ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 691 - (show annotations)
Sun Sep 11 14:31:21 2011 UTC (5 years, 4 months ago) by ph10
File MIME type: text/html
File size: 20361 byte(s)
Final source and document tidies for 8.20-RC1.
1 <html>
2 <head>
3 <title>pcrepartial specification</title>
4 </head>
5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 <h1>pcrepartial man page</h1>
7 <p>
8 Return to the <a href="index.html">PCRE index page</a>.
9 </p>
10 <p>
11 This page is part of the PCRE HTML documentation. It was generated automatically
12 from the original man page. If there is any nonsense in it, please consult the
13 man page, in case the conversion went wrong.
14 <br>
15 <ul>
16 <li><a name="TOC1" href="#SEC1">PARTIAL MATCHING IN PCRE</a>
17 <li><a name="TOC2" href="#SEC2">PARTIAL MATCHING USING pcre_exec()</a>
18 <li><a name="TOC3" href="#SEC3">PARTIAL MATCHING USING pcre_dfa_exec()</a>
20 <li><a name="TOC5" href="#SEC5">FORMERLY RESTRICTED PATTERNS</a>
22 <li><a name="TOC7" href="#SEC7">MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()</a>
23 <li><a name="TOC8" href="#SEC8">MULTI-SEGMENT MATCHING WITH pcre_exec()</a>
24 <li><a name="TOC9" href="#SEC9">ISSUES WITH MULTI-SEGMENT MATCHING</a>
25 <li><a name="TOC10" href="#SEC10">AUTHOR</a>
26 <li><a name="TOC11" href="#SEC11">REVISION</a>
27 </ul>
28 <br><a name="SEC1" href="#TOC1">PARTIAL MATCHING IN PCRE</a><br>
29 <P>
30 In normal use of PCRE, if the subject string that is passed to
31 <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b> matches as far as it goes, but is
32 too short to match the entire pattern, PCRE_ERROR_NOMATCH is returned. There
33 are circumstances where it might be helpful to distinguish this case from other
34 cases in which there is no match.
35 </P>
36 <P>
37 Consider, for example, an application where a human is required to type in data
38 for a field with specific formatting requirements. An example might be a date
39 in the form <i>ddmmmyy</i>, defined by this pattern:
40 <pre>
41 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
42 </pre>
43 If the application sees the user's keystrokes one by one, and can check that
44 what has been typed so far is potentially valid, it is able to raise an error
45 as soon as a mistake is made, by beeping and not reflecting the character that
46 has been typed, for example. This immediate feedback is likely to be a better
47 user interface than a check that is delayed until the entire string has been
48 entered. Partial matching can also be useful when the subject string is very
49 long and is not all available at once.
50 </P>
51 <P>
52 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
53 PCRE_PARTIAL_HARD options, which can be set when calling <b>pcre_exec()</b> or
54 <b>pcre_dfa_exec()</b>. For backwards compatibility, PCRE_PARTIAL is a synonym
55 for PCRE_PARTIAL_SOFT. The essential difference between the two options is
56 whether or not a partial match is preferred to an alternative complete match,
57 though the details differ between the two matching functions. If both options
58 are set, PCRE_PARTIAL_HARD takes precedence.
59 </P>
60 <P>
61 Setting a partial matching option for <b>pcre_exec()</b> disables the use of any
62 just-in-time code that was set up by calling <b>pcre_study()</b> with the
63 PCRE_STUDY_JIT_COMPILE option. It also disables two of PCRE's standard
64 optimizations. PCRE remembers the last literal byte in a pattern, and abandons
65 matching immediately if such a byte is not present in the subject string. This
66 optimization cannot be used for a subject string that might match only
67 partially. If the pattern was studied, PCRE knows the minimum length of a
68 matching string, and does not bother to run the matching function on shorter
69 strings. This optimization is also disabled for partial matching.
