This module is intended to test your regular expressions. Given a subject string and a regular expression
(aka pattern), the module not only tests whether the regular expression complete matches the subject
string, it performs a "utf8::upgrade" or "utf8::downgrade" on the subject string and performs the tests
again, if necessary. Furthermore, given a pattern with capturing parenthesis, it checks whether all
captures are present, and in the right order. Both named and unnamed captures are checked.
By default, the module exports two subroutines, "match" and "no_match". The latter is actually a thin
wrapper around "match", calling it with "match => 0".
"Completematching"
A match is only considered to successfully match if the entire string is matched - that is, if $& matches
the subject string. So:
Subject Pattern
"aaabb" qr /a+b+/ # Considered ok
"aaabb" qr /a+/ # Not considered ok
For efficiency reasons, when the matching is performed the pattern is actually anchored at the start.
It's not anchored at the end as that would potentially influence the matching.
UTF8matching
Certain regular expression constructs match differently depending on whether UTF8 matching is in effect
or not. This is only relevant if the subject string has characters with code points between 128 and 255,
and no characters above 255 -- in such a case, matching may be different depending on whether the subject
string has the UTF8 flag on or not. "Test::Regexp" detects such a case, and will then run the tests
twice; once with the subject string "utf8::downgraded", and once with the subject string
"utf8::upgraded".
Numberoftests
There's no fixed number of tests that is run. The number of tests depends on the number of captures, the
number of different names of captures, and whether there is the need to up- or downgrade the subject
string.
It is therefore recommended to use "use Text::Regexp tests => 'no_plan';". In a later version,
"Test::Regexp" will use a version of "Test::Builder" that allows for nested tests.
Details
The number of tests is as follows:
If no match is expected ("no_match => 0", or "no_match" is used), only one test is performed.
Otherwise (we are expecting a match), if "pattern" is used, there will be three tests.
For "keep_pattern", there will be four tests, plus one tests for each capture, an additional test for
each named capture, and a test for each name used in the set of named captures. So, if there are "N"
captures, there will be at least "4 + N" tests, and at most "4 + 3 * N" tests.
If both "pattern" and "keep_pattern" are used, the number of tests add up.
If "Test::Regexp" decides to upgrade or downgrade, the number of tests double.
"use"options
When using "Test::Regexp", there are a few options you can give it.
"tests => 'no_plan'", "tests => 123"
The number of tests you are going to run. Since takes some work to figure out how many tests will be
run, for now the recommendation is to use "tests => 'no_plan'".
"import => [methods]"
By default, the subroutines "match" and "no_match" are exported. If you want to import a subset, use
the "import" tag, and give it an arrayref with the names of the subroutines to import.
"match"
The subroutine "match" is the workhorse of the module. It takes a number of named arguments, most of them
optional, and runs one or more tests. It returns 1 if all tests were run successfully, and 0 if one or
more tests failed. The following options are available:
"subject => STRING"
The string against which the pattern is tested is passed to "match" using the "subject" option. It's
an error to not pass in a subject.
"pattern => PATTERN", "keep_pattern => PATTERN"
A pattern (aka regular expression) to test can be passed with one of "pattern" or "keep_pattern". The
former should be used if the pattern does not have any matching parenthesis; the latter if the
pattern does have capturing parenthesis. If both "pattern" and "keep_pattern" are provided, the
subject is tested against both. It's an error to not give either "pattern" or "keep_pattern".
"captures => [LIST]"
If a regular expression is passed with "keep_pattern" you should pass in a list of captures using the
"captures" option.
This list should contain all the captures, in order. For unnamed captures, this should just be the
string matched by the capture; for a named capture, this should be a two element array, the first
element being the name of the capture, the second element the capture. Named and unnamed captures may
be mixed, and the same name for a capture may be repeated.
Example:
match subject => "Eland Wapiti Caribou",
keep_pattern => qr /(\w+)\s+(?<a>\w+)\s+(\w+)/,
captures => ["Eland", [a => "Wapiti"], "Caribou"];
"name => NAME"
The "name" of the test. It's being used in the test comment.
"comment => NAME"
An alternative for "name". If both are present, "comment" is used.
"utf8_upgrade => 0", "utf8_downgrade => 0"
As explained in "UTF8 matching", "Test::Regexp" detects whether a subject may provoke different
matching depending on its UTF8 flag, and then it "utf8::upgrades" or "utf8::downgrades" the subject
string and runs the test again. Setting "utf8_upgrade" to 0 prevents "Test::Regexp" from downgrading
the subject string, while setting "utf8_upgrade" to 0 prevents "Test::Regexp" from upgrading the
subject string.
"match => BOOLEAN"
By default, "match" assumes the pattern should match. But it also important to test which strings do
not match a regular expression. This can be done by calling "match" with "match => 0" as parameter.
(Or by calling "no_match" instead of "match"). In this case, the test is a failure if the pattern
completely matches the subject string. A "captures" argument is ignored.
"reason => STRING"
If the match is expected to fail (so, when "match => 0" is passed, or if "no_match" is called), a
reason may be provided with the "reason" option. The reason is then printed in the comment of the
test.
"test => STRING"
If the match is expected to pass (when "match" is called, without "match" being false), and this
option is passed, a message is printed indicating what this specific test is testing (the argument to
"test").
"todo => STRING"
If the "todo" parameter is used (with a defined value), the tests are assumed to be TODO tests. The
argument is used as the TODO message.
"full_text => BOOL"
By default, long test messages are truncated; if a true value is passed, the message will not get
truncated.
"escape => INTEGER"
Controls how non-ASCII and non-printables are displayed in generated test messages:
0 No characters are escape, everything is displayed as is.
1 Show newlines, linefeeds and tabs using their usual escape sequences ("\n", "\r", and "\t").
2 Show any character outside of the printable ASCII characters as named escapes ("\N{UNICODE NAME}"),
or a hex escape if the unicode name is not found ("\x{XX}"). This is the default if "-CO" is not in
effect ("${^UNICODE}" is false).
Newlines, linefeeds and tabs are displayed as above.
3 Show any character outside of the printable ASCII characters as hext escapes ("\x{XX}").
Newlines, linefeeds and tabs are displayed as above.
4 Show the non-printable ASCII characters as hex escapes ("\x{XX}"); any non-ASCII character is
displayed as is. This is the default if "-CO" is in effect ("${^UNICODE}" is true).
Newlines, linefeeds and tabs are displayed as above.
"no_keep_message => BOOL"
If matching against a keeping pattern, a message "(with -Keep)" is added to the comment. Setting this
parameter suppresses this message. Mostly useful for "Regexp::Common510".
"no_match"
Similar to "match", except that it tests whether a pattern does not match a string. Accepts the same
arguments as "match", except for "match".
OOinterface
Since one typically checks a pattern with multiple strings, and it can be tiresome to repeatedly call
"match" or "no_match" with the same arguments, there's also an OO interface. Using a pattern, one
constructs an object and can then repeatedly call the object to match a string.
To construct and initialize the object, call the following:
my $checker = Test::Regexp -> new -> init (
pattern => qr /PATTERN/,
keep_pattern => qr /(PATTERN)/,
...
);
"init" takes exactly the same arguments as "match", with the exception of "subject" and "captures". To
perform a match, all "match" (or "no_match") on the object. The first argument should be the subject the
pattern should match against (see the "subject" argument of "match" discussed above). If there is a match
against a capturing pattern, the second argument is a reference to an array with the matches (see the
"captures" argument of "match" discussed above).
Both "match" and "no_match" can take additional (named) arguments, identical to the none-OO "match" and
"no_match" routines.
Install Now
PNG Icons
