.xx meta.keywords="regular expression pattern match regression test"

.MT 4

.TL

AT&T Research regex(3) regression tests

.AF "AT&T Research - Florham Park NJ"

.AU "Glenn Fowler <gsf@research.att.com>"

.H 1

.xx link="testregex.c testregex.c 2004-05-31"

is the latest source for the AT&T Research regression test

harness for the

.xx link="http://www.opengroup.org/onlinepubs/007904975/functions/regcomp.html X/Open regex"

pattern match interface.

See

.BR testregex (1)

for option and test input details.

The source and test data posted here are license free.

.P

.B testregex

can:

.BL

.LI

verify stability for a particular implementation in the face of

source code and/or compilation environment changes

.LI

verify standard compliance for all implementations

.LI

provide a basis for discussions on what

.I compliance

means

.LE

.P

See

.xx link="re-interpretation.html An Interpretation of the POSIX regex Standards"

for an analysis of the POSIX-X/Open

.B regex

standards.

.H 1 "Reference Implementations"

.B testregex

is currently built against these reference implementations:

.TS

center box;

rb cb lb

r c l.

NAME LABEL AUTHORS

AT&T ast \h'0*\w"http://www.research.att.com/sw/download/"'A\h'0' Glenn Fowler and Doug McIlroy

bsd \h'0*\w"ftp://ftp.netbsd.org/pub/NetBSD/NetBSD-1.5.2/source/sets/src.tgz"'B\h'0' \|

Bell Labs \h'0*\w"http://www.bell-labs.com/"'D\h'0' Doug McIlroy

old gnu \h'0*\w"http://www.gnu.org"'G\h'0' \|

gnu \h'0*\w"http://www.gnu.org"'H\h'0' Isamu Hasegawa

irix \h'0*\w"http://www.sgi.com"'I\h'0' \|

boost \h'0*\w"http://www.boost.org/libs/regex/"'J\h'0' John Maddock

regex++ \h'0*\w"http://ourworld.compuserve.com/homepages/John_Maddock/regexpp.htm"'M\h'0' John Maddock

pcre perl compatible \h'0*\w"http://www.pcre.org/"'P\h'0' Philip Hazel

rx \h'0*\w"ftp://regexps.com/pub/src/hackerlab/"'R\h'0' Tom Lord

spencer \h'0*\w"http://arglist.com/regex/rxspencer-alpha3.8.g2.tar.gz"'S\h'0' Henry Spencer

libtre \h'0*\w"http://kouli.iki.fi/~vlaurika/libtre/"'T\h'0' Ville Laurikari

unix caldera \h'0*\w"http://unixtools.sourceforge.net/"'U\h'0' \|

.TE

.H 1 "Test Data Repository"

.TS

center box;

r l.

\h'0*\w"basic.dat"'basic.dat\h'0' \|\|basic regex(3) -- all implementations should pass these

\h'0*\w"categorize.dat"'categorize.dat\h'0' \|\|\h'0*\w"./re-categorize.html"'implementation categorization\h'0'

\h'0*\w"nullsubexpr.dat"'nullsubexpr.dat\h'0' \|\|\h'0*\w"./re-nullsubexpr.html"'null (...)* tests\h'0'

\h'0*\w"leftassoc.dat"'leftassoc.dat\h'0' \|\|\h'0*\w"./re-assoc.html"'left associative catenation implementation must pass these\h'0'

\h'0*\w"rightassoc.dat"'rightassoc.dat\h'0' \|\|\h'0*\w"./re-assoc.html"'right associative catenation implementation must pass these\h'0'

\h'0*\w"forcedassoc.dat"'forcedassoc.dat\h'0' \|\|\h'0*\w"./re-assoc.html"'subexpression grouping to force associativity\h'0'

\h'0*\w"repetition.dat"'repetition.dat\h'0' \|\|\h'0*\w"./re-repetition.html"'explicit vs. implicit repetitions\h'0'

.TE

.H 1 "Usage"

To run the

.B basic.dat

tests:

.EX

testregex < basic.dat

.EE

.P

If the local implementation hangs or dumps on some tests then run with

the \fB-c\fP option.

The \fB-h\fP option lists the test data format details.

The test data files exercise all features;

the test harness detects and ignores features not

supported by the local implementation.

.H 1 "Reference Implementation Notes"

.H 2 "D: diet libc"

The

.xx link="http://www.fefe.de/dietlibc/ diet libc"

implementation is currently omitted because it fails all but one

.B basic.dat

test.

.H 2 "P: PCRE"

The

.B P

implementation emulates

.BR perl (1)

and is not X/Open compliant by design.

The main differences are:

.BL

.LI

.B P

.I "leftmost-first"

matching as opposed to the X/Open

.IR "leftmost-longest" .

.LI

.B REG_EXTENDED

patterns only.

.LE

.P

However, the

.B P

package regression tests, and

.BR perl (1)

features creeping into other implementations,

make it reasonable to include here.

.H 1 "testregex Notes"

Extensions to the standard terminology are derived from the AT&T

implementation, unified under

.B <regex.h>

with these modes:

.TS

center allbox;

cb lb lb

r l l.

MODE FLAGS DESCRIPTION

BRE 0 basic RE

ERE REG_EXTENDED egrep RE with perl (...) extensions

ARE REG_AUGMENTED ERE with ! negation, <> word boundaries

SRE REG_SHELL sh patterns

KRE REG_SHELL|REG_AUGMENTED ksh93 patterns: ! @ ( | & ) { }

LRE REG_LITERAL fgrep patterns

.TE

.P

and a few flags to handle

.BR fnmatch (3):

.TS

center allbox;

lb lb

l l.

regex FLAG fnmatch FLAG

REG_SHELL_ESCAPED FNM_NOESCAPE

REG_SHELL_PATH FNM_PATHNAME

REG_SHELL_DOT FNM_PERIOD

.TE

.P

The original

.L testregex.c

was done by Doug McIlroy at Bell Labs.

The current implementation is maintained by Glenn Fowler <gsf@research.att.com>.