PostgreSQL中的User subroutines有什么作用

本篇内容介绍了“postgresql中的User subroutines有什么作用”的有关知识，在实际案例的操作过程中，不少人都会遇到这样的困境，接下来就让小编带领大家学习一下如何处理这些情况吧！希望大家仔细阅读，能够学有所成！

Flex输入文件由四部分组成:

%{
Declarations
%}
Definitions
%%
Rules
%%
User subroutines

User subroutines

在规则之后是自定义例程,在scan.l中定义的例程主要是对输入的sql语句进行解析以及执行初始化和事后清理工作等.

#undef yyextra
#define yyextra  (((struct yyguts_t *) yyscanner)->yyextra_r)

//定义其他需要的东西:yylloc/yyleng
#undef yylloc
#define yylloc    (((struct yyguts_t *) yyscanner)->yylloc_r)
#undef yyleng
#define yyleng    (((struct yyguts_t *) yyscanner)->yyleng_r)

int
scanner_errposition(int location, core_yyscan_t yyscanner)
{
    int            pos;
    if (location < 0)
        return 0;                
    
    pos = pg_mbstrlen_with_len(yyextra->scanbuf, location) + 1;
    
    return errposition(pos);
}

void
scanner_yyerror(const char *message, core_yyscan_t yyscanner)
{
    const char *loc = yyextra->scanbuf + *yylloc;
    if (*loc == YY_END_OF_BUFFER_CHAR)
    {
        ereport(ERROR,
                (errcode(ERRCODE_SYNTAX_ERROR),
        
                 errmsg("%s at end of input", _(message)),
                 lexer_errposition()));
    }
    else
    {
        ereport(ERROR,
                (errcode(ERRCODE_SYNTAX_ERROR),
        
                 errmsg("%s at or near \"%s\"", _(message), loc),
                 lexer_errposition()));
    }
}

core_yyscan_t
scanner_init(const char *str,
             core_yy_extra_type *yyext,
             const ScanKeyWord *keywords,
             int num_keywords)
{
    Size        slen = strlen(str);
    yyscan_t    scanner;
    if (yylex_init(&scanner) != 0)
        elog(ERROR, "yylex_init() failed: %m");
    core_yyset_extra(yyext, scanner);
    yyext->keywords = keywords;
    yyext->num_keywords = num_keywords;
    yyext->backslash_quote = backslash_quote;
    yyext->escape_string_warning = escape_string_warning;
    yyext->standard_confORMing_strings = standard_conforming_strings;
    
    yyext->scanbuf = (char *) palloc(slen + 2);
    yyext->scanbuflen = slen;
    memcpy(yyext->scanbuf, str, slen);
    yyext->scanbuf[slen] = yyext->scanbuf[slen + 1] = YY_END_OF_BUFFER_CHAR;
    yy_scan_buffer(yyext->scanbuf, slen + 2, scanner);
    
    yyext->literalalloc = 1024;
    yyext->literalbuf = (char *) palloc(yyext->literalalloc);
    yyext->literallen = 0;
    return scanner;
}

void
scanner_finish(core_yyscan_t yyscanner)
{
    
    if (yyextra->scanbuflen >= 8192)
        pfree(yyextra->scanbuf);
    if (yyextra->literalalloc >= 8192)
        pfree(yyextra->literalbuf);
}
static void
addlit(char *ytext, int yleng, core_yyscan_t yyscanner)
{
    
    //增大缓存
    if ((yyextra->literallen + yleng) >= yyextra->literalalloc)
    {
        do
        {
            yyextra->literalalloc *= 2;
        } while ((yyextra->literallen + yleng) >= yyextra->literalalloc);
        yyextra->literalbuf = (char *) repalloc(yyextra->literalbuf,
                                                yyextra->literalalloc);
    }
    
    //追加新数据
    memcpy(yyextra->literalbuf + yyextra->literallen, ytext, yleng);
    yyextra->literallen += yleng;
}
static void
addlitchar(unsigned char ychar, core_yyscan_t yyscanner)
{
    
    if ((yyextra->literallen + 1) >= yyextra->literalalloc)
    {
        yyextra->literalalloc *= 2;
        yyextra->literalbuf = (char *) repalloc(yyextra->literalbuf,
                                                yyextra->literalalloc);
    }
    
    yyextra->literalbuf[yyextra->literallen] = ychar;
    yyextra->literallen += 1;
}