70 </P>
71 <br><a name="SEC2" href="#TOC1">PARTIAL MATCHING USING pcre_exec()</a><br>
72 <P>
73 A partial match occurs during a call to <b>pcre_exec()</b> when the end of the
74 subject string is reached successfully, but matching cannot continue because
75 more characters are needed. However, at least one character in the subject must
76 have been inspected. This character need not form part of the final matched
77 string; lookbehind assertions and the \K escape sequence provide ways of
78 inspecting characters before the start of a matched substring. The requirement
79 for inspecting at least one character exists because an empty string can always
80 be matched; without such a restriction there would always be a partial match of
81 an empty string at the end of the subject.
82 </P>
83 <P>
84 If there are at least two slots in the offsets vector when <b>pcre_exec()</b>
85 returns with a partial match, the first slot is set to the offset of the
86 earliest character that was inspected when the partial match was found. For
87 convenience, the second offset points to the end of the subject so that a
88 substring can easily be identified.
89 </P>
90 <P>
91 For the majority of patterns, the first offset identifies the start of the
92 partially matched string. However, for patterns that contain lookbehind
93 assertions, or \K, or begin with \b or \B, earlier characters have been
94 inspected while carrying out the match. For example:
95 <pre>
96 /(?&#60;=abc)123/
97 </pre>
98 This pattern matches "123", but only if it is preceded by "abc". If the subject
99 string is "xyzabc12", the offsets after a partial match are for the substring
100 "abc12", because all these characters are needed if another match is tried
101 with extra characters added to the subject.
102 </P>
103 <P>
104 What happens when a partial match is identified depends on which of the two
105 partial matching options are set.
106 </P>
107 <br><b>
108 PCRE_PARTIAL_SOFT with pcre_exec()
109 </b><br>
110 <P>
111 If PCRE_PARTIAL_SOFT is set when <b>pcre_exec()</b> identifies a partial match,
112 the partial match is remembered, but matching continues as normal, and other
113 alternatives in the pattern are tried. If no complete match can be found,
114 <b>pcre_exec()</b> returns PCRE_ERROR_PARTIAL instead of PCRE_ERROR_NOMATCH.
115 </P>
116 <P>
117 This option is "soft" because it prefers a complete match over a partial match.
118 All the various matching items in a pattern behave as if the subject string is
119 potentially complete. For example, \z, \Z, and $ match at the end of the
120 subject, as normal, and for \b and \B the end of the subject is treated as a
121 non-alphanumeric.
122 </P>
123 <P>
124 If there is more than one partial match, the first one that was found provides
125 the data that is returned. Consider this pattern:
126 <pre>
127 /123\w+X|dogY/
128 </pre>
129 If this is matched against the subject string "abc123dog", both
130 alternatives fail to match, but the end of the subject is reached during
131 matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3 and 9,
132 identifying "123dog" as the first partial match that was found. (In this
133 example, there are two partial matches, because "dog" on its own partially
134 matches the second alternative.)
135 </P>
136 <br><b>
137 PCRE_PARTIAL_HARD with pcre_exec()
138 </b><br>
139 <P>
140 If PCRE_PARTIAL_HARD is set for <b>pcre_exec()</b>, it returns
141 PCRE_ERROR_PARTIAL as soon as a partial match is found, without continuing to
142 search for possible complete matches. This option is "hard" because it prefers
143 an earlier partial match over a later complete match. For this reason, the
144 assumption is made that the end of the supplied subject string may not be the
145 true end of the available data, and so, if \z, \Z, \b, \B, or $ are
146 encountered at the end of the subject, the result is PCRE_ERROR_PARTIAL.
