mirror of
https://github.com/superseriousbusiness/gotosocial.git
synced 2024-11-30 07:32:45 +00:00
637a57f2de
Bumps [github.com/jackc/pgx/v5](https://github.com/jackc/pgx) from 5.5.1 to 5.5.2. - [Changelog](https://github.com/jackc/pgx/blob/master/CHANGELOG.md) - [Commits](https://github.com/jackc/pgx/compare/v5.5.1...v5.5.2) --- updated-dependencies: - dependency-name: github.com/jackc/pgx/v5 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
270 lines
5.8 KiB
Go
270 lines
5.8 KiB
Go
package pgx
|
|
|
|
import (
|
|
"context"
|
|
"strconv"
|
|
"strings"
|
|
"unicode/utf8"
|
|
)
|
|
|
|
// NamedArgs can be used as the first argument to a query method. It will replace every '@' named placeholder with a '$'
|
|
// ordinal placeholder and construct the appropriate arguments.
|
|
//
|
|
// For example, the following two queries are equivalent:
|
|
//
|
|
// conn.Query(ctx, "select * from widgets where foo = @foo and bar = @bar", pgx.NamedArgs{"foo": 1, "bar": 2})
|
|
// conn.Query(ctx, "select * from widgets where foo = $1 and bar = $2", 1, 2)
|
|
//
|
|
// Named placeholders are case sensitive and must start with a letter or underscore. Subsequent characters can be
|
|
// letters, numbers, or underscores.
|
|
type NamedArgs map[string]any
|
|
|
|
// RewriteQuery implements the QueryRewriter interface.
|
|
func (na NamedArgs) RewriteQuery(ctx context.Context, conn *Conn, sql string, args []any) (newSQL string, newArgs []any, err error) {
|
|
l := &sqlLexer{
|
|
src: sql,
|
|
stateFn: rawState,
|
|
nameToOrdinal: make(map[namedArg]int, len(na)),
|
|
}
|
|
|
|
for l.stateFn != nil {
|
|
l.stateFn = l.stateFn(l)
|
|
}
|
|
|
|
sb := strings.Builder{}
|
|
for _, p := range l.parts {
|
|
switch p := p.(type) {
|
|
case string:
|
|
sb.WriteString(p)
|
|
case namedArg:
|
|
sb.WriteRune('$')
|
|
sb.WriteString(strconv.Itoa(l.nameToOrdinal[p]))
|
|
}
|
|
}
|
|
|
|
newArgs = make([]any, len(l.nameToOrdinal))
|
|
for name, ordinal := range l.nameToOrdinal {
|
|
newArgs[ordinal-1] = na[string(name)]
|
|
}
|
|
|
|
return sb.String(), newArgs, nil
|
|
}
|
|
|
|
type namedArg string
|
|
|
|
type sqlLexer struct {
|
|
src string
|
|
start int
|
|
pos int
|
|
nested int // multiline comment nesting level.
|
|
stateFn stateFn
|
|
parts []any
|
|
|
|
nameToOrdinal map[namedArg]int
|
|
}
|
|
|
|
type stateFn func(*sqlLexer) stateFn
|
|
|
|
func rawState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case 'e', 'E':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune == '\'' {
|
|
l.pos += width
|
|
return escapeStringState
|
|
}
|
|
case '\'':
|
|
return singleQuoteState
|
|
case '"':
|
|
return doubleQuoteState
|
|
case '@':
|
|
nextRune, _ := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if isLetter(nextRune) || nextRune == '_' {
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos-width])
|
|
}
|
|
l.start = l.pos
|
|
return namedArgState
|
|
}
|
|
case '-':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune == '-' {
|
|
l.pos += width
|
|
return oneLineCommentState
|
|
}
|
|
case '/':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune == '*' {
|
|
l.pos += width
|
|
return multilineCommentState
|
|
}
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|
|
|
|
func isLetter(r rune) bool {
|
|
return (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z')
|
|
}
|
|
|
|
func namedArgState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
if r == utf8.RuneError {
|
|
if l.pos-l.start > 0 {
|
|
na := namedArg(l.src[l.start:l.pos])
|
|
if _, found := l.nameToOrdinal[na]; !found {
|
|
l.nameToOrdinal[na] = len(l.nameToOrdinal) + 1
|
|
}
|
|
l.parts = append(l.parts, na)
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
} else if !(isLetter(r) || (r >= '0' && r <= '9') || r == '_') {
|
|
l.pos -= width
|
|
na := namedArg(l.src[l.start:l.pos])
|
|
if _, found := l.nameToOrdinal[na]; !found {
|
|
l.nameToOrdinal[na] = len(l.nameToOrdinal) + 1
|
|
}
|
|
l.parts = append(l.parts, namedArg(na))
|
|
l.start = l.pos
|
|
return rawState
|
|
}
|
|
}
|
|
}
|
|
|
|
func singleQuoteState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case '\'':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune != '\'' {
|
|
return rawState
|
|
}
|
|
l.pos += width
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|
|
|
|
func doubleQuoteState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case '"':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune != '"' {
|
|
return rawState
|
|
}
|
|
l.pos += width
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|
|
|
|
func escapeStringState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case '\\':
|
|
_, width = utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
case '\'':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune != '\'' {
|
|
return rawState
|
|
}
|
|
l.pos += width
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|
|
|
|
func oneLineCommentState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case '\\':
|
|
_, width = utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
case '\n', '\r':
|
|
return rawState
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|
|
|
|
func multilineCommentState(l *sqlLexer) stateFn {
|
|
for {
|
|
r, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
l.pos += width
|
|
|
|
switch r {
|
|
case '/':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune == '*' {
|
|
l.pos += width
|
|
l.nested++
|
|
}
|
|
case '*':
|
|
nextRune, width := utf8.DecodeRuneInString(l.src[l.pos:])
|
|
if nextRune != '/' {
|
|
continue
|
|
}
|
|
|
|
l.pos += width
|
|
if l.nested == 0 {
|
|
return rawState
|
|
}
|
|
l.nested--
|
|
|
|
case utf8.RuneError:
|
|
if l.pos-l.start > 0 {
|
|
l.parts = append(l.parts, l.src[l.start:l.pos])
|
|
l.start = l.pos
|
|
}
|
|
return nil
|
|
}
|
|
}
|
|
}
|