Description
The current logic parsing PAX extension headers fails if the number of bytes used to encode an entry is different from the number of characters - i.e. for any character outside of the ASCII range as the headers are UTF-8 encoded. E.g.
11 path=รค
takes 11 bytes (one has to account for the trailing newline) for 10 characters and the parser fails with "Expected 3 chars, read 2"
Attachments
Issue Links
- blocks
-
COMPRESS-183 Support for de/encoding of tar entry names other than plain 8BIT conversion.
- Resolved