0
Fork 0
mirror of https://github.com/caddyserver/caddy.git synced 2024-12-23 22:27:38 -05:00
caddy/config/lexer.go
2015-01-21 12:09:01 -07:00

114 lines
1.9 KiB
Go

package config
import (
"bufio"
"io"
"unicode"
)
type (
// lexer is a utility which can get values, token by
// token, from a reader. A token is a word, and tokens
// are separated by whitespace. A word can be enclosed in
// quotes if it contains whitespace.
lexer struct {
reader *bufio.Reader
token token
line int
}
// token represents a single processable unit.
token struct {
line int
text string
}
)
// load prepares the lexer to scan a file for tokens.
func (l *lexer) load(file io.Reader) error {
l.reader = bufio.NewReader(file)
l.line = 1
return nil
}
// next loads the next token into the lexer.
// A token is delimited by whitespace, unless
// the token starts with a quotes character (")
// in which case the token goes until the closing
// quotes (the enclosing quotes are not included).
// The rest of the line is skipped if a "#"
// character is read in. Returns true if a token
// was loaded; false otherwise.
func (l *lexer) next() bool {
var val []rune
var comment, quoted, escaped bool
makeToken := func() bool {
l.token.text = string(val)
return true
}
for {
ch, _, err := l.reader.ReadRune()
if err != nil {
if len(val) > 0 {
return makeToken()
}
if err == io.EOF {
return false
} else {
panic(err)
}
}
if quoted {
if !escaped {
if ch == '\\' {
escaped = true
continue
} else if ch == '"' {
quoted = false
return makeToken()
}
}
if ch == '\n' {
l.line++
}
val = append(val, ch)
escaped = false
continue
}
if unicode.IsSpace(ch) {
if ch == '\r' {
continue
}
if ch == '\n' {
l.line++
comment = false
}
if len(val) > 0 {
return makeToken()
}
continue
}
if ch == '#' {
comment = true
}
if comment {
continue
}
if len(val) == 0 {
l.token = token{line: l.line}
if ch == '"' {
quoted = true
continue
}
}
val = append(val, ch)
}
}