static char *
litbufdup(core_yyscan_t yyscanner)
{
    int            llen = yyextra->literallen;
    char       *new;
    new = palloc(llen + 1);
    memcpy(new, yyextra->literalbuf, llen);
    new[llen] = '\0';
    return new;
}
static int
process_integer_literal(const char *token, YYSTYPE *lval)
{
    //处理整型字面值
    int            val;
    char       *endptr;
    errno = 0;
    val = strtoint(token, &endptr, 10);
    if (*endptr != '\0' || errno == ERANGE)
    {
        
        lval->str = pstrdup(token);
        return FCONST;
    }
    lval->ival = val;
    return ICONST;
}
static unsigned int
hexval(unsigned char c)
{
    //十六进制
    if (c >= '0' && c <= '9')
        return c - '0';
    if (c >= 'a' && c <= 'f')
        return c - 'a' + 0xA;
    if (c >= 'A' && c <= 'F')
        return c - 'A' + 0xA;
    elog(ERROR, "invalid hexadecimal digit");
    return 0;                    
}
static void
check_unicode_value(pg_wchar c, char *loc, core_yyscan_t yyscanner)
{
    if (GetDatabaseEncoding() == PG_UTF8)
        return;
    if (c > 0x7F)
    {
        ADVANCE_YYLLOC(loc - yyextra->literalbuf + 3);    
        yyerror("Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8");
    }
}
static bool
is_utf16_surrogate_first(pg_wchar c)
{
    return (c >= 0xD800 && c <= 0xDBFF);
}
static bool
is_utf16_surrogate_second(pg_wchar c)
{
    return (c >= 0xDC00 && c <= 0xDFFF);
}
static pg_wchar
surrogate_pair_to_codepoint(pg_wchar first, pg_wchar second)
{
    return ((first & 0x3FF) << 10) + 0x10000 + (second & 0x3FF);
}
static void
addunicode(pg_wchar c, core_yyscan_t yyscanner)
{
    char        buf[8];
    if (c == 0 || c > 0x10FFFF)
        yyerror("invalid Unicode escape value");
    if (c > 0x7F)
    {
        if (GetDatabaseEncoding() != PG_UTF8)
            yyerror("Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8");
        yyextra->saw_non_ascii = true;
    }
    unicode_to_utf8(c, (unsigned char *) buf);
    addlit(buf, pg_mblen(buf), yyscanner);
}

static bool
check_uescapechar(unsigned char escape)
{
    if (isxdigit(escape)
        || escape == '+'
        || escape == '\''
        || escape == '"'
        || scanner_isspace(escape))
    {
        return false;
    }
    else
        return true;
}

static char *
litbuf_udeescape(unsigned char escape, core_yyscan_t yyscanner)
{
    char       *new;
    char       *litbuf,
               *in,
               *out;
    pg_wchar    pair_first = 0;
    
    litbuf = yyextra->literalbuf;
    litbuf[yyextra->literallen] = '\0';
    
