[COMPRESS-182] Support big or even negative numbers in all numeric TAR headers

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.3
• Fix Version/s: 1.4
• Component/s: Archivers
• Assignee: Unassigned
• Reporter: Stefan Bodewig
• Created: 01/Mar/12 11:48
• Updated: 10/Mar/12 07:03
• Resolved: 10/Mar/12 07:03

Description

This is a superset of the functionality that addressed COMPRESS-175

Jörg Schillig's star and GNU tar may use binary encoding for all numeric fields, PAX/POSIX also provides them inside the extension headers.

The timestamp field may even contain negative numbers.

IMHO Commons Compress should:

• be able to parse numeric fields using binary encoding (positive and negative)

• fix the current binary parser (see discussion in COMPRESS-16) and add a workaround for broken writers (see COMPRESS-181)

• be able to parse all standard fields of PAX headers, including the numeric ones (I haven't checked, maybe it already does)

• have an option to write numbers too big/small in binary encoding much like BIGFILE_STAR does for the file size in trunk

• have an option to write numbers too big/small in PAX headers much like BIGFILE_POSIX does for the file size in trunk

• replace bigFileMode and the constants with a more generic property that controls all numeric fields. We can remove the bigFileMode stuff as it has been added after the 1.3 release.

Issue Links

• incorporates: COMPRESS-175 GNU Tar sometimes uses binary encoding for UID and GID
• relates to: COMPRESS-181 Tar files created by AIX native tar, and which contain symlinks, cannot be read by TarArchiveInputStream