package parser import ( "bufio" "errors" "fmt" "io" "unicode" "unicode/utf8" ) // position represents the current position of the reader in the buffer. It has // both a row and column, as well as a previous column. We need that previous // column to figure out our position when backtracking with reader.unreadRune(). type position struct { col uint row uint prevCol uint } // String returns the position of the last-processed rune in the buffer in the // standard ‘row:col’ format, while starting at 1:1 to be vim-compatible. func (p position) String() string { return fmt.Sprintf("%d:%d", p.row+1, p.col) } // reader represents the actual parser. It has both an underlying buffered // reader, and a position. type reader struct { r *bufio.Reader pos position } // peakRune returns the next rune in the buffer without actually moving the // parser forwards. func (reader *reader) peekRune() (rune, error) { bytes, _ := reader.r.Peek(4) r, size := utf8.DecodeRune(bytes) switch { case r == utf8.RuneError && size == 0: return 0, io.EOF case r == utf8.RuneError && size == 1: return 0, errors.New("Tried to decode malformed UTF-8") } return r, nil } // unreadRune moves the parser one rune back, allowing for basic backtracking. // You can only safely call reader.unreadRune() once before reading again. If // not you will risk messing up the parsers position tracking if you unread // multiple newlines. func (reader *reader) unreadRune() error { if reader.pos.col == 0 { reader.pos.col = reader.pos.prevCol reader.pos.row-- } else { reader.pos.col-- } return reader.r.UnreadRune() } // readRune reads and returns the next rune in the readers buffer. It is just a // wrapper around ‘reader.r.ReadRune()’ that updates the readers position. func (reader *reader) readRune() (rune, error) { r, _, err := reader.r.ReadRune() if r == '\n' { reader.pos.prevCol = reader.pos.col reader.pos.col = 0 reader.pos.row++ } else { reader.pos.col++ } return r, err } // readNonSpaceRune is identical to reader.readRune(), except it skips any runes // representing spaces (as defined by unicode). func (reader *reader) readNonSpaceRune() (rune, error) { if err := reader.skipSpaces(); err != nil { return 0, err } return reader.readRune() } // skipSpaces moves the parser forwards so that the next call to // ‘reader.readRune()’ is guaranteed to read a non-space rune as defined by // Unicode. func (reader *reader) skipSpaces() error { for { if r, err := reader.readRune(); err != nil { if err == io.EOF { return nil } return err } else if !unicode.IsSpace(r) { return reader.unreadRune() } } }