| import 'dart:io'; | |
| import 'package:path/path.dart' as p; | |
| import 'package:tar/tar.dart'; | |
| const outputName = 'self.tar.gz'; | |
| /// This example creates a `.tar.gz` file of the directory its running in. | |
| Future<void> main() { | |
| final entries = findEntries(); | |
| final output = File(outputName); | |
| return entries | |
| .transform(tarWriter) | |
| .transform(gzip.encoder) | |
| .pipe(output.openWrite()); | |
| } | |
| Stream<TarEntry> findEntries() async* { | |
| // Build a stream of tar entries by going through each file system entity in | |
| // the current directory. | |
| final root = Directory.current; | |
| await for (final entry in root.list(recursive: true)) { | |
| // We could write directories too, but we only care about files for | |
| // simplicity. | |
| if (entry is! File) continue; | |
| final name = p.relative(entry.path, from: root.path); | |
| // Let's also ignore hidden directories and files. | |
| if (name.split(p.separator).any((part) => part.startsWith('.'))) continue; | |
| // Finally, we should ignore the output file since weird things may happen | |
| // otherwise. | |
| if (name == outputName) continue; | |
| final stat = entry.statSync(); | |
| yield TarEntry( | |
| TarHeader( | |
| name: name, | |
| typeFlag: TypeFlag.reg, // It's a regular file | |
| // Apart from that, copy over meta information | |
| mode: stat.mode, | |
| modified: stat.modified, | |
| accessed: stat.accessed, | |
| changed: stat.changed, | |
| // This assumes that the file won't change until we're writing it into | |
| // the archive later, since then the size might be wrong. It's more | |
| // efficient though, since the tar writer would otherwise have to buffer | |
| // everything to find out the size. | |
| size: stat.size, | |
| ), | |
| // Use entry.openRead() to obtain an input stream for the file that the | |
| // writer will use later. | |
| entry.openRead(), | |
| ); | |
| } | |
| } |