X-Git-Url: http://www.git.cypherpunks.ru/?a=blobdiff_plain;f=r.go;h=195a7256aa9ad0f9d276b9f39ad79c956c162c60;hb=ed427fe339b6cbf1bb1c3f4e61625418fb115347;hp=33f3fb959e348bbc95410ce8221d89aba185fdf8;hpb=88188862917fb5f831158bf4a8ce30b62f87e370;p=gorecfile.git diff --git a/r.go b/r.go index 33f3fb9..195a725 100644 --- a/r.go +++ b/r.go @@ -1,6 +1,6 @@ /* recfile -- GNU recutils'es recfiles parser on pure Go -Copyright (C) 2020-2021 Sergey Matveev +Copyright (C) 2020-2022 Sergey Matveev This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -21,12 +21,9 @@ import ( "bufio" "errors" "io" - "regexp" "strings" ) -var KeyValRe = regexp.MustCompile(`([a-zA-Z%][a-zA-Z0-9_]*):\s*(.*)$`) - type Reader struct { scanner *bufio.Scanner } @@ -37,6 +34,31 @@ func NewReader(r io.Reader) *Reader { return &Reader{bufio.NewScanner(r)} } +func getKeyValue(text string) (string, string) { + cols := strings.SplitN(text, ":", 2) + if len(cols) != 2 { + return "", "" + } + k := cols[0] + if len(k) == 0 { + return "", "" + } + if !((k[0] == '%') || + ('a' <= k[0] && k[0] <= 'z') || + ('A' <= k[0] && k[0] <= 'Z')) { + return "", "" + } + for _, c := range k[1:] { + if !((c == '_') || + ('a' <= c && c <= 'z') || + ('A' <= c && c <= 'Z') || + ('0' <= c && c <= '9')) { + return "", "" + } + } + return k, strings.TrimPrefix(cols[1], " ") +} + // Get next record. Each record is just a collection of fields. io.EOF // is returned if there is nothing to read more. func (r *Reader) Next() ([]Field, error) { @@ -94,12 +116,10 @@ func (r *Reader) Next() ([]Field, error) { break } - matches := KeyValRe.FindStringSubmatch(text) - if len(matches) == 0 { + name, line = getKeyValue(text) + if name == "" { return fields, errors.New("invalid field format") } - name = matches[1] - line = matches[2] if len(line) > 0 && line[len(line)-1] == '\\' { continuation = true