    new = palloc(yyextra->literallen + 1);
    in = litbuf;
    out = new;
    while (*in)
    {
        if (in[0] == escape)
        {
            if (in[1] == escape)
            {
                if (pair_first)
                {
                    ADVANCE_YYLLOC(in - litbuf + 3);    
                    yyerror("invalid Unicode surrogate pair");
                }
                *out++ = escape;
                in += 2;
            }
            else if (isxdigit((unsigned char) in[1]) &&
                     isxdigit((unsigned char) in[2]) &&
                     isxdigit((unsigned char) in[3]) &&
                     isxdigit((unsigned char) in[4]))
            {
                pg_wchar    unicode;
                unicode = (hexval(in[1]) << 12) +
                    (hexval(in[2]) << 8) +
                    (hexval(in[3]) << 4) +
                    hexval(in[4]);
                check_unicode_value(unicode, in, yyscanner);
                if (pair_first)
                {
                    if (is_utf16_surrogate_second(unicode))
                    {
                        unicode = surrogate_pair_to_codepoint(pair_first, unicode);
                        pair_first = 0;
                    }
                    else
                    {
                        ADVANCE_YYLLOC(in - litbuf + 3);        
                        yyerror("invalid Unicode surrogate pair");
                    }
                }
                else if (is_utf16_surrogate_second(unicode))
                    yyerror("invalid Unicode surrogate pair");
                if (is_utf16_surrogate_first(unicode))
                    pair_first = unicode;
                else
                {
                    unicode_to_utf8(unicode, (unsigned char *) out);
                    out += pg_mblen(out);
                }
                in += 5;
            }
            else if (in[1] == '+' &&
                     isxdigit((unsigned char) in[2]) &&
                     isxdigit((unsigned char) in[3]) &&
                     isxdigit((unsigned char) in[4]) &&
                     isxdigit((unsigned char) in[5]) &&
                     isxdigit((unsigned char) in[6]) &&
                     isxdigit((unsigned char) in[7]))
            {
                pg_wchar    unicode;
                unicode = (hexval(in[2]) << 20) +
                    (hexval(in[3]) << 16) +
                    (hexval(in[4]) << 12) +
                    (hexval(in[5]) << 8) +
                    (hexval(in[6]) << 4) +
                    hexval(in[7]);
                check_unicode_value(unicode, in, yyscanner);
                if (pair_first)
                {
                    if (is_utf16_surrogate_second(unicode))
                    {
                        unicode = surrogate_pair_to_codepoint(pair_first, unicode);
                        pair_first = 0;
                    }
                    else
                    {
                        ADVANCE_YYLLOC(in - litbuf + 3);        
                        yyerror("invalid Unicode surrogate pair");
                    }
                }
                else if (is_utf16_surrogate_second(unicode))
                    yyerror("invalid Unicode surrogate pair");
                if (is_utf16_surrogate_first(unicode))
                    pair_first = unicode;
                else
                {
                    unicode_to_utf8(unicode, (unsigned char *) out);
                    out += pg_mblen(out);
                }
                in += 8;
            }
            else
            {
                ADVANCE_YYLLOC(in - litbuf + 3);        
                yyerror("invalid Unicode escape value");
            }
        }
        else
        {
            if (pair_first)
            {
                ADVANCE_YYLLOC(in - litbuf + 3);        
                yyerror("invalid Unicode surrogate pair");
            }
            *out++ = *in++;
        }
    }
    
    if (pair_first)
    {
        ADVANCE_YYLLOC(in - litbuf + 3);                
        yyerror("invalid Unicode surrogate pair");
    }
    *out = '\0';
    
    pg_verifymbstr(new, out - new, false);
    return new;
}
static unsigned char
unescape_single_char(unsigned char c, core_yyscan_t yyscanner)
{
    switch (c)
    {
        case 'b':
            return '\b';
        case 'f':
            return '\f';
        case 'n':
            return '\n';
        case 'r':
            return '\r';
        case 't':
            return '\t';
        default:
            
            if (c == '\0' || IS_HIGHBIT_SET(c))
                yyextra->saw_non_ascii = true;
            return c;
    }
}
static void
check_string_escape_warning(unsigned char ychar, core_yyscan_t yyscanner)
{
    if (ychar == '\'')
    {
        if (yyextra->warn_on_first_escape && yyextra->escape_string_warning)
            ereport(WARNING,
                    (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER),
                     errmsg("nonstandard use of \\' in a string literal"),
                     errhint("Use '' to write quotes in strings, or use the escape string syntax (E'...')."),
                     lexer_errposition()));
        yyextra->warn_on_first_escape = false;    
    }
    else if (ychar == '\\')
    {
        if (yyextra->warn_on_first_escape && yyextra->escape_string_warning)
            ereport(WARNING,
                    (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER),
                     errmsg("nonstandard use of \\\\ in a string literal"),
                     errhint("Use the escape string syntax for backslashes, e.g., E'\\\\'."),
                     lexer_errposition()));
        yyextra->warn_on_first_escape = false;    
    }
    else
        check_escape_warning(yyscanner);
}
static void
check_escape_warning(core_yyscan_t yyscanner)
{
    if (yyextra->warn_on_first_escape && yyextra->escape_string_warning)
        ereport(WARNING,
                (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER),
                 errmsg("nonstandard use of escape in a string literal"),
        errhint("Use the escape string syntax for escapes, e.g., E'\\r\\n'."),
                 lexer_errposition()));
    yyextra->warn_on_first_escape = false;        
}

void *
core_yyalloc(yy_size_t bytes, core_yyscan_t yyscanner)
{
    return palloc(bytes);
}
void *
core_yyrealloc(void *ptr, yy_size_t bytes, core_yyscan_t yyscanner)
{
    if (ptr)
        return repalloc(ptr, bytes);
    else
        return palloc(bytes);
}
void
core_yyfree(void *ptr, core_yyscan_t yyscanner)
{
    if (ptr)
        pfree(ptr);
}

“PostgreSQL中的User subroutines有什么作用”的内容就介绍到这里了，感谢大家的阅读。如果想了解更多行业相关的知识可以关注编程网网站，小编将为大家输出更多高质量的实用文章！