *ptr; + + len1 = VARSIZE_ANY_EXHDR(t1); + len2 = VARSIZE_ANY_EXHDR(t2); + + /* paranoia ... probably should throw error instead? */ + if (len1 < 0) + len1 = 0; + if (len2 < 0) + len2 = 0; + + len = len1 + len2 + VARHDRSZ; + result = (bytea *) palloc(len); + + /* Set size of result string... */ + SET_VARSIZE(result, len); + + /* Fill data field of result string... */ + ptr = VARDATA(result); + if (len1 > 0) + memcpy(ptr, VARDATA_ANY(t1), len1); + if (len2 > 0) + memcpy(ptr + len1, VARDATA_ANY(t2), len2); + + return result; +} + +#define PG_STR_GET_BYTEA(str_) \ + DatumGetByteaPP(DirectFunctionCall1(byteain, CStringGetDatum(str_))) + +static bytea * +bytea_substring(Datum str, + int S, + int L, + bool length_not_specified) +{ + int32 S1; /* adjusted start position */ + int32 L1; /* adjusted substring length */ + int32 E; /* end position */ + + /* + * The logic here should generally match text_substring(). + */ + S1 = Max(S, 1); + + if (length_not_specified) + { + /* + * Not passed a length - DatumGetByteaPSlice() grabs everything to the + * end of the string if we pass it a negative value for length. + */ + L1 = -1; + } + else if (L < 0) + { + /* SQL99 says to throw an error for E < S, i.e., negative length */ + ereport(ERROR, + (errcode(ERRCODE_SUBSTRING_ERROR), + errmsg("negative substring length not allowed"))); + L1 = -1; /* silence stupider compilers */ + } + else if (pg_add_s32_overflow(S, L, &E)) + { + /* + * L could be large enough for S + L to overflow, in which case the + * substring must run to end of string. + */ + L1 = -1; + } + else + { + /* + * A zero or negative value for the end position can happen if the + * start was negative or one. SQL99 says to return a zero-length + * string. + */ + if (E < 1) + return PG_STR_GET_BYTEA(""); + + L1 = E - S1; + } + + /* + * If the start position is past the end of the string, SQL99 says to + * return a zero-length string -- DatumGetByteaPSlice() will do that for + * us. We need only convert S1 to zero-based starting position. + */ + return DatumGetByteaPSlice(str, S1 - 1, L1); +} + +static bytea * +bytea_overlay(bytea *t1, bytea *t2, int sp, int sl) +{ + bytea *result; + bytea *s1; + bytea *s2; + int sp_pl_sl; + + /* + * Check for possible integer-overflow cases. For negative sp, throw a + * "substring length" error because that's what should be expected + * according to the spec's definition of OVERLAY(). + */ + if (sp <= 0) + ereport(ERROR, + (errcode(ERRCODE_SUBSTRING_ERROR), + errmsg("negative substring length not allowed"))); + if (pg_add_s32_overflow(sp, sl, &sp_pl_sl)) + ereport(ERROR, + (errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), + errmsg("integer out of range"))); + + s1 = bytea_substring(PointerGetDatum(t1), 1, sp - 1, false); + s2 = bytea_substring(PointerGetDatum(t1), sp_pl_sl, -1, true); + result = bytea_catenate(s1, t2); + result = bytea_catenate(result, s2); + + return result; +} + +/***************************************************************************** + * USER I/O ROUTINES * + *****************************************************************************/ + +#define VAL(CH) ((CH) - '0') +#define DIG(VAL) ((VAL) + '0') + +/* + * byteain - converts from printable representation of byte array + * + * Non-printable characters must be passed as '\nnn' (octal) and are + * converted to internal form. '\' must be passed as '\\'. + * ereport(ERROR, ...) if bad form. + * + * BUGS: + * The input is scanned twice. + * The error checking of input is minimal. + */ +Datum +byteain(PG_FUNCTION_ARGS) +{ + char *inputText = PG_GETARG_CSTRING(0); + Node *escontext = fcinfo->context; + char *tp; + char *rp; + int bc; + bytea *result; + + /* Recognize hex input */ + if (inputText[0] == '\\' && inputText[1] == 'x') + { + size_t len = strlen(inputText); + + bc = (len - 2) / 2 + VARHDRSZ; /* maximum possible length */ + result = palloc(bc); + bc = hex_decode_safe(inputText + 2, len - 2, VARDATA(result), + escontext); + SET_VARSIZE(result, bc + VARHDRSZ); /* actual length */ + + PG_RETURN_BYTEA_P(result); + } + + /* Else, it's the traditional escaped style */ + for (bc = 0, tp = inputText; *tp != '\0'; bc++) + { + if (tp[0] != '\\') + tp++; + else if ((tp[0] == '\\') && + (tp[1] >= '0' && tp[1] <= '3') && + (tp[2] >= '0' && tp[2] <= '7') && + (tp[3] >= '0' && tp[3] <= '7')) + tp += 4; + else if ((tp[0] == '\\') && + (tp[1] == '\\')) + tp += 2; + else + { + /* + * one backslash, not followed by another or ### valid octal + */ + ereturn(escontext, (Datum) 0, + (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION), + errmsg("invalid input syntax for type %s", "bytea"))); + } + } + + bc += VARHDRSZ; + + result = (bytea *) palloc(bc); + SET_VARSIZE(result, bc); + + tp = inputText; + rp = VARDATA(result); + while (*tp != '\0') + { + if (tp[0] != '\\') + *rp++ = *tp++; + else if ((tp[0] == '\\') && + (tp[1] >= '0' && tp[1] <= '3') && + (tp[2] >= '0' && tp[2] <= '7') && + (tp[3] >= '0' && tp[3] <= '7')) + { + bc = VAL(tp[1]); + bc <<= 3; + bc += VAL(tp[2]); + bc <<= 3; + *rp++ = bc + VAL(tp[3]); + + tp += 4; + } + else if ((tp[0] == '\\') && + (tp[1] == '\\')) + { + *rp++ = '\\'; + tp += 2; + } + else + { + /* + * We should never get here. The first pass should not allow it. + */ + ereturn(escontext, (Datum) 0, + (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION), + errmsg("invalid input syntax for type %s", "bytea"))); + } + } + + PG_RETURN_BYTEA_P(result); +} + +/* + * byteaout - converts to printable representation of byte array + * + * In the traditional escaped format, non-printable characters are + * printed as '\nnn' (octal) and '\' as '\\'. + */ +Datum +byteaout(PG_FUNCTION_ARGS) +{ + bytea *vlena = PG_GETARG_BYTEA_PP(0); + char *result; + char *rp; + + if (bytea_output == BYTEA_OUTPUT_HEX) + { + /* Print hex format */ + rp = result = palloc(VARSIZE_ANY_EXHDR(vlena) * 2 + 2 + 1); + *rp++ = '\\'; + *rp++ = 'x'; + rp += hex_encode(VARDATA_ANY(vlena), VARSIZE_ANY_EXHDR(vlena), rp); + } + else if (bytea_output == BYTEA_OUTPUT_ESCAPE) + { + /* Print traditional escaped format */ + char *vp; + uint64 len; + int i; + + len = 1; /* empty string has 1 char */ + vp = VARDATA_ANY(vlena); + for (i = VARSIZE_ANY_EXHDR(vlena); i != 0; i--, vp++) + { + if (*vp == '\\') + len += 2; + else if ((unsigned char) *vp < 0x20 || (unsigned char) *vp > 0x7e) + len += 4; + else + len++; + } + + /* + * In principle len can't overflow uint32 if the input fit in 1GB, but + * for safety let's check rather than relying on palloc's internal + * check. + */ + if (len > MaxAllocSize) + ereport(ERROR, + (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED), + errmsg_internal("result of bytea output conversion is too large"))); + rp = result = (char *) palloc(len); + + vp = VARDATA_ANY(vlena); + for (i = VARSIZE_ANY_EXHDR(vlena); i != 0; i--, vp++) + { + if (*vp == '\\') + { + *rp++ = '\\'; + *rp++ = '\\'; + } + else if ((unsigned char) *vp < 0x20 || (unsigned char) *vp > 0x7e) + { + int val; /* holds unprintable chars */ + + val = *vp; + rp[0] = '\\'; + rp[3] = DIG(val & 07); + val >>= 3; + rp[2] = DIG(val & 07); + val >>= 3; + rp[1] = DIG(val & 03); + rp += 4; + } + else + *rp++ = *vp; + } + } + else + { + elog(ERROR, "unrecognized \"bytea_output\" setting: %d", + bytea_output); + rp = result = NULL; /* keep compiler quiet */ + } + *rp = '\0'; + PG_RETURN_CSTRING(result); +} + +/* + * bytearecv - converts external binary format to bytea + */ +Datum +bytearecv(PG_FUNCTION_ARGS) +{ + StringInfo buf = (StringInfo) PG_GETARG_POINTER(0); + bytea *result; + int nbytes; + + nbytes = buf->len - buf->cursor; + result = (bytea *) palloc(nbytes + VARHDRSZ); + SET_VARSIZE(result, nbytes + VARHDRSZ); + pq_copymsgbytes(buf, VARDATA(result), nbytes); + PG_RETURN_BYTEA_P(result); +} + +/* + * byteasend - converts bytea to binary format + * + * This is a special case: just copy the input... + */ +Datum +byteasend(PG_FUNCTION_ARGS) +{ + bytea *vlena = PG_GETARG_BYTEA_P_COPY(0); + + PG_RETURN_BYTEA_P(vlena); +} + +Datum +bytea_string_agg_transfn(PG_FUNCTION_ARGS) +{ + StringInfo state; + + state = PG_ARGISNULL(0) ? NULL : (StringInfo) PG_GETARG_POINTER(0); + + /* Append the value unless null, preceding it with the delimiter. */ + if (!PG_ARGISNULL(1)) + { + bytea *value = PG_GETARG_BYTEA_PP(1); + bool isfirst = false; + + /* + * You might think we can just throw away the first delimiter, however + * we must keep it as we may be a parallel worker doing partial + * aggregation building a state to send to the main process. We need + * to keep the delimiter of every aggregation so that the combine + * function can properly join up the strings of two separately + * partially aggregated results. The first delimiter is only stripped + * off in the final function. To know how much to strip off the front + * of the string, we store the length of the first delimiter in the + * StringInfo's cursor field, which we don't otherwise need here. + */ + if (state == NULL) + { + MemoryContext aggcontext; + MemoryContext oldcontext; + + if (!AggCheckCallContext(fcinfo, &aggcontext)) + { + /* cannot be called directly because of internal-type argument */ + elog(ERROR, "bytea_string_agg_transfn called in non-aggregate context"); + } + + /* + * Create state in aggregate context. It'll stay there across + * subsequent calls. + */ + oldcontext = MemoryContextSwitchTo(aggcontext); + state = makeStringInfo(); + MemoryContextSwitchTo(oldcontext); + + isfirst = true; + } + + if (!PG_ARGISNULL(2)) + { + bytea *delim = PG_GETARG_BYTEA_PP(2); + + appendBinaryStringInfo(state, VARDATA_ANY(delim), + VARSIZE_ANY_EXHDR(delim)); + if (isfirst) + state->cursor = VARSIZE_ANY_EXHDR(delim); + } + + appendBinaryStringInfo(state, VARDATA_ANY(value), + VARSIZE_ANY_EXHDR(value)); + } + + /* + * The transition type for string_agg() is declared to be "internal", + * which is a pass-by-value type the same size as a pointer. + */ + if (state) + PG_RETURN_POINTER(state); + PG_RETURN_NULL(); +} + +Datum +bytea_string_agg_finalfn(PG_FUNCTION_ARGS) +{ + StringInfo state; + + /* cannot be called directly because of internal-type argument */ + Assert(AggCheckCallContext(fcinfo, NULL)); + + state = PG_ARGISNULL(0) ? NULL : (StringInfo) PG_GETARG_POINTER(0); + + if (state != NULL) + { + /* As per comment in transfn, strip data before the cursor position */ + bytea *result; + int strippedlen = state->len - state->cursor; + + result = (bytea *) palloc(strippedlen + VARHDRSZ); + SET_VARSIZE(result, strippedlen + VARHDRSZ); + memcpy(VARDATA(result), &state->data[state->cursor], strippedlen); + PG_RETURN_BYTEA_P(result); + } + else + PG_RETURN_NULL(); +} + +/*------------------------------------------------------------- + * byteaoctetlen + * + * get the number of bytes contained in an instance of type 'bytea' + *------------------------------------------------------------- + */ +Datum +byteaoctetlen(PG_FUNCTION_ARGS) +{ + Datum str = PG_GETARG_DATUM(0); + + /* We need not detoast the input at all */ + PG_RETURN_INT32(toast_raw_datum_size(str) - VARHDRSZ); +} + +/* + * byteacat - + * takes two bytea* and returns a bytea* that is the concatenation of + * the two. + * + * Cloned from textcat and modified as required. + */ +Datum +byteacat(PG_FUNCTION_ARGS) +{ + bytea *t1 = PG_GETARG_BYTEA_PP(0); + bytea *t2 = PG_GETARG_BYTEA_PP(1); + + PG_RETURN_BYTEA_P(bytea_catenate(t1, t2)); +} + +/* + * byteaoverlay + * Replace specified substring of first string with second + * + * The SQL standard defines OVERLAY() in terms of substring and concatenation. + * This code is a direct implementation of what the standard says. + */ +Datum +byteaoverlay(PG_FUNCTION_ARGS) +{ + bytea *t1 = PG_GETARG_BYTEA_PP(0); + bytea *t2 = PG_GETARG_BYTEA_PP(1); + int sp = PG_GETARG_INT32(2); /* substring start position */ + int sl = PG_GETARG_INT32(3); /* substring length */ + + PG_RETURN_BYTEA_P(bytea_overlay(t1, t2, sp, sl)); +} + +Datum +byteaoverlay_no_len(PG_FUNCTION_ARGS) +{ + bytea *t1 = PG_GETARG_BYTEA_PP(0); + bytea *t2 = PG_GETARG_BYTEA_PP(1); + int sp = PG_GETARG_INT32(2); /* substring start position */ + int sl; + + sl = VARSIZE_ANY_EXHDR(t2); /* defaults to length(t2) */ + PG_RETURN_BYTEA_P(bytea_overlay(t1, t2, sp, sl)); +} + +/* + * bytea_substr() + * Return a substring starting at the specified position. + * Cloned from text_substr and modified as required. + * + * Input: + * - string + * - starting position (is one-based) + * - string length (optional) + * + * If the starting position is zero or less, then return from the start of the string + * adjusting the length to be consistent with the "negative start" per SQL. + * If the length is less than zero, an ERROR is thrown. If no third argument + * (length) is provided, the length to the end of the string is assumed. + */ +Datum +bytea_substr(PG_FUNCTION_ARGS) +{ + PG_RETURN_BYTEA_P(bytea_substring(PG_GETARG_DATUM(0), + PG_GETARG_INT32(1), + PG_GETARG_INT32(2), + false)); +} + +/* + * bytea_substr_no_len - + * Wrapper to avoid opr_sanity failure due to + * one function accepting a different number of args. + */ +Datum +bytea_substr_no_len(PG_FUNCTION_ARGS) +{ + PG_RETURN_BYTEA_P(bytea_substring(PG_GETARG_DATUM(0), + PG_GETARG_INT32(1), + -1, + true)); +} + +/* + * bit_count + */ +Datum +bytea_bit_count(PG_FUNCTION_ARGS) +{ + bytea *t1 = PG_GETARG_BYTEA_PP(0); + + PG_RETURN_INT64(pg_popcount(VARDATA_ANY(t1), VARSIZE_ANY_EXHDR(t1))); +} + +/* + * byteapos - + * Return the position of the specified