147 </P>
148 <P>
149 Setting PCRE_PARTIAL_HARD also affects the way <b>pcre_exec()</b> checks UTF-8
150 subject strings for validity. Normally, an invalid UTF-8 sequence causes the
151 error PCRE_ERROR_BADUTF8. However, in the special case of a truncated UTF-8
152 character at the end of the subject, PCRE_ERROR_SHORTUTF8 is returned when
154 </P>
155 <br><b>
156 Comparing hard and soft partial matching
157 </b><br>
158 <P>
159 The difference between the two partial matching options can be illustrated by a
160 pattern such as:
161 <pre>
162 /dog(sbody)?/
163 </pre>
164 This matches either "dog" or "dogsbody", greedily (that is, it prefers the
165 longer string if possible). If it is matched against the string "dog" with
166 PCRE_PARTIAL_SOFT, it yields a complete match for "dog". However, if
167 PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL. On the other hand,
168 if the pattern is made ungreedy the result is different:
169 <pre>
170 /dog(sbody)??/
171 </pre>
172 In this case the result is always a complete match because <b>pcre_exec()</b>
173 finds that first, and it never continues after finding a match. It might be
174 easier to follow this explanation by thinking of the two patterns like this:
175 <pre>
176 /dog(sbody)?/ is the same as /dogsbody|dog/
177 /dog(sbody)??/ is the same as /dog|dogsbody/
178 </pre>
179 The second pattern will never match "dogsbody" when <b>pcre_exec()</b> is
180 used, because it will always find the shorter match first.
181 </P>
182 <br><a name="SEC3" href="#TOC1">PARTIAL MATCHING USING pcre_dfa_exec()</a><br>
183 <P>
184 The <b>pcre_dfa_exec()</b> function moves along the subject string character by
185 character, without backtracking, searching for all possible matches
186 simultaneously. If the end of the subject is reached before the end of the
187 pattern, there is the possibility of a partial match, again provided that at
188 least one character has been inspected.
189 </P>
190 <P>
191 When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if there
192 have been no complete matches. Otherwise, the complete matches are returned.
193 However, if PCRE_PARTIAL_HARD is set, a partial match takes precedence over any
194 complete matches. The portion of the string that was inspected when the longest
195 partial match was found is set as the first matching string, provided there are
196 at least two slots in the offsets vector.
197 </P>
198 <P>
199 Because <b>pcre_dfa_exec()</b> always searches for all possible matches, and
200 there is no difference between greedy and ungreedy repetition, its behaviour is
201 different from <b>pcre_exec</b> when PCRE_PARTIAL_HARD is set. Consider the
202 string "dog" matched against the ungreedy pattern shown above:
203 <pre>
204 /dog(sbody)??/
205 </pre>
206 Whereas <b>pcre_exec()</b> stops as soon as it finds the complete match for
207 "dog", <b>pcre_dfa_exec()</b> also finds the partial match for "dogsbody", and
208 so returns that when PCRE_PARTIAL_HARD is set.
209 </P>
210 <br><a name="SEC4" href="#TOC1">PARTIAL MATCHING AND WORD BOUNDARIES</a><br>
211 <P>
212 If a pattern ends with one of sequences \b or \B, which test for word
213 boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-intuitive
214 results. Consider this pattern:
215 <pre>
216 /\bcat\b/
217 </pre>
218 This matches "cat", provided there is a word boundary at either end. If the
219 subject string is "the cat", the comparison of the final "t" with a following
220 character cannot take place, so a partial match is found. However,
221 <b>pcre_exec()</b> carries on with normal matching, which matches \b at the end
222 of the subject when the last character is a letter, thus finding a complete
223 match. The result, therefore, is <i>not</i> PCRE_ERROR_PARTIAL. The same thing
224 happens with <b>pcre_dfa_exec()</b>, because it also finds the complete match.
225 </P>
226 <P>
227 Using PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL, because
228 then the partial match takes precedence.
229 </P>
230 <br><a name="SEC5" href="#TOC1">FORMERLY RESTRICTED PATTERNS</a><br>
231 <P>
232 For releases of PCRE prior to 8.00, because of the way certain internal
233 optimizations were implemented in the <b>pcre_exec()</b> function, the
234 PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be used with
235 all patterns. From release 8.00 onwards, the restrictions no longer apply, and
236 partial matching with <b>pcre_exec()</b> can be requested for any pattern.
237 </P>
238 <P>
239 Items that were formerly restricted were repeated single characters and
240 repeated metasequences. If PCRE_PARTIAL was set for a pattern that did not
241 conform to the restrictions, <b>pcre_exec()</b> returned the error code
242 PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
243 PCRE_INFO_OKPARTIAL call to <b>pcre_fullinfo()</b> to find out if a compiled
244 pattern can be used for partial matching now always returns 1.
