GNU Tar Include and Exclude Behavior

This table represents the results of installcheck/gnutar.pl across multiple GNU Tar versions. Note that this page only deals with include and exclude behavior; see the GNU Tar FAQ entry for other undesirable behaviors.

pat	file	include									exclude
		no args			-wc			-no-wc			no args			-wc			-no-wc
		<1.16	1.16-22	>1.22	<1.16	1.16-22	>1.22	<1.16	1.16-22	>1.22	<1.16	1.16-22	>1.22	<1.16	1.16-22	>1.22	<1.16	1.16-22	>1.22
./A*A	A*A
./A*A	AxA
./A\*A	A*A
./A\*A	AxA
./B?B	B?B
./B?B	BxB
./B\?B	B?B
./B\?B	BxB
./C[C	C[C
./C\[C	C[C
./D\]D	D]D
./D]D	D]D
./E\E	E\E
./E\\E	E\E
./F'F	F'F
./F\'F	F'F
./G"G	G"G
./G\"G	G"G
./H H	H H
./H\ H	H H

This was tested against tar versions 1.15, 1.15.1, 1.16, 1.17, 1.18, 1.19, 1.20, 1.21, 1.22, and 1.23.

Summary

This is the most concise summary I can invent. Yes, there are *five* different matching schemes implemented in GNU tar.

• Single quotes ('), double quotes ("), and spaces always match themselves exactly, regardless of wildcards.
• Includes
  • The default behavior is identical to --no-wildcards
  • Behavior changed with version 1.16:
    • In versions before 1.16, the wildcard option is ignored for includes, and type α wildcard matching is always applied.
    • In versions 1.16 and higher, when wildcard matching is enabled, type β wildcard matching is applied. When wildcard matching is disabled, type ε matching is applied (!).
• Excludes
  • The default behavior is identical to --wildcards
  • When wildcards are disabled, they are truly disabled: only literal matches are accepted.
  • Behavior changed with version 1.23:
    • In versions up to 1.23, when wildcards are enabled, type γ matching is applied.
    • In versions 1.23 and higher, when wildcards are enabled, type δ matching is applied.

Matching types mentioned above:

type α

Only *?[\ are special, and only special characters can be escaped by \ - otherwise, the escaping backslash is treated literally (e.g., E\E matches against itself, but not against EE). There is a bug with \?, which is treated as \0177 internally.

type β

Only *?[\ are special, and \ can escape any character, so \X and X will both match X. However, both \? and \\ are buggy and will not match ? and \, respectively.

type γ

Only *?[\ are special, and \ can escape any character. There is no bug with \? or \\.

type δ

Only *?[ are special, and no escaping is possible.

type ε

Only literal matches are accepted, except that both \ and \\ will match \.

To Do

• Explore the inside of character classes: how do you specify ] or \ in a character class? Negation?
• Look at the source to figure out what's going on with backslashes
• Is type δ really that limited?