substring. + * Implements the SQL POSITION() function. + * Cloned from textpos and modified as required. + */ +Datum +byteapos(PG_FUNCTION_ARGS) +{ + bytea *t1 = PG_GETARG_BYTEA_PP(0); + bytea *t2 = PG_GETARG_BYTEA_PP(1); + int pos; + int px, + p; + int len1, + len2; + char *p1, + *p2; + + len1 = VARSIZE_ANY_EXHDR(t1); + len2 = VARSIZE_ANY_EXHDR(t2); + + if (len2 <= 0) + PG_RETURN_INT32(1); /* result for empty pattern */ + + p1 = VARDATA_ANY(t1); + p2 = VARDATA_ANY(t2); + + pos = 0; + px = (len1 - len2); + for (p = 0; p <= px; p++) + { + if ((*p2 == *p1) && (memcmp(p1, p2, len2) == 0)) + { + pos = p + 1; + break; + }; + p1++; + }; + + PG_RETURN_INT32(pos); +} + +/*------------------------------------------------------------- + * byteaGetByte + * + * this routine treats "bytea" as an array of bytes. + * It returns the Nth byte (a number between 0 and 255). + *------------------------------------------------------------- + */ +Datum +byteaGetByte(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + int32 n = PG_GETARG_INT32(1); + int len; + int byte; + + len = VARSIZE_ANY_EXHDR(v); + + if (n < 0 || n >= len) + ereport(ERROR, + (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), + errmsg("index %d out of valid range, 0..%d", + n, len - 1))); + + byte = ((unsigned char *) VARDATA_ANY(v))[n]; + + PG_RETURN_INT32(byte); +} + +/*------------------------------------------------------------- + * byteaGetBit + * + * This routine treats a "bytea" type like an array of bits. + * It returns the value of the Nth bit (0 or 1). + * + *------------------------------------------------------------- + */ +Datum +byteaGetBit(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + int64 n = PG_GETARG_INT64(1); + int byteNo, + bitNo; + int len; + int byte; + + len = VARSIZE_ANY_EXHDR(v); + + if (n < 0 || n >= (int64) len * 8) + ereport(ERROR, + (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), + errmsg("index %" PRId64 " out of valid range, 0..%" PRId64, + n, (int64) len * 8 - 1))); + + /* n/8 is now known < len, so safe to cast to int */ + byteNo = (int) (n / 8); + bitNo = (int) (n % 8); + + byte = ((unsigned char *) VARDATA_ANY(v))[byteNo]; + + if (byte & (1 << bitNo)) + PG_RETURN_INT32(1); + else + PG_RETURN_INT32(0); +} + +/*------------------------------------------------------------- + * byteaSetByte + * + * Given an instance of type 'bytea' creates a new one with + * the Nth byte set to the given value. + * + *------------------------------------------------------------- + */ +Datum +byteaSetByte(PG_FUNCTION_ARGS) +{ + bytea *res = PG_GETARG_BYTEA_P_COPY(0); + int32 n = PG_GETARG_INT32(1); + int32 newByte = PG_GETARG_INT32(2); + int len; + + len = VARSIZE(res) - VARHDRSZ; + + if (n < 0 || n >= len) + ereport(ERROR, + (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), + errmsg("index %d out of valid range, 0..%d", + n, len - 1))); + + /* + * Now set the byte. + */ + ((unsigned char *) VARDATA(res))[n] = newByte; + + PG_RETURN_BYTEA_P(res); +} + +/*------------------------------------------------------------- + * byteaSetBit + * + * Given an instance of type 'bytea' creates a new one with + * the Nth bit set to the given value. + * + *------------------------------------------------------------- + */ +Datum +byteaSetBit(PG_FUNCTION_ARGS) +{ + bytea *res = PG_GETARG_BYTEA_P_COPY(0); + int64 n = PG_GETARG_INT64(1); + int32 newBit = PG_GETARG_INT32(2); + int len; + int oldByte, + newByte; + int byteNo, + bitNo; + + len = VARSIZE(res) - VARHDRSZ; + + if (n < 0 || n >= (int64) len * 8) + ereport(ERROR, + (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), + errmsg("index %" PRId64 " out of valid range, 0..%" PRId64, + n, (int64) len * 8 - 1))); + + /* n/8 is now known < len, so safe to cast to int */ + byteNo = (int) (n / 8); + bitNo = (int) (n % 8); + + /* + * sanity check! + */ + if (newBit != 0 && newBit != 1) + ereport(ERROR, + (errcode(ERRCODE_INVALID_PARAMETER_VALUE), + errmsg("new bit must be 0 or 1"))); + + /* + * Update the byte. + */ + oldByte = ((unsigned char *) VARDATA(res))[byteNo]; + + if (newBit == 0) + newByte = oldByte & (~(1 << bitNo)); + else + newByte = oldByte | (1 << bitNo); + + ((unsigned char *) VARDATA(res))[byteNo] = newByte; + + PG_RETURN_BYTEA_P(res); +} + +/* + * Return reversed bytea + */ +Datum +bytea_reverse(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + const char *p = VARDATA_ANY(v); + int len = VARSIZE_ANY_EXHDR(v); + const char *endp = p + len; + bytea *result = palloc(len + VARHDRSZ); + char *dst = (char *) VARDATA(result) + len; + + SET_VARSIZE(result, len + VARHDRSZ); + + while (p < endp) + *(--dst) = *p++; + + PG_RETURN_BYTEA_P(result); +} + + +/***************************************************************************** + * Comparison Functions used for bytea + * + * Note: btree indexes need these routines not to leak memory; therefore, + * be careful to free working copies of toasted datums. Most places don't + * need to be so careful. + *****************************************************************************/ + +Datum +byteaeq(PG_FUNCTION_ARGS) +{ + Datum arg1 = PG_GETARG_DATUM(0); + Datum arg2 = PG_GETARG_DATUM(1); + bool result; + Size len1, + len2; + + /* + * We can use a fast path for unequal lengths, which might save us from + * having to detoast one or both values. + */ + len1 = toast_raw_datum_size(arg1); + len2 = toast_raw_datum_size(arg2); + if (len1 != len2) + result = false; + else + { + bytea *barg1 = DatumGetByteaPP(arg1); + bytea *barg2 = DatumGetByteaPP(arg2); + + result = (memcmp(VARDATA_ANY(barg1), VARDATA_ANY(barg2), + len1 - VARHDRSZ) == 0); + + PG_FREE_IF_COPY(barg1, 0); + PG_FREE_IF_COPY(barg2, 1); + } + + PG_RETURN_BOOL(result); +} + +Datum +byteane(PG_FUNCTION_ARGS) +{ + Datum arg1 = PG_GETARG_DATUM(0); + Datum arg2 = PG_GETARG_DATUM(1); + bool result; + Size len1, + len2; + + /* + * We can use a fast path for unequal lengths, which might save us from + * having to detoast one or both values. + */ + len1 = toast_raw_datum_size(arg1); + len2 = toast_raw_datum_size(arg2); + if (len1 != len2) + result = true; + else + { + bytea *barg1 = DatumGetByteaPP(arg1); + bytea *barg2 = DatumGetByteaPP(arg2); + + result = (memcmp(VARDATA_ANY(barg1), VARDATA_ANY(barg2), + len1 - VARHDRSZ) != 0); + + PG_FREE_IF_COPY(barg1, 0); + PG_FREE_IF_COPY(barg2, 1); + } + + PG_RETURN_BOOL(result); +} + +Datum +bytealt(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + + PG_FREE_IF_COPY(arg1, 0); + PG_FREE_IF_COPY(arg2, 1); + + PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 < len2))); +} + +Datum +byteale(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + + PG_FREE_IF_COPY(arg1, 0); + PG_FREE_IF_COPY(arg2, 1); + + PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 <= len2))); +} + +Datum +byteagt(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + + PG_FREE_IF_COPY(arg1, 0); + PG_FREE_IF_COPY(arg2, 1); + + PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 > len2))); +} + +Datum +byteage(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + + PG_FREE_IF_COPY(arg1, 0); + PG_FREE_IF_COPY(arg2, 1); + + PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 >= len2))); +} + +Datum +byteacmp(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + if ((cmp == 0) && (len1 != len2)) + cmp = (len1 < len2) ? -1 : 1; + + PG_FREE_IF_COPY(arg1, 0); + PG_FREE_IF_COPY(arg2, 1); + + PG_RETURN_INT32(cmp); +} + +Datum +bytea_larger(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + bytea *result; + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + result = ((cmp > 0) || ((cmp == 0) && (len1 > len2)) ? arg1 : arg2); + + PG_RETURN_BYTEA_P(result); +} + +Datum +bytea_smaller(PG_FUNCTION_ARGS) +{ + bytea *arg1 = PG_GETARG_BYTEA_PP(0); + bytea *arg2 = PG_GETARG_BYTEA_PP(1); + bytea *result; + int len1, + len2; + int cmp; + + len1 = VARSIZE_ANY_EXHDR(arg1); + len2 = VARSIZE_ANY_EXHDR(arg2); + + cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); + result = ((cmp < 0) || ((cmp == 0) && (len1 < len2)) ? arg1 : arg2); + + PG_RETURN_BYTEA_P(result); +} + +Datum +bytea_sortsupport(PG_FUNCTION_ARGS) +{ + SortSupport ssup = (SortSupport) PG_GETARG_POINTER(0); + MemoryContext oldcontext; + + oldcontext = MemoryContextSwitchTo(ssup->ssup_cxt); + + /* Use generic string SortSupport, forcing "C" collation */ + varstr_sortsupport(ssup, BYTEAOID, C_COLLATION_OID); + + MemoryContextSwitchTo(oldcontext); + + PG_RETURN_VOID(); +} + +/* Cast bytea -> int2 */ +Datum +bytea_int2(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + int len = VARSIZE_ANY_EXHDR(v); + uint16 result; + + /* Check that the byte array is not too long */ + if (len > sizeof(result)) + ereport(ERROR, + errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), + errmsg("smallint out of range")); + + /* Convert it to an integer; most significant bytes come first */ + result = 0; + for (int i = 0; i < len; i++) + { + result <<= BITS_PER_BYTE; + result |= ((unsigned char *) VARDATA_ANY(v))[i]; + } + + PG_RETURN_INT16(result); +} + +/* Cast bytea -> int4 */ +Datum +bytea_int4(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + int len = VARSIZE_ANY_EXHDR(v); + uint32 result; + + /* Check that the byte array is not too long */ + if (len > sizeof(result)) + ereport(ERROR, + errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), + errmsg("integer out of range")); + + /* Convert it to an integer; most significant bytes come first */ + result = 0; + for (int i = 0; i < len; i++) + { + result <<= BITS_PER_BYTE; + result |= ((unsigned char *) VARDATA_ANY(v))[i]; + } + + PG_RETURN_INT32(result); +} + +/* Cast bytea -> int8 */ +Datum +bytea_int8(PG_FUNCTION_ARGS) +{ + bytea *v = PG_GETARG_BYTEA_PP(0); + int len = VARSIZE_ANY_EXHDR(v); + uint64 result; + + /* Check that the byte array is not too long */ + if (len > sizeof(result)) + ereport(ERROR, + errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), + errmsg("bigint out of range")); + + /* Convert it to an integer; most significant bytes come first */ + result = 0; + for (int i = 0; i < len; i++) + { + result <<= BITS_PER_BYTE; + result |= ((unsigned char *) VARDATA_ANY(v))[i]; + } + + PG_RETURN_INT64(result); +} + +/* Cast int2 -> bytea; can just use int2send() */ +Datum +int2_bytea(PG_FUNCTION_ARGS) +{ + return int2send(fcinfo); +} + +/* Cast int4 -> bytea; can just use int4send() */ +Datum +int4_bytea(PG_FUNCTION_ARGS) +{ + return int4send(fcinfo); +} + +/* Cast int8 -> bytea; can just use int8send() */ +Datum +int8_bytea(PG_FUNCTION_ARGS) +{ + return int8send(fcinfo); +} diff --git a/src/backend/utils/adt/meson.build b/src/backend/utils/adt/meson.build index 244f48f4fd7..ed9bbd7b926 100644 --- a/src/backend/utils/adt/meson.build +++ b/src/backend/utils/adt/meson.build @@ -12,6 +12,7 @@ backend_sources += files( 'arrayutils.c', 'ascii.c', 'bool.c', + 'bytea.c', 'cash.c', 'char.c', 'cryptohashfuncs.c', diff --git a/src/backend/utils/adt/varlena.c b/src/backend/utils/adt/varlena.c index 3e4d5568bde..ffae8c23abf 100644 --- a/src/backend/utils/adt/varlena.c +++ b/src/backend/utils/adt/varlena.c @@ -35,7 +35,6 @@ #include "port/pg_bswap.h" #include "regex/regex.h" #include "utils/builtins.h" -#include "utils/bytea.h" #include "utils/guc.h" #include "utils/lsyscache.h" #include "utils/memutils.h" @@ -43,10 +42,6 @@ #include "utils/sortsupport.h" #include "utils/varlena.h" - -/* GUC variable */ -int bytea_output = BYTEA_OUTPUT_HEX; - typedef struct varlena VarString; /* @@ -148,12 +143,6 @@ static int text_position_get_match_pos(TextPositionState *state); static void text_position_cleanup(TextPositionState *state); static void check_collation_set(Oid collid); static int text_cmp(text *arg1, text *arg2, Oid collid); -static bytea *bytea_catenate(bytea *t1, bytea *t2); -static bytea *bytea_substring(Datum str, - int S, - int L, - bool length_not_specified); -static bytea *bytea_overlay(bytea *t1, bytea *t2, int sp, int sl); static void appendStringInfoText(StringInfo str, const text *t); static bool split_text(FunctionCallInfo fcinfo, SplitTextOutputData *tstate); static void split_text_accum_result(SplitTextOutputData *tstate, @@ -279,307 +268,6 @@ text_to_cstring_buffer(const text *src, char *dst, size_t dst_len) * USER I/O ROUTINES * *****************************************************************************/ - -#define VAL(CH) ((CH) - '0') -#define DIG(VAL) ((VAL) + '0') - -/* - * byteain - converts from printable representation of byte array - * - * Non-printable characters must be passed as '\nnn' (octal) and are - * converted to internal form. '\' must be passed as '\\'. - * ereport(ERROR, ...) if bad form. - * - * BUGS: - * The input is scanned twice. - * The error checking of input is minimal. - */ -Datum -byteain(PG_FUNCTION_ARGS) -{ - char *inputText = PG_GETARG_CSTRING(0); - Node *escontext = fcinfo->context; - char *tp; - char *rp; - int bc; - bytea *result; - - /* Recognize hex input */ - if (inputText[0] == '\\' && inputText[1] == 'x') - { - size_t len = strlen(inputText); - - bc = (len - 2) / 2 + VARHDRSZ; /* maximum possible length */ - result = palloc(bc); - bc = hex_decode_safe(inputText + 2, len - 2, VARDATA(result), - escontext); - SET_VARSIZE(result, bc + VARHDRSZ); /* actual length */ - - PG_RETURN_BYTEA_P(result); - } - - /* Else, it's the traditional escaped style */ - for (bc = 0, tp = inputText; *tp != '\0'; bc++) - { - if (tp[0] != '\\') - tp++; - else if ((tp[0] == '\\') && - (tp[1] >= '0' && tp[1] <= '3') && - (tp[2] >= '0' && tp[2] <= '7') && - (tp[3] >= '0' && tp[3] <= '7')) - tp += 4; - else if ((tp[0] == '\\') && - (tp[1] == '\\')) - tp += 2; - else - { - /* - * one backslash, not followed by another or ### valid octal - */ - ereturn(escontext, (Datum) 0, - (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION), - errmsg("invalid input syntax for type %s", "bytea"))); - } - } - - bc += VARHDRSZ; - - result = (bytea *) palloc(bc); - SET_VARSIZE(result, bc); - - tp = inputText; - rp = VARDATA(result); - while (*tp != '\0') - { - if (tp[0] != '\\') - *rp++ = *tp++; - else if ((tp[0] == '\\') && - (tp[1] >= '0' && tp[1] <= '3') && - (tp[2] >= '0' && tp[2] <= '7') && - (tp[3] >= '0' && tp[3] <= '7')) - { - bc = VAL(tp[1]); - bc <<= 3; - bc += VAL(tp[2]); - bc <<= 3; - *rp++ = bc + VAL(tp[3]); - - tp += 4; - } - else if ((tp[0] == '\\') && - (tp[1] == '\\')) - { - *rp++ = '\\'; - tp += 2; - } - else - { - /* - * We should never get here. The first pass should not allow it. - */ - ereturn(escontext, (Datum) 0, - (errcode(ERRCODE_INVALID_TEXT_REPRESENTATION), - errmsg("invalid input syntax for type %s", "bytea"))); - } - } - - PG_RETURN_BYTEA_P(result); -} - -/* - * byteaout - converts to printable representation of byte array - * - * In the traditional escaped format, non-printable characters are - * printed as '\nnn' (octal) and '\' as '\\'. - */ -Datum -byteaout(PG_FUNCTION_ARGS) -{ - bytea *vlena = PG_GETARG_BYTEA_PP(0); - char *result; - char *rp; - - if (bytea_output == BYTEA_OUTPUT_HEX) - { - /* Print hex format */ - rp = result = palloc(VARSIZE_ANY_EXHDR(vlena) * 2 + 2 + 1); - *rp++ = '\\'; - *rp++ = 'x'; - rp += hex_encode(VARDATA_ANY(vlena), VARSIZE_ANY_EXHDR(vlena), rp); - } - else if (bytea_output == BYTEA_OUTPUT_ESCAPE) - { - /* Print traditional escaped format */ - char *vp; - uint64 len; - int i; - - len = 1; /* empty string has 1 char */ - vp = VARDATA_ANY(vlena); - for (i = VARSIZE_ANY_EXHDR(vlena); i != 0; i--, vp++) - { - if (*vp == '\\') - len += 2; - else if ((unsigned char) *vp < 0x20 || (unsigned char) *vp > 0x7e) - len += 4; - else - len++; - } - - /* - * In principle len can't overflow uint32 if the input fit in 1GB, but - * for safety let's check rather than relying on palloc's internal - * check. - */ - if (len > MaxAllocSize) - ereport(ERROR, - (errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED), - errmsg_internal("result of bytea output conversion is too large"))); - rp = result = (char *) palloc(len); - - vp = VARDATA_ANY(vlena); - for (i = VARSIZE_ANY_EXHDR(vlena); i != 0; i--, vp++) - { - if (*vp == '\\') - { - *rp++ = '\\'; - *rp++ = '\\'; - } - else if ((unsigned char) *vp < 0x20 || (unsigned char) *vp > 0x7e) - { - int val; /* holds unprintable chars */ - - val = *vp; - rp[0] = '\\'; - rp[3] = DIG(val & 07); - val >>= 3; - rp[2] = DIG(val & 07); - val >>= 3; - rp[1] = DIG(val & 03); - rp += 4; - } - else - *rp++ = *vp; - } - } - else - { - elog(ERROR, "unrecognized \"bytea_output\" setting: %d", - bytea_output); - rp = result = NULL; /* keep compiler quiet */ - } - *rp = '\0'; - PG_RETURN_CSTRING(result); -} - -/* - * bytearecv - converts external binary format to bytea - */ -Datum -bytearecv(PG_FUNCTION_ARGS) -{ - StringInfo buf = (StringInfo) PG_GETARG_POINTER(0); - bytea *result; - int nbytes; - - nbytes = buf->len - buf->cursor; - result = (bytea *) palloc(nbytes + VARHDRSZ); - SET_VARSIZE(result, nbytes + VARHDRSZ); - pq_copymsgbytes(buf, VARDATA(result), nbytes); - PG_RETURN_BYTEA_P(result); -} - -/* - * byteasend - converts bytea to binary format - * - * This is a special case: just copy the input... - */ -Datum -byteasend(PG_FUNCTION_ARGS) -{ - bytea *vlena = PG_GETARG_BYTEA_P_COPY(0); - - PG_RETURN_BYTEA_P(vlena); -} - -Datum -bytea_string_agg_transfn(PG_FUNCTION_ARGS) -{ - StringInfo state; - - state = PG_ARGISNULL(0) ? NULL : (StringInfo) PG_GETARG_POINTER(0); - - /* Append the value unless null, preceding it with the delimiter. */ - if (!PG_ARGISNULL(1)) - { - bytea *value = PG_GETARG_BYTEA_PP(1); - bool isfirst = false; - - /* - * You might think we can just throw away the first delimiter, however - * we must keep it as we may be a parallel worker doing partial - * aggregation building a state to send to the main process. We need - * to keep the delimiter of every aggregation so that the combine - * function can properly join up the strings of two separately - * partially aggregated results. The first delimiter is only stripped - * off in the final function. To know how much to strip off the front - * of the string, we store the length of the first delimiter in the - * StringInfo's cursor field, which we don't otherwise need here. - */ - if (state == NULL) - { - state = makeStringAggState(fcinfo); - isfirst = true; - } - - if (!PG_ARGISNULL(2)) - { - bytea *delim = PG_GETARG_BYTEA_PP(2); - - appendBinaryStringInfo(state, VARDATA_ANY(delim), - VARSIZE_ANY_EXHDR(delim)); - if (isfirst) - state->cursor = VARSIZE_ANY_EXHDR(delim); - } - - appendBinaryStringInfo(state, VARDATA_ANY(value), - VARSIZE_ANY_EXHDR(value)); - } - - /* - * The transition type for string_agg() is declared to be "internal", - * which is a pass-by-value type the same size as a pointer. - */ - if (state) - PG_RETURN_POINTER(state); - PG_RETURN_NULL(); -} - -Datum -bytea_string_agg_finalfn(PG_FUNCTION_ARGS) -{ - StringInfo state; - - /* cannot be called directly because of internal-type argument */ - Assert(AggCheckCallContext(fcinfo, NULL)); - - state = PG_ARGISNULL(0) ? NULL : (StringInfo) PG_GETARG_POINTER(0); - - if (state != NULL) - { - /* As per comment in transfn, strip data before the cursor position */ - bytea *result; - int strippedlen = state->len - state->cursor; - - result = (bytea *) palloc(strippedlen + VARHDRSZ); - SET_VARSIZE(result, strippedlen + VARHDRSZ); - memcpy(VARDATA(result), &state->data[state->cursor], strippedlen); - PG_RETURN_BYTEA_P(result); - } - else - PG_RETURN_NULL(); -} - /* * textin - converts cstring to internal representation */ @@ -2959,552 +2647,91 @@ bttext_pattern_sortsupport(PG_FUNCTION_ARGS) } -/*------------------------------------------------------------- - * byteaoctetlen - * - * get the number of bytes contained in an instance of type 'bytea' - *------------------------------------------------------------- +/* text_name() + * Converts a text type to a Name type. */ Datum -byteaoctetlen(PG_FUNCTION_ARGS) +text_name(PG_FUNCTION_ARGS) { - Datum str = PG_GETARG_DATUM(0); + text *s = PG_GETARG_TEXT_PP(0); + Name result; + int len; - /* We need not detoast the input at all */ - PG_RETURN_INT32(toast_raw_datum_size(str) - VARHDRSZ); + len = VARSIZE_ANY_EXHDR(s); + + /* Truncate oversize input */ + if (len >= NAMEDATALEN) + len = pg_mbcliplen(VARDATA_ANY(s), len, NAMEDATALEN - 1); + + /* We use palloc0 here to ensure result is zero-padded */ + result = (Name) palloc0(NAMEDATALEN); + memcpy(NameStr(*result), VARDATA_ANY(s), len); + + PG_RETURN_NAME(result); } -/* - * byteacat - - * takes two bytea* and returns a bytea* that is the concatenation of - * the two. - * - * Cloned from textcat and modified as required. +/* name_text() + * Converts a Name type to a text type. */ Datum -byteacat(PG_FUNCTION_ARGS) +name_text(PG_FUNCTION_ARGS) { - bytea *t1 = PG_GETARG_BYTEA_PP(0); - bytea *t2 = PG_GETARG_BYTEA_PP(1); + Name s = PG_GETARG_NAME(0); - PG_RETURN_BYTEA_P(bytea_catenate(t1, t2)); + PG_RETURN_TEXT_P(cstring_to_text(NameStr(*s))); } + /* - * bytea_catenate - * Guts of byteacat(), broken out so it can be used by other functions + * textToQualifiedNameList - convert a text object to list of names * - * Arguments can be in short-header form, but not compressed or out-of-line + * This implements the input parsing needed by nextval() and other + * functions that take a text parameter representing a qualified name. + * We split the name at dots, downcase if not double-quoted, and + * truncate names if they're too long. */ -static bytea * -bytea_catenate(bytea *t1, bytea *t2) +List * +textToQualifiedNameList(text *textval) { - bytea *result; - int len1, - len2, - len; - char *ptr; + char *rawname; + List *result = NIL; + List *namelist; + ListCell *l; - len1 = VARSIZE_ANY_EXHDR(t1); - len2 = VARSIZE_ANY_EXHDR(t2); + /* Convert to C string (handles possible detoasting). */ + /* Note we rely on being able to modify rawname below. */ + rawname = text_to_cstring(textval); - /* paranoia ... probably should throw error instead? */ - if (len1 < 0) - len1 = 0; - if (len2 < 0) - len2 = 0; + if (!SplitIdentifierString(rawname, '.', &namelist)) + ereport(ERROR, + (errcode(ERRCODE_INVALID_NAME), + errmsg("invalid name syntax"))); - len = len1 + len2 + VARHDRSZ; - result = (bytea *) palloc(len); + if (namelist == NIL) + ereport(ERROR, + (errcode(ERRCODE_INVALID_NAME), + errmsg("invalid name syntax"))); - /* Set size of result string... */ - SET_VARSIZE(result, len); + foreach(l, namelist) + { + char *curname = (char *) lfirst(l); - /* Fill data field of result string... */ - ptr = VARDATA(result); - if (len1 > 0) - memcpy(ptr, VARDATA_ANY(t1), len1); - if (len2 > 0) - memcpy(ptr + len1, VARDATA_ANY(t2), len2); + result = lappend(result, makeString(pstrdup(curname))); + } + + pfree(rawname); + list_free(namelist); return result; } -#define PG_STR_GET_BYTEA(str_) \ - DatumGetByteaPP(DirectFunctionCall1(byteain, CStringGetDatum(str_))) - /* - * bytea_substr() - * Return a substring starting at the specified position. - * Cloned from text_substr and modified as required. + * SplitIdentifierString --- parse a string containing identifiers * - * Input: - * - string - * - starting position (is one-based) - * - string length (optional) - * - * If the starting position is zero or less, then return from the start of the string - * adjusting the length to be consistent with the "negative start" per SQL. - * If the length is less than zero, an ERROR is thrown. If no third argument - * (length) is provided, the length to the end of the string is assumed. - */ -Datum -bytea_substr(PG_FUNCTION_ARGS) -{ - PG_RETURN_BYTEA_P(bytea_substring(PG_GETARG_DATUM(0), - PG_GETARG_INT32(1), - PG_GETARG_INT32(2), - false)); -} - -/* - * bytea_substr_no_len - - * Wrapper to avoid opr_sanity failure due to - * one function accepting a different number of args. - */ -Datum -bytea_substr_no_len(PG_FUNCTION_ARGS) -{ - PG_RETURN_BYTEA_P(bytea_substring(PG_GETARG_DATUM(0), - PG_GETARG_INT32(1), - -1, - true)); -} - -static bytea * -bytea_substring(Datum str, - int S, - int L, - bool length_not_specified) -{ - int32 S1; /* adjusted start position */ - int32 L1; /* adjusted substring length */ - int32 E; /* end position */ - - /* - * The logic here should generally match text_substring(). - */ - S1 = Max(S, 1); - - if (length_not_specified) - { - /* - * Not passed a length - DatumGetByteaPSlice() grabs everything to the - * end of the string if we pass it a negative value for length. - */ - L1 = -1; - } - else if (L < 0) - { - /* SQL99 says to throw an error for E < S, i.e., negative length */ - ereport(ERROR, - (errcode(ERRCODE_SUBSTRING_ERROR), - errmsg("negative substring length not allowed"))); - L1 = -1; /* silence stupider compilers */ - } - else if (pg_add_s32_overflow(S, L, &E)) - { - /* - * L could be large enough for S + L to overflow, in which case the - * substring must run to end of string. - */ - L1 = -1; - } - else - { - /* - * A zero or negative value for the end position can happen if the - * start was negative or one. SQL99 says to return a zero-length - * string. - */ - if (E < 1) - return PG_STR_GET_BYTEA(""); - - L1 = E - S1; - } - - /* - * If the start position is past the end of the string, SQL99 says to - * return a zero-length string -- DatumGetByteaPSlice() will do that for - * us. We need only convert S1 to zero-based starting position. - */ - return DatumGetByteaPSlice(str, S1 - 1, L1); -} - -/* - * byteaoverlay - * Replace specified substring of first string with second - * - * The SQL standard defines OVERLAY() in terms of substring and concatenation. - * This code is a direct implementation of what the standard says. - */ -Datum -byteaoverlay(PG_FUNCTION_ARGS) -{ - bytea *t1 = PG_GETARG_BYTEA_PP(0); - bytea *t2 = PG_GETARG_BYTEA_PP(1); - int sp = PG_GETARG_INT32(2); /* substring start position */ - int sl = PG_GETARG_INT32(3); /* substring length */ - - PG_RETURN_BYTEA_P(bytea_overlay(t1, t2, sp, sl)); -} - -Datum -byteaoverlay_no_len(PG_FUNCTION_ARGS) -{ - bytea *t1 = PG_GETARG_BYTEA_PP(0); - bytea *t2 = PG_GETARG_BYTEA_PP(1); - int sp = PG_GETARG_INT32(2); /* substring start position */ - int sl; - - sl = VARSIZE_ANY_EXHDR(t2); /* defaults to length(t2) */ - PG_RETURN_BYTEA_P(bytea_overlay(t1, t2, sp, sl)); -} - -static bytea * -bytea_overlay(bytea *t1, bytea *t2, int sp, int sl) -{ - bytea *result; - bytea *s1; - bytea *s2; - int sp_pl_sl; - - /* - * Check for possible integer-overflow cases. For negative sp, throw a - * "substring length" error because that's what should be expected - * according to the spec's definition of OVERLAY(). - */ - if (sp <= 0) - ereport(ERROR, - (errcode(ERRCODE_SUBSTRING_ERROR), - errmsg("negative substring length not allowed"))); - if (pg_add_s32_overflow(sp, sl, &sp_pl_sl)) - ereport(ERROR, - (errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), - errmsg("integer out of range"))); - - s1 = bytea_substring(PointerGetDatum(t1), 1, sp - 1, false); - s2 = bytea_substring(PointerGetDatum(t1), sp_pl_sl, -1, true); - result = bytea_catenate(s1, t2); - result = bytea_catenate(result, s2); - - return result; -} - -/* - * bit_count - */ -Datum -bytea_bit_count(PG_FUNCTION_ARGS) -{ - bytea *t1 = PG_GETARG_BYTEA_PP(0); - - PG_RETURN_INT64(pg_popcount(VARDATA_ANY(t1), VARSIZE_ANY_EXHDR(t1))); -} - -/* - * byteapos - - * Return the position of the specified substring. - * Implements the SQL POSITION() function. - * Cloned from textpos and modified as required. - */ -Datum -byteapos(PG_FUNCTION_ARGS) -{ - bytea *t1 = PG_GETARG_BYTEA_PP(0); - bytea *t2 = PG_GETARG_BYTEA_PP(1); - int pos; - int px, - p; - int len1, - len2; - char *p1, - *p2; - - len1 = VARSIZE_ANY_EXHDR(t1); - len2 = VARSIZE_ANY_EXHDR(t2); - - if (len2 <= 0) - PG_RETURN_INT32(1); /* result for empty pattern */ - - p1 = VARDATA_ANY(t1); - p2 = VARDATA_ANY(t2); - - pos = 0; - px = (len1 - len2); - for (p = 0; p <= px; p++) - { - if ((*p2 == *p1) && (memcmp(p1, p2, len2) == 0)) - { - pos = p + 1; - break; - }; - p1++; - }; - - PG_RETURN_INT32(pos); -} - -/*------------------------------------------------------------- - * byteaGetByte - * - * this routine treats "bytea" as an array of bytes. - * It returns the Nth byte (a number between 0 and 255). - *------------------------------------------------------------- - */ -Datum -byteaGetByte(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - int32 n = PG_GETARG_INT32(1); - int len; - int byte; - - len = VARSIZE_ANY_EXHDR(v); - - if (n < 0 || n >= len) - ereport(ERROR, - (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), - errmsg("index %d out of valid range, 0..%d", - n, len - 1))); - - byte = ((unsigned char *) VARDATA_ANY(v))[n]; - - PG_RETURN_INT32(byte); -} - -/*------------------------------------------------------------- - * byteaGetBit - * - * This routine treats a "bytea" type like an array of bits. - * It returns the value of the Nth bit (0 or 1). - * - *------------------------------------------------------------- - */ -Datum -byteaGetBit(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - int64 n = PG_GETARG_INT64(1); - int byteNo, - bitNo; - int len; - int byte; - - len = VARSIZE_ANY_EXHDR(v); - - if (n < 0 || n >= (int64) len * 8) - ereport(ERROR, - (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), - errmsg("index %" PRId64 " out of valid range, 0..%" PRId64, - n, (int64) len * 8 - 1))); - - /* n/8 is now known < len, so safe to cast to int */ - byteNo = (int) (n / 8); - bitNo = (int) (n % 8); - - byte = ((unsigned char *) VARDATA_ANY(v))[byteNo]; - - if (byte & (1 << bitNo)) - PG_RETURN_INT32(1); - else - PG_RETURN_INT32(0); -} - -/*------------------------------------------------------------- - * byteaSetByte - * - * Given an instance of type 'bytea' creates a new one with - * the Nth byte set to the given value. - * - *------------------------------------------------------------- - */ -Datum -byteaSetByte(PG_FUNCTION_ARGS) -{ - bytea *res = PG_GETARG_BYTEA_P_COPY(0); - int32 n = PG_GETARG_INT32(1); - int32 newByte = PG_GETARG_INT32(2); - int len; - - len = VARSIZE(res) - VARHDRSZ; - - if (n < 0 || n >= len) - ereport(ERROR, - (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), - errmsg("index %d out of valid range, 0..%d", - n, len - 1))); - - /* - * Now set the byte. - */ - ((unsigned char *) VARDATA(res))[n] = newByte; - - PG_RETURN_BYTEA_P(res); -} - -/*------------------------------------------------------------- - * byteaSetBit - * - * Given an instance of type 'bytea' creates a new one with - * the Nth bit set to the given value. - * - *------------------------------------------------------------- - */ -Datum -byteaSetBit(PG_FUNCTION_ARGS) -{ - bytea *res = PG_GETARG_BYTEA_P_COPY(0); - int64 n = PG_GETARG_INT64(1); - int32 newBit = PG_GETARG_INT32(2); - int len; - int oldByte, - newByte; - int byteNo, - bitNo; - - len = VARSIZE(res) - VARHDRSZ; - - if (n < 0 || n >= (int64) len * 8) - ereport(ERROR, - (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), - errmsg("index %" PRId64 " out of valid range, 0..%" PRId64, - n, (int64) len * 8 - 1))); - - /* n/8 is now known < len, so safe to cast to int */ - byteNo = (int) (n / 8); - bitNo = (int) (n % 8); - - /* - * sanity check! - */ - if (newBit != 0 && newBit != 1) - ereport(ERROR, - (errcode(ERRCODE_INVALID_PARAMETER_VALUE), - errmsg("new bit must be 0 or 1"))); - - /* - * Update the byte. - */ - oldByte = ((unsigned char *) VARDATA(res))[byteNo]; - - if (newBit == 0) - newByte = oldByte & (~(1 << bitNo)); - else - newByte = oldByte | (1 << bitNo); - - ((unsigned char *) VARDATA(res))[byteNo] = newByte; - - PG_RETURN_BYTEA_P(res); -} - -/* - * Return reversed bytea - */ -Datum -bytea_reverse(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - const char *p = VARDATA_ANY(v); - int len = VARSIZE_ANY_EXHDR(v); - const char *endp = p + len; - bytea *result = palloc(len + VARHDRSZ); - char *dst = (char *) VARDATA(result) + len; - - SET_VARSIZE(result, len + VARHDRSZ); - - while (p < endp) - *(--dst) = *p++; - - PG_RETURN_BYTEA_P(result); -} - - -/* text_name() - * Converts a text type to a Name type. - */ -Datum -text_name(PG_FUNCTION_ARGS) -{ - text *s = PG_GETARG_TEXT_PP(0); - Name result; - int len; - - len = VARSIZE_ANY_EXHDR(s); - - /* Truncate oversize input */ - if (len >= NAMEDATALEN) - len = pg_mbcliplen(VARDATA_ANY(s), len, NAMEDATALEN - 1); - - /* We use palloc0 here to ensure result is zero-padded */ - result = (Name) palloc0(NAMEDATALEN); - memcpy(NameStr(*result), VARDATA_ANY(s), len); - - PG_RETURN_NAME(result); -} - -/* name_text() - * Converts a Name type to a text type. - */ -Datum -name_text(PG_FUNCTION_ARGS) -{ - Name s = PG_GETARG_NAME(0); - - PG_RETURN_TEXT_P(cstring_to_text(NameStr(*s))); -} - - -/* - * textToQualifiedNameList - convert a text object to list of names - * - * This implements the input parsing needed by nextval() and other - * functions that take a text parameter representing a qualified name. - * We split the name at dots, downcase if not double-quoted, and - * truncate names if they're too long. - */ -List * -textToQualifiedNameList(text *textval) -{ - char *rawname; - List *result = NIL; - List *namelist; - ListCell *l; - - /* Convert to C string (handles possible detoasting). */ - /* Note we rely on being able to modify rawname below. */ - rawname = text_to_cstring(textval); - - if (!SplitIdentifierString(rawname, '.', &namelist)) - ereport(ERROR, - (errcode(ERRCODE_INVALID_NAME), - errmsg("invalid name syntax"))); - - if (namelist == NIL) - ereport(ERROR, - (errcode(ERRCODE_INVALID_NAME), - errmsg("invalid name syntax"))); - - foreach(l, namelist) - { - char *curname = (char *) lfirst(l); - - result = lappend(result, makeString(pstrdup(curname))); - } - - pfree(rawname); - list_free(namelist); - - return