245 </P>
246 <br><a name="SEC6" href="#TOC1">EXAMPLE OF PARTIAL MATCHING USING PCRETEST</a><br>
247 <P>
248 If the escape sequence \P is present in a <b>pcretest</b> data line, the
249 PCRE_PARTIAL_SOFT option is used for the match. Here is a run of <b>pcretest</b>
250 that uses the date example quoted above:
251 <pre>
252 re&#62; /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
253 data&#62; 25jun04\P
254 0: 25jun04
255 1: jun
256 data&#62; 25dec3\P
257 Partial match: 23dec3
258 data&#62; 3ju\P
259 Partial match: 3ju
260 data&#62; 3juj\P
261 No match
262 data&#62; j\P
263 No match
264 </pre>
265 The first data string is matched completely, so <b>pcretest</b> shows the
266 matched substrings. The remaining four strings do not match the complete
267 pattern, but the first two are partial matches. Similar output is obtained
268 when <b>pcre_dfa_exec()</b> is used.
269 </P>
270 <P>
271 If the escape sequence \P is present more than once in a <b>pcretest</b> data
272 line, the PCRE_PARTIAL_HARD option is set for the match.
273 </P>
274 <br><a name="SEC7" href="#TOC1">MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()</a><br>
275 <P>
276 When a partial match has been found using <b>pcre_dfa_exec()</b>, it is possible
277 to continue the match by providing additional subject data and calling
278 <b>pcre_dfa_exec()</b> again with the same compiled regular expression, this
279 time setting the PCRE_DFA_RESTART option. You must pass the same working
280 space as before, because this is where details of the previous partial match
281 are stored. Here is an example using <b>pcretest</b>, using the \R escape
282 sequence to set the PCRE_DFA_RESTART option (\D specifies the use of
283 <b>pcre_dfa_exec()</b>):
284 <pre>
285 re&#62; /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
286 data&#62; 23ja\P\D
287 Partial match: 23ja
288 data&#62; n05\R\D
289 0: n05
290 </pre>
291 The first call has "23ja" as the subject, and requests partial matching; the
292 second call has "n05" as the subject for the continued (restarted) match.
293 Notice that when the match is complete, only the last part is shown; PCRE does
294 not retain the previously partially-matched string. It is up to the calling
295 program to do that if it needs to.
296 </P>
297 <P>
298 You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
299 PCRE_DFA_RESTART to continue partial matching over multiple segments. This
300 facility can be used to pass very long subject strings to
301 <b>pcre_dfa_exec()</b>.
302 </P>
303 <br><a name="SEC8" href="#TOC1">MULTI-SEGMENT MATCHING WITH pcre_exec()</a><br>
304 <P>
305 From release 8.00, <b>pcre_exec()</b> can also be used to do multi-segment
306 matching. Unlike <b>pcre_dfa_exec()</b>, it is not possible to restart the
307 previous match with a new segment of data. Instead, new data must be added to
308 the previous subject string, and the entire match re-run, starting from the
309 point where the partial match occurred. Earlier data can be discarded. It is
310 best to use PCRE_PARTIAL_HARD in this situation, because it does not treat the
311 end of a segment as the end of the subject when matching \z, \Z, \b, \B,
312 and $. Consider an unanchored pattern that matches dates:
313 <pre>
314 re&#62; /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
315 data&#62; The date is 23ja\P\P
316 Partial match: 23ja
317 </pre>
318 At this stage, an application could discard the text preceding "23ja", add on
319 text from the next segment, and call <b>pcre_exec()</b> again. Unlike
320 <b>pcre_dfa_exec()</b>, the entire matching string must always be available, and
321 the complete matching process occurs for each call, so more memory and more
322 processing time is needed.
323 </P>
324 <P>
325 <b>Note:</b> If the pattern contains lookbehind assertions, or \K, or starts
326 with \b or \B, the string that is returned for a partial match will include
327 characters that precede the partially matched string itself, because these must
328 be retained when adding on more characters for a subsequent matching attempt.
