trunk/doc/pcrematching.3

.TH PCREMATCHING 3
.SH NAME
PCRE - Perl-compatible regular expressions
.SH "PCRE MATCHING ALGORITHMS"
.rs
.sp
This document describes the two different algorithms that are available in PCRE
for matching a compiled regular expression against a given subject string. The
"standard" algorithm is the one provided by the \fBpcre_exec()\fP function.
This works in the same was as Perl's matching function, and provides a
Perl-compatible matching operation.
.P
An alternative algorithm is provided by the \fBpcre_dfa_exec()\fP function;
this operates in a different way, and is not Perl-compatible. It has advantages
and disadvantages compared with the standard algorithm, and these are described
below.
.P
When there is only one possible way in which a given subject string can match a
pattern, the two algorithms give the same answer. A difference arises, however,
when there are multiple possibilities. For example, if the pattern
.sp
  ^<.*>
.sp
is matched against the string
.sp
  <something> <something else> <something further>
.sp
there are three possible answers. The standard algorithm finds only one of
them, whereas the alternative algorithm finds all three.
.
.SH "REGULAR EXPRESSIONS AS TREES"
.rs
.sp
The set of strings that are matched by a regular expression can be represented
as a tree structure. An unlimited repetition in the pattern makes the tree of
infinite size, but it is still a tree. Matching the pattern to a given subject
string (from a given starting point) can be thought of as a search of the tree.
There are two ways to search a tree: depth-first and breadth-first, and these
correspond to the two matching algorithms provided by PCRE.
.
.SH "THE STANDARD MATCHING ALGORITHM"
.rs
.sp
In the terminology of Jeffrey Friedl's book "Mastering Regular
Expressions", the standard algorithm is an "NFA algorithm". It conducts a
depth-first search of the pattern tree. That is, it proceeds along a single
path through the tree, checking that the subject matches what is required. When
there is a mismatch, the algorithm tries any alternatives at the current point,
and if they all fail, it backs up to the previous branch point in the tree, and
tries the next alternative branch at that level. This often involves backing up
(moving to the left) in the subject string as well. The order in which
repetition branches are tried is controlled by the greedy or ungreedy nature of
the quantifier.
.P
If a leaf node is reached, a matching string has been found, and at that point
the algorithm stops. Thus, if there is more than one possible match, this
algorithm returns the first one that it finds. Whether this is the shortest,
the longest, or some intermediate length depends on the way the greedy and
ungreedy repetition quantifiers are specified in the pattern.
.P
Because it ends up with a single path through the tree, it is relatively
straightforward for this algorithm to keep track of the substrings that are
matched by portions of the pattern in parentheses. This provides support for
capturing parentheses and back references.
.
.SH "THE ALTERNATIVE MATCHING ALGORITHM"
.rs
.sp
This algorithm conducts a breadth-first search of the tree. Starting from the
first matching point in the subject, it scans the subject string from left to
right, once, character by character, and as it does this, it remembers all the
paths through the tree that represent valid matches. In Friedl's terminology,
this is a kind of "DFA algorithm", though it is not implemented as a
traditional finite state machine (it keeps multiple states active
simultaneously).
.P
The scan continues until either the end of the subject is reached, or there are
no more unterminated paths. At this point, terminated paths represent the
different matching possibilities (if there are none, the match has failed).
Thus, if there is more than one possible match, this algorithm finds all of
them, and in particular, it finds the longest. In PCRE, there is an option to
stop the algorithm after the first match (which is necessarily the shortest)
has been found.
.P
Note that all the matches that are found start at the same point in the
subject. If the pattern
.sp
  cat(er(pillar)?)
.sp
is matched against the string "the caterpillar catchment", the result will be
the three strings "cat", "cater", and "caterpillar" that start at the fourth
character of the subject. The algorithm does not automatically move on to find
matches that start at later positions.
.P
Although the general principle of this matching algorithm is that it scans the 
subject string only once, without backtracking, there is one exception: when a 
lookbehind assertion is encountered, the preceding characters have to be
re-inspected.
.P
There are a number of features of PCRE regular expressions that are not
supported by the alternative matching algorithm. They are as follows:
.P
1. Because the algorithm finds all possible matches, the greedy or ungreedy
nature of repetition quantifiers is not relevant. Greedy and ungreedy
quantifiers are treated in exactly the same way. However, possessive
quantifiers can make a difference when what follows could also match what is
quantified, for example in a pattern like this:
.sp
  ^a++\ew!
.sp
This pattern matches "aaab!" but not "aaa!", which would be matched by a
non-possessive quantifier. Similarly, if an atomic group is present, it is
matched as if it were a standalone pattern at the current point, and the
longest match is then "locked in" for the rest of the overall pattern.
.P
2. When dealing with multiple paths through the tree simultaneously, it is not
straightforward to keep track of captured substrings for the different matching
possibilities, and PCRE's implementation of this algorithm does not attempt to
do this. This means that no captured substrings are available.
.P
3. Because no substrings are captured, back references within the pattern are
not supported, and cause errors if encountered.