result; -} - -/* - * SplitIdentifierString --- parse a string containing identifiers - * - * This is the guts of textToQualifiedNameList, and is exported for use in - * other situations such as parsing GUC variables. In the GUC case, it's - * important to avoid memory leaks, so the API is designed to minimize the - * amount of stuff that needs to be allocated and freed. + * This is the guts of textToQualifiedNameList, and is exported for use in + * other situations such as parsing GUC variables. In the GUC case, it's + * important to avoid memory leaks, so the API is designed to minimize the + * amount of stuff that needs to be allocated and freed. * * Inputs: * rawstring: the input string; must be overwritable! On return, it's @@ -3849,331 +3076,6 @@ SplitGUCList(char *rawstring, char separator, return true; } - -/***************************************************************************** - * Comparison Functions used for bytea - * - * Note: btree indexes need these routines not to leak memory; therefore, - * be careful to free working copies of toasted datums. Most places don't - * need to be so careful. - *****************************************************************************/ - -Datum -byteaeq(PG_FUNCTION_ARGS) -{ - Datum arg1 = PG_GETARG_DATUM(0); - Datum arg2 = PG_GETARG_DATUM(1); - bool result; - Size len1, - len2; - - /* - * We can use a fast path for unequal lengths, which might save us from - * having to detoast one or both values. - */ - len1 = toast_raw_datum_size(arg1); - len2 = toast_raw_datum_size(arg2); - if (len1 != len2) - result = false; - else - { - bytea *barg1 = DatumGetByteaPP(arg1); - bytea *barg2 = DatumGetByteaPP(arg2); - - result = (memcmp(VARDATA_ANY(barg1), VARDATA_ANY(barg2), - len1 - VARHDRSZ) == 0); - - PG_FREE_IF_COPY(barg1, 0); - PG_FREE_IF_COPY(barg2, 1); - } - - PG_RETURN_BOOL(result); -} - -Datum -byteane(PG_FUNCTION_ARGS) -{ - Datum arg1 = PG_GETARG_DATUM(0); - Datum arg2 = PG_GETARG_DATUM(1); - bool result; - Size len1, - len2; - - /* - * We can use a fast path for unequal lengths, which might save us from - * having to detoast one or both values. - */ - len1 = toast_raw_datum_size(arg1); - len2 = toast_raw_datum_size(arg2); - if (len1 != len2) - result = true; - else - { - bytea *barg1 = DatumGetByteaPP(arg1); - bytea *barg2 = DatumGetByteaPP(arg2); - - result = (memcmp(VARDATA_ANY(barg1), VARDATA_ANY(barg2), - len1 - VARHDRSZ) != 0); - - PG_FREE_IF_COPY(barg1, 0); - PG_FREE_IF_COPY(barg2, 1); - } - - PG_RETURN_BOOL(result); -} - -Datum -bytealt(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - - PG_FREE_IF_COPY(arg1, 0); - PG_FREE_IF_COPY(arg2, 1); - - PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 < len2))); -} - -Datum -byteale(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - - PG_FREE_IF_COPY(arg1, 0); - PG_FREE_IF_COPY(arg2, 1); - - PG_RETURN_BOOL((cmp < 0) || ((cmp == 0) && (len1 <= len2))); -} - -Datum -byteagt(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - - PG_FREE_IF_COPY(arg1, 0); - PG_FREE_IF_COPY(arg2, 1); - - PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 > len2))); -} - -Datum -byteage(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - - PG_FREE_IF_COPY(arg1, 0); - PG_FREE_IF_COPY(arg2, 1); - - PG_RETURN_BOOL((cmp > 0) || ((cmp == 0) && (len1 >= len2))); -} - -Datum -byteacmp(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - if ((cmp == 0) && (len1 != len2)) - cmp = (len1 < len2) ? -1 : 1; - - PG_FREE_IF_COPY(arg1, 0); - PG_FREE_IF_COPY(arg2, 1); - - PG_RETURN_INT32(cmp); -} - -Datum -bytea_larger(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - bytea *result; - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - result = ((cmp > 0) || ((cmp == 0) && (len1 > len2)) ? arg1 : arg2); - - PG_RETURN_BYTEA_P(result); -} - -Datum -bytea_smaller(PG_FUNCTION_ARGS) -{ - bytea *arg1 = PG_GETARG_BYTEA_PP(0); - bytea *arg2 = PG_GETARG_BYTEA_PP(1); - bytea *result; - int len1, - len2; - int cmp; - - len1 = VARSIZE_ANY_EXHDR(arg1); - len2 = VARSIZE_ANY_EXHDR(arg2); - - cmp = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2)); - result = ((cmp < 0) || ((cmp == 0) && (len1 < len2)) ? arg1 : arg2); - - PG_RETURN_BYTEA_P(result); -} - -Datum -bytea_sortsupport(PG_FUNCTION_ARGS) -{ - SortSupport ssup = (SortSupport) PG_GETARG_POINTER(0); - MemoryContext oldcontext; - - oldcontext = MemoryContextSwitchTo(ssup->ssup_cxt); - - /* Use generic string SortSupport, forcing "C" collation */ - varstr_sortsupport(ssup, BYTEAOID, C_COLLATION_OID); - - MemoryContextSwitchTo(oldcontext); - - PG_RETURN_VOID(); -} - -/* Cast bytea -> int2 */ -Datum -bytea_int2(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - int len = VARSIZE_ANY_EXHDR(v); - uint16 result; - - /* Check that the byte array is not too long */ - if (len > sizeof(result)) - ereport(ERROR, - errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), - errmsg("smallint out of range")); - - /* Convert it to an integer; most significant bytes come first */ - result = 0; - for (int i = 0; i < len; i++) - { - result <<= BITS_PER_BYTE; - result |= ((unsigned char *) VARDATA_ANY(v))[i]; - } - - PG_RETURN_INT16(result); -} - -/* Cast bytea -> int4 */ -Datum -bytea_int4(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - int len = VARSIZE_ANY_EXHDR(v); - uint32 result; - - /* Check that the byte array is not too long */ - if (len > sizeof(result)) - ereport(ERROR, - errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), - errmsg("integer out of range")); - - /* Convert it to an integer; most significant bytes come first */ - result = 0; - for (int i = 0; i < len; i++) - { - result <<= BITS_PER_BYTE; - result |= ((unsigned char *) VARDATA_ANY(v))[i]; - } - - PG_RETURN_INT32(result); -} - -/* Cast bytea -> int8 */ -Datum -bytea_int8(PG_FUNCTION_ARGS) -{ - bytea *v = PG_GETARG_BYTEA_PP(0); - int len = VARSIZE_ANY_EXHDR(v); - uint64 result; - - /* Check that the byte array is not too long */ - if (len > sizeof(result)) - ereport(ERROR, - errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), - errmsg("bigint out of range")); - - /* Convert it to an integer; most significant bytes come first */ - result = 0; - for (int i = 0; i < len; i++) - { - result <<= BITS_PER_BYTE; - result |= ((unsigned char *) VARDATA_ANY(v))[i]; - } - - PG_RETURN_INT64(result); -} - -/* Cast int2 -> bytea; can just use int2send() */ -Datum -int2_bytea(PG_FUNCTION_ARGS) -{ - return int2send(fcinfo); -} - -/* Cast int4 -> bytea; can just use int4send() */ -Datum -int4_bytea(PG_FUNCTION_ARGS) -{ - return int4send(fcinfo); -} - -/* Cast int8 -> bytea; can just use int8send() */ -Datum -int8_bytea(PG_FUNCTION_ARGS) -{ - return int8send(fcinfo); -} - /* * appendStringInfoText * -- 2.39.5