329 </P>
330 <br><a name="SEC9" href="#TOC1">ISSUES WITH MULTI-SEGMENT MATCHING</a><br>
331 <P>
332 Certain types of pattern may give problems with multi-segment matching,
333 whichever matching function is used.
334 </P>
335 <P>
336 1. If the pattern contains a test for the beginning of a line, you need to pass
337 the PCRE_NOTBOL option when the subject string for any call does start at the
338 beginning of a line. There is also a PCRE_NOTEOL option, but in practice when
339 doing multi-segment matching you should be using PCRE_PARTIAL_HARD, which
340 includes the effect of PCRE_NOTEOL.
341 </P>
342 <P>
343 2. Lookbehind assertions at the start of a pattern are catered for in the
344 offsets that are returned for a partial match. However, in theory, a lookbehind
345 assertion later in the pattern could require even earlier characters to be
346 inspected, and it might not have been reached when a partial match occurs. This
347 is probably an extremely unlikely case; you could guard against it to a certain
348 extent by always including extra characters at the start.
349 </P>
350 <P>
351 3. Matching a subject string that is split into multiple segments may not
352 always produce exactly the same result as matching over one single long string,
353 especially when PCRE_PARTIAL_SOFT is used. The section "Partial Matching and
354 Word Boundaries" above describes an issue that arises if the pattern ends with
355 \b or \B. Another kind of difference may occur when there are multiple
356 matching possibilities, because (for PCRE_PARTIAL_SOFT) a partial match result
357 is given only when there are no completed matches. This means that as soon as
358 the shortest match has been found, continuation to a new subject segment is no
359 longer possible. Consider again this <b>pcretest</b> example:
360 <pre>
361 re&#62; /dog(sbody)?/
362 data&#62; dogsb\P
363 0: dog
364 data&#62; do\P\D
365 Partial match: do
366 data&#62; gsb\R\P\D
367 0: g
368 data&#62; dogsbody\D
369 0: dogsbody
370 1: dog
371 </pre>
372 The first data line passes the string "dogsb" to <b>pcre_exec()</b>, setting the
373 PCRE_PARTIAL_SOFT option. Although the string is a partial match for
374 "dogsbody", the result is not PCRE_ERROR_PARTIAL, because the shorter string
375 "dog" is a complete match. Similarly, when the subject is presented to
376 <b>pcre_dfa_exec()</b> in several parts ("do" and "gsb" being the first two) the
377 match stops when "dog" has been found, and it is not possible to continue. On
378 the other hand, if "dogsbody" is presented as a single string,
379 <b>pcre_dfa_exec()</b> finds both matches.
380 </P>
381 <P>
382 Because of these problems, it is best to use PCRE_PARTIAL_HARD when matching
383 multi-segment data. The example above then behaves differently:
384 <pre>
385 re&#62; /dog(sbody)?/
386 data&#62; dogsb\P\P
387 Partial match: dogsb
388 data&#62; do\P\D
389 Partial match: do
390 data&#62; gsb\R\P\P\D
391 Partial match: gsb
392 </pre>
393 4. Patterns that contain alternatives at the top level which do not all
394 start with the same pattern item may not work as expected when
395 PCRE_DFA_RESTART is used with <b>pcre_dfa_exec()</b>. For example, consider this
396 pattern:
397 <pre>
398 1234|3789
399 </pre>
400 If the first part of the subject is "ABC123", a partial match of the first
401 alternative is found at offset 3. There is no partial match for the second
402 alternative, because such a match does not start at the same point in the
403 subject string. Attempting to continue with the string "7890" does not yield a
404 match because only those alternatives that match at one point in the subject
405 are remembered. The problem arises because the start of the second alternative
406 matches within the first alternative. There is no problem with anchored
407 patterns or patterns such as:
408 <pre>
409 1234|ABCD
410 </pre>
411 where no string can be a partial match for both alternatives. This is not a
412 