ViewVC logotype

Contents of /code/trunk/doc/pcrepartial.3

Parent Directory Parent Directory | Revision Log Revision Log

Revision 553 - (show annotations)
Fri Oct 22 15:57:50 2010 UTC (10 years, 6 months ago) by ph10
File size: 18014 byte(s)
Change the way PCRE_PARTIAL_HARD handles \z, \Z, \b, \B, and $.
3 PCRE - Perl-compatible regular expressions
5 .rs
6 .sp
7 In normal use of PCRE, if the subject string that is passed to
8 \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP matches as far as it goes, but is
9 too short to match the entire pattern, PCRE_ERROR_NOMATCH is returned. There
10 are circumstances where it might be helpful to distinguish this case from other
11 cases in which there is no match.
12 .P
13 Consider, for example, an application where a human is required to type in data
14 for a field with specific formatting requirements. An example might be a date
15 in the form \fIddmmmyy\fP, defined by this pattern:
16 .sp
17 ^\ed?\ed(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\ed\ed$
18 .sp
19 If the application sees the user's keystrokes one by one, and can check that
20 what has been typed so far is potentially valid, it is able to raise an error
21 as soon as a mistake is made, by beeping and not reflecting the character that
22 has been typed, for example. This immediate feedback is likely to be a better
23 user interface than a check that is delayed until the entire string has been
24 entered. Partial matching can also be useful when the subject string is very
25 long and is not all available at once.
26 .P
27 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
28 PCRE_PARTIAL_HARD options, which can be set when calling \fBpcre_exec()\fP or
29 \fBpcre_dfa_exec()\fP. For backwards compatibility, PCRE_PARTIAL is a synonym
30 for PCRE_PARTIAL_SOFT. The essential difference between the two options is
31 whether or not a partial match is preferred to an alternative complete match,
32 though the details differ between the two matching functions. If both options
33 are set, PCRE_PARTIAL_HARD takes precedence.
34 .P
35 Setting a partial matching option disables two of PCRE's optimizations. PCRE
36 remembers the last literal byte in a pattern, and abandons matching immediately
37 if such a byte is not present in the subject string. This optimization cannot
38 be used for a subject string that might match only partially. If the pattern
39 was studied, PCRE knows the minimum length of a matching string, and does not
40 bother to run the matching function on shorter strings. This optimization is
41 also disabled for partial matching.
42 .
43 .
45 .rs
46 .sp
47 A partial match occurs during a call to \fBpcre_exec()\fP when the end of the
48 subject string is reached successfully, but matching cannot continue because
49 more characters are needed. However, at least one character in the subject must
50 have been inspected. This character need not form part of the final matched
51 string; lookbehind assertions and the \eK escape sequence provide ways of
52 inspecting characters before the start of a matched substring. The requirement
53 for inspecting at least one character exists because an empty string can always
54 be matched; without such a restriction there would always be a partial match of
55 an empty string at the end of the subject.
56 .P
57 If there are at least two slots in the offsets vector when \fBpcre_exec()\fP
58 returns with a partial match, the first slot is set to the offset of the
59 earliest character that was inspected when the partial match was found. For
60 convenience, the second offset points to the end of the subject so that a
61 substring can easily be identified.
62 .P
63 For the majority of patterns, the first offset identifies the start of the
64 partially matched string. However, for patterns that contain lookbehind
65 assertions, or \eK, or begin with \eb or \eB, earlier characters have been
66 inspected while carrying out the match. For example:
67 .sp
68 /(?<=abc)123/
69 .sp
70 This pattern matches "123", but only if it is preceded by "abc". If the subject
71 string is "xyzabc12", the offsets after a partial match are for the substring
72 "abc12", because all these characters are needed if another match is tried
73 with extra characters added to the subject.
74 .P
75 What happens when a partial match is identified depends on which of the two
76 partial matching options are set.
77 .
78 .
79 .SS "PCRE_PARTIAL_SOFT with pcre_exec()"
80 .rs
81 .sp
82 If PCRE_PARTIAL_SOFT is set when \fBpcre_exec()\fP identifies a partial match,
83 the partial match is remembered, but matching continues as normal, and other
84 alternatives in the pattern are tried. If no complete match can be found,
85 \fBpcre_exec()\fP returns PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH.
86 .P
87 This option is "soft" because it prefers a complete match over a partial match.
88 All the various matching items in a pattern behave as if the subject string is
89 potentially complete. For example, \ez, \eZ, and $ match at the end of the
90 subject, as normal, and for \eb and \eB the end of the subject is treated as a
91 non-alphanumeric.