.P
4. For the same reason, conditional expressions that use a backreference as the
condition or test for a specific group recursion are not supported.
.P
5. Because many paths through the tree may be active, the \eK escape sequence,
which resets the start of the match when encountered (but may be on some paths
and not on others), is not supported. It causes an error if encountered.
.P
6. Callouts are supported, but the value of the \fIcapture_top\fP field is
always 1, and the value of the \fIcapture_last\fP field is always -1.
.P
7. The \eC escape sequence, which (in the standard algorithm) matches a single
byte, even in UTF-8 mode, is not supported because the alternative algorithm
moves through the subject string one character at a time, for all active paths
through the tree.
.P
8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE) are not
supported. (*FAIL) is supported, and behaves like a failing negative assertion.
.
.SH "ADVANTAGES OF THE ALTERNATIVE ALGORITHM"
.rs
.sp
Using the alternative matching algorithm provides the following advantages:
.P
1. All possible matches (at a single point in the subject) are automatically
found, and in particular, the longest match is found. To find more than one
match using the standard algorithm, you have to do kludgy things with
callouts.
.P
2. Because the alternative algorithm scans the subject string just once, and
never needs to backtrack, it is possible to pass very long subject strings to
the matching function in several pieces, checking for partial matching each
time.
.
.SH "DISADVANTAGES OF THE ALTERNATIVE ALGORITHM"
.rs
.sp
The alternative algorithm suffers from a number of disadvantages:
.P
1. It is substantially slower than the standard algorithm. This is partly
because it has to search for all possible matches, but is also because it is
less susceptible to optimization.
.P
2. Capturing parentheses and back references are not supported.
.P
3. Although atomic groups are supported, their use does not provide the
performance advantage that it does for the standard algorithm.
.
.
.SH AUTHOR
.rs
.sp
.nf
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
.fi
.
.
.SH REVISION
.rs
.sp
.nf
Last updated: 05 September 2009
Copyright (c) 1997-2009 University of Cambridge.
.fi
1	.TH PCREMATCHING 3
2	.SH NAME
3	PCRE - Perl-compatible regular expressions
4	.SH "PCRE MATCHING ALGORITHMS"
5	.rs
6	.sp
7	This document describes the two different algorithms that are available in PCRE
8	for matching a compiled regular expression against a given subject string. The
9	"standard" algorithm is the one provided by the \fBpcre_exec()\fP function.
10	This works in the same was as Perl's matching function, and provides a
11	Perl-compatible matching operation.
12	.P
13	An alternative algorithm is provided by the \fBpcre_dfa_exec()\fP function;
14	this operates in a different way, and is not Perl-compatible. It has advantages
15	and disadvantages compared with the standard algorithm, and these are described
16	below.
17	.P
18	When there is only one possible way in which a given subject string can match a
19	pattern, the two algorithms give the same answer. A difference arises, however,
20	when there are multiple possibilities. For example, if the pattern
21	.sp
22	^<.*>
23	.sp
24	is matched against the string
25	.sp
26	<something> <something else> <something further>
27	.sp
28	there are three possible answers. The standard algorithm finds only one of
29	them, whereas the alternative algorithm finds all three.
30	.
31	.SH "REGULAR EXPRESSIONS AS TREES"
32	.rs
33	.sp
34	The set of strings that are matched by a regular expression can be represented
35	as a tree structure. An unlimited repetition in the pattern makes the tree of
36	infinite size, but it is still a tree. Matching the pattern to a given subject
37	string (from a given starting point) can be thought of as a search of the tree.
38	There are two ways to search a tree: depth-first and breadth-first, and these
39	correspond to the two matching algorithms provided by PCRE.
40	.
41	.SH "THE STANDARD MATCHING ALGORITHM"
42	.rs
43	.sp
44	In the terminology of Jeffrey Friedl's book "Mastering Regular
45	Expressions", the standard algorithm is an "NFA algorithm". It conducts a
46	depth-first search of the pattern tree. That is, it proceeds along a single
47	path through the tree, checking that the subject matches what is required. When
48	there is a mismatch, the algorithm tries any alternatives at the current point,
49	and if they all fail, it backs up to the previous branch point in the tree, and
50	tries the next alternative branch at that level. This often involves backing up
51	(moving to the left) in the subject string as well. The order in which
52	repetition branches are tried is controlled by the greedy or ungreedy nature of
53	the quantifier.
54	.P
55	If a leaf node is reached, a matching string has been found, and at that point
56	the algorithm stops. Thus, if there is more than one possible match, this
57	algorithm returns the first one that it finds. Whether this is the shortest,
58	the longest, or some intermediate length depends on the way the greedy and
59	ungreedy repetition quantifiers are specified in the pattern.
60	.P
61	Because it ends up with a single path through the tree, it is relatively
62	straightforward for this algorithm to keep track of the substrings that are
63	matched by portions of the pattern in parentheses. This provides support for
64	capturing parentheses and back references.
65	.