92 .P
93 If there is more than one partial match, the first one that was found provides
94 the data that is returned. Consider this pattern:
95 .sp
96 /123\ew+X|dogY/
97 .sp
98 If this is matched against the subject string "abc123dog", both
99 alternatives fail to match, but the end of the subject is reached during
100 matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3 and 9,
101 identifying "123dog" as the first partial match that was found. (In this
102 example, there are two partial matches, because "dog" on its own partially
103 matches the second alternative.)
104 .
105 .
106 .SS "PCRE_PARTIAL_HARD with pcre_exec()"
107 .rs
108 .sp
109 If PCRE_PARTIAL_HARD is set for \fBpcre_exec()\fP, it returns
110 PCRE_ERROR_PARTIAL as soon as a partial match is found, without continuing to
111 search for possible complete matches. This option is "hard" because it prefers
112 an earlier partial match over a later complete match. For this reason, the
113 assumption is made that the end of the supplied subject string may not be the
114 true end of the available data, and so, if \ez, \eZ, \eb, \eB, or $ are
115 encountered at the end of the subject, the result is PCRE_ERROR_PARTIAL.
116 .
117 .
118 .SS "Comparing hard and soft partial matching"
119 .rs
120 .sp
121 The difference between the two partial matching options can be illustrated by a
122 pattern such as:
123 .sp
124 /dog(sbody)?/
125 .sp
126 This matches either "dog" or "dogsbody", greedily (that is, it prefers the
127 longer string if possible). If it is matched against the string "dog" with
128 PCRE_PARTIAL_SOFT, it yields a complete match for "dog". However, if
129 PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL. On the other hand,
130 if the pattern is made ungreedy the result is different:
131 .sp
132 /dog(sbody)??/
133 .sp
134 In this case the result is always a complete match because \fBpcre_exec()\fP
135 finds that first, and it never continues after finding a match. It might be
136 easier to follow this explanation by thinking of the two patterns like this:
137 .sp
138 /dog(sbody)?/ is the same as /dogsbody|dog/
139 /dog(sbody)??/ is the same as /dog|dogsbody/
140 .sp
141 The second pattern will never match "dogsbody" when \fBpcre_exec()\fP is
142 used, because it will always find the shorter match first.
143 .
144 .
145 .SH "PARTIAL MATCHING USING pcre_dfa_exec()"
146 .rs
147 .sp
148 The \fBpcre_dfa_exec()\fP function moves along the subject string character by
149 character, without backtracking, searching for all possible matches
150 simultaneously. If the end of the subject is reached before the end of the
151 pattern, there is the possibility of a partial match, again provided that at
152 least one character has been inspected.
153 .P
154 When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if there
155 have been no complete matches. Otherwise, the complete matches are returned.
156 However, if PCRE_PARTIAL_HARD is set, a partial match takes precedence over any
157 complete matches. The portion of the string that was inspected when the longest
158 partial match was found is set as the first matching string, provided there are
159 at least two slots in the offsets vector.
160 .P
161 Because \fBpcre_dfa_exec()\fP always searches for all possible matches, and
162 there is no difference between greedy and ungreedy repetition, its behaviour is
163 different from \fBpcre_exec\fP when PCRE_PARTIAL_HARD is set. Consider the
164 string "dog" matched against the ungreedy pattern shown above:
165 .sp
166 /dog(sbody)??/
167 .sp
168 Whereas \fBpcre_exec()\fP stops as soon as it finds the complete match for
169 "dog", \fBpcre_dfa_exec()\fP also finds the partial match for "dogsbody", and
170 so returns that when PCRE_PARTIAL_HARD is set.
171 .
172 .
174 .rs
175 .sp
176 If a pattern ends with one of sequences \eb or \eB, which test for word
177 boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-intuitive
178 results. Consider this pattern:
179 .sp
180 /\ebcat\eb/
181 .sp
182 This matches "cat", provided there is a word boundary at either end. If the
183 subject string is "the cat", the comparison of the final "t" with a following
184 character cannot take place, so a partial match is found. However,
185 \fBpcre_exec()\fP carries on with normal matching, which matches \eb at the end
186 of the subject when the last character is a letter, thus finding a complete
187 match. The result, therefore, is \fInot\fP PCRE_ERROR_PARTIAL. The same thing
188 happens with \fBpcre_dfa_exec()\fP, because it also finds the complete match.
189 .P
190 Using PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL, because
191 then the partial match takes precedence.
192 .
193 .
195 .rs
196 .sp
197 For releases of PCRE prior to 8.00, because of the way certain internal
198 optimizations were implemented in the \fBpcre_exec()\fP function, the
199 PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be used with
200 all patterns. From release 8.00 onwards, the restrictions no longer apply, and
201 partial matching with \fBpcre_exec()\fP can be requested for any pattern.
202 .P
203 Items that were formerly restricted were repeated single characters and
204 repeated metasequences. If PCRE_PARTIAL was set for a pattern that did not
205 conform to the restrictions, \fBpcre_exec()\fP returned the error code
206 PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
207 PCRE_INFO_OKPARTIAL call to \fBpcre_fullinfo()\fP to find out if a compiled
208 pattern can be used for partial matching now always returns 1.
209 .
210 .
212 .rs
213 .sp
214 If the escape sequence \eP is present in a \fBpcretest\fP data line, the
215 PCRE_PARTIAL_SOFT option is used for the match. Here is a run of \fBpcretest\fP
216 that uses the date example quoted above:
217 .sp
218 re> /^\ed?\ed(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\ed\ed$/
219 data> 25jun04\eP
220 0: 25jun04
221 1: jun
222 data> 25dec3\eP
223 Partial match: 23dec3
224 data> 3ju\eP
225 Partial match: 3ju
226 data> 3juj\eP
227 No match
228 data> j\eP
229 No match
230 .sp
231 The first data string is matched completely, so \fBpcretest\fP shows the
232 matched substrings. The remaining four strings do not match the complete
233 pattern, but the first two are partial matches. Similar output is obtained
234 when \fBpcre_dfa_exec()\fP is used.
235 .P
236 If the escape sequence \eP is present more than once in a \fBpcretest\fP data
237 line, the PCRE_PARTIAL_HARD option is set for the match.
238 .
239 .
240 .SH "MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()"
241 .rs
242 .sp
243 When a partial match has been found using \fBpcre_dfa_exec()\fP, it is possible
244 to continue the match by providing additional subject data and calling
245 \fBpcre_dfa_exec()\fP again with the same compiled regular expression, this
246 time setting the PCRE_DFA_RESTART option. You must pass the same working
247 space as before, because this is where details of the previous partial match
248 are stored. Here is an example using \fBpcretest\fP, using the \eR escape
249 sequence to set the PCRE_DFA_RESTART option (\eD specifies the use of
250 \fBpcre_dfa_exec()\fP):
251 .sp
252 re> /^\ed?\ed(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\ed\ed$/
253 data> 23ja\eP\eD
254 Partial match: 23ja
255 data> n05\eR\eD
256 0: n05
257 .sp
258 The first call has "23ja" as the subject, and requests partial matching; the
259 second call has "n05" as the subject for the continued (restarted) match.
260 Notice that when the match is complete, only the last part is shown; PCRE does
261 not retain the previously partially-matched string. It is up to the calling
262 program to do that if it needs to.
263 .P
264 You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
265 PCRE_DFA_RESTART to continue partial matching over multiple segments. This
266 facility can be used to pass very long subject strings to
267 \fBpcre_dfa_exec()\fP.
268 .
269 .
271 .rs
272 .sp
273 From release 8.00, \fBpcre_exec()\fP can also be used to do multi-segment
274 matching. Unlike \fBpcre_dfa_exec()\fP, it is not possible to restart the
275 previous match with a new segment of data. Instead, new data must be added to
276 the previous subject string, and the entire match re-run, starting from the
277 point where the partial match occurred. Earlier data can be discarded. It is
278 best to use PCRE_PARTIAL_HARD in this situation, because it does not treat the
279 end of a segment as the end of the subject when matching \ez, \eZ, \eb, \eB,
280 and $. Consider an unanchored pattern that matches dates:
281 .sp
282 re> /\ed?\ed(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\ed\ed/
283 data> The date is 23ja\eP\eP
284 Partial match: 23ja
285 .sp
286 At this stage, an application could discard the text preceding "23ja", add on
287 text from the next segment, and call \fBpcre_exec()\fP again. Unlike
288 \fBpcre_dfa_exec()\fP, the entire matching string must always be available, and
289 the complete matching process occurs for each call, so more memory and more
290 processing time is needed.
291 .P
292 \fBNote:\fP If the pattern contains lookbehind assertions, or \eK, or starts
293 with \eb or \eB, the string that is returned for a partial match will include
294 characters that precede the partially matched string itself, because these must
295 be retained when adding on more characters for a subsequent matching attempt.
296 .
297 .
299 .rs
300 .sp
301 Certain types of pattern may give problems with multi-segment matching,
302 whichever matching function is used.