66	.SH "THE ALTERNATIVE MATCHING ALGORITHM"
67	.rs
68	.sp
69	This algorithm conducts a breadth-first search of the tree. Starting from the
70	first matching point in the subject, it scans the subject string from left to
71	right, once, character by character, and as it does this, it remembers all the
72	paths through the tree that represent valid matches. In Friedl's terminology,
73	this is a kind of "DFA algorithm", though it is not implemented as a
74	traditional finite state machine (it keeps multiple states active
75	simultaneously).
76	.P
77	The scan continues until either the end of the subject is reached, or there are
78	no more unterminated paths. At this point, terminated paths represent the
79	different matching possibilities (if there are none, the match has failed).
80	Thus, if there is more than one possible match, this algorithm finds all of
81	them, and in particular, it finds the longest. In PCRE, there is an option to
82	stop the algorithm after the first match (which is necessarily the shortest)
83	has been found.
84	.P
85	Note that all the matches that are found start at the same point in the
86	subject. If the pattern
87	.sp
88	cat(er(pillar)?)
89	.sp
90	is matched against the string "the caterpillar catchment", the result will be
91	the three strings "cat", "cater", and "caterpillar" that start at the fourth
92	character of the subject. The algorithm does not automatically move on to find
93	matches that start at later positions.
94	.P
95	Although the general principle of this matching algorithm is that it scans the
96	subject string only once, without backtracking, there is one exception: when a
97	lookbehind assertion is encountered, the preceding characters have to be
98	re-inspected.
99	.P
100	There are a number of features of PCRE regular expressions that are not
101	supported by the alternative matching algorithm. They are as follows:
102	.P
103	1. Because the algorithm finds all possible matches, the greedy or ungreedy
104	nature of repetition quantifiers is not relevant. Greedy and ungreedy
105	quantifiers are treated in exactly the same way. However, possessive
106	quantifiers can make a difference when what follows could also match what is
107	quantified, for example in a pattern like this:
108	.sp
109	^a++\ew!
110	.sp
111	This pattern matches "aaab!" but not "aaa!", which would be matched by a
112	non-possessive quantifier. Similarly, if an atomic group is present, it is
113	matched as if it were a standalone pattern at the current point, and the
114	longest match is then "locked in" for the rest of the overall pattern.
115	.P
116	2. When dealing with multiple paths through the tree simultaneously, it is not
117	straightforward to keep track of captured substrings for the different matching
118	possibilities, and PCRE's implementation of this algorithm does not attempt to
119	do this. This means that no captured substrings are available.
120	.P
121	3. Because no substrings are captured, back references within the pattern are
122	not supported, and cause errors if encountered.
123	.P
124	4. For the same reason, conditional expressions that use a backreference as the
125	condition or test for a specific group recursion are not supported.
126	.P
127	5. Because many paths through the tree may be active, the \eK escape sequence,
128	which resets the start of the match when encountered (but may be on some paths
129	and not on others), is not supported. It causes an error if encountered.
130	.P
131	6. Callouts are supported, but the value of the \fIcapture_top\fP field is
132	always 1, and the value of the \fIcapture_last\fP field is always -1.
133	.P
134	7. The \eC escape sequence, which (in the standard algorithm) matches a single
135	byte, even in UTF-8 mode, is not supported because the alternative algorithm
136	moves through the subject string one character at a time, for all active paths
137	through the tree.
138	.P
139	8. Except for (FAIL), the backtracking control verbs such as (PRUNE) are not
140	supported. (*FAIL) is supported, and behaves like a failing negative assertion.
141	.
142	.SH "ADVANTAGES OF THE ALTERNATIVE ALGORITHM"
143	.rs
144	.sp
145	Using the alternative matching algorithm provides the following advantages:
146	.P
147	1. All possible matches (at a single point in the subject) are automatically
148	found, and in particular, the longest match is found. To find more than one
149	match using the standard algorithm, you have to do kludgy things with
150	callouts.
151	.P
152	2. Because the alternative algorithm scans the subject string just once, and
153	never needs to backtrack, it is possible to pass very long subject strings to
154	the matching function in several pieces, checking for partial matching each
155	time.
156	.
157	.SH "DISADVANTAGES OF THE ALTERNATIVE ALGORITHM"
158	.rs
159	.sp
160	The alternative algorithm suffers from a number of disadvantages:
161	.P
162	1. It is substantially slower than the standard algorithm. This is partly
163	because it has to search for all possible matches, but is also because it is
164	less susceptible to optimization.
165	.P
166	2. Capturing parentheses and back references are not supported.
167	.P
168	3. Although atomic groups are supported, their use does not provide the
169	performance advantage that it does for the standard algorithm.
170	.
171	.
172	.SH AUTHOR
173	.rs
174	.sp
175	.nf
176	Philip Hazel
177	University Computing Service
178	Cambridge CB2 3QH, England.
179	.fi
180	.
181	.
182	.SH REVISION
183	.rs
184	.sp
185	.nf
186	Last updated: 05 September 2009
187	Copyright (c) 1997-2009 University of Cambridge.
188	.fi
Name	Value
svn:eol-style	native
svn:keywords	"Author Date Id Revision Url"