303 .P
304 1. If the pattern contains a test for the beginning of a line, you need to pass
305 the PCRE_NOTBOL option when the subject string for any call does start at the
306 beginning of a line. There is also a PCRE_NOTEOL option, but in practice when
307 doing multi-segment matching you should be using PCRE_PARTIAL_HARD, which
308 includes the effect of PCRE_NOTEOL.
309 .P
310 2. Lookbehind assertions at the start of a pattern are catered for in the
311 offsets that are returned for a partial match. However, in theory, a lookbehind
312 assertion later in the pattern could require even earlier characters to be
313 inspected, and it might not have been reached when a partial match occurs. This
314 is probably an extremely unlikely case; you could guard against it to a certain
315 extent by always including extra characters at the start.
316 .P
317 3. Matching a subject string that is split into multiple segments may not
318 always produce exactly the same result as matching over one single long string,
319 especially when PCRE_PARTIAL_SOFT is used. The section "Partial Matching and
320 Word Boundaries" above describes an issue that arises if the pattern ends with
321 \eb or \eB. Another kind of difference may occur when there are multiple
322 matching possibilities, because (for PCRE_PARTIAL_SOFT) a partial match result
323 is given only when there are no completed matches. This means that as soon as
324 the shortest match has been found, continuation to a new subject segment is no
325 longer possible. Consider again this \fBpcretest\fP example:
326 .sp
327 re> /dog(sbody)?/
328 data> dogsb\eP
329 0: dog
330 data> do\eP\eD
331 Partial match: do
332 data> gsb\eR\eP\eD
333 0: g
334 data> dogsbody\eD
335 0: dogsbody
336 1: dog
337 .sp
338 The first data line passes the string "dogsb" to \fBpcre_exec()\fP, setting the
339 PCRE_PARTIAL_SOFT option. Although the string is a partial match for
340 "dogsbody", the result is not PCRE_ERROR_PARTIAL, because the shorter string
341 "dog" is a complete match. Similarly, when the subject is presented to
342 \fBpcre_dfa_exec()\fP in several parts ("do" and "gsb" being the first two) the
343 match stops when "dog" has been found, and it is not possible to continue. On
344 the other hand, if "dogsbody" is presented as a single string,
345 \fBpcre_dfa_exec()\fP finds both matches.
346 .P
347 Because of these problems, it is best to use PCRE_PARTIAL_HARD when matching
348 multi-segment data. The example above then behaves differently:
349 .sp
350 re> /dog(sbody)?/
351 data> dogsb\eP\eP
352 Partial match: dogsb
353 data> do\eP\eD
354 Partial match: do
355 data> gsb\eR\eP\eP\eD
356 Partial match: gsb
357 .sp
358 .P
359 4. Patterns that contain alternatives at the top level which do not all
360 start with the same pattern item may not work as expected when
361 PCRE_DFA_RESTART is used with \fBpcre_dfa_exec()\fP. For example, consider this
362 pattern:
363 .sp
364 1234|3789
365 .sp
366 If the first part of the subject is "ABC123", a partial match of the first
367 alternative is found at offset 3. There is no partial match for the second
368 alternative, because such a match does not start at the same point in the
369 subject string. Attempting to continue with the string "7890" does not yield a
370 match because only those alternatives that match at one point in the subject
371 are remembered. The problem arises because the start of the second alternative
372 matches within the first alternative. There is no problem with anchored
373 patterns or patterns such as:
374 .sp
375 1234|ABCD
376 .sp
377 where no string can be a partial match for both alternatives. This is not a
378 problem if \fBpcre_exec()\fP is used, because the entire match has to be rerun
379 each time:
380 .sp
381 re> /1234|3789/
382 data> ABC123\eP\eP
383 Partial match: 123
384 data> 1237890
385 0: 3789
386 .sp
387 Of course, instead of using PCRE_DFA_RESTART, the same technique of re-running
388 the entire match can also be used with \fBpcre_dfa_exec()\fP. Another
389 possibility is to work with two buffers. If a partial match at offset \fIn\fP
390 in the first buffer is followed by "no match" when PCRE_DFA_RESTART is used on
391 the second buffer, you can then try a new match starting at offset \fIn+1\fP in
392 the first buffer.
393 .
394 .
396 .rs
397 .sp
398 .nf
399 Philip Hazel
400 University Computing Service
401 Cambridge CB2 3QH, England.
402 .fi
403 .
404 .
406 .rs
407 .sp
408 .nf
409 Last updated: 22 October 2010
410 Copyright (c) 1997-2010 University of Cambridge.
411 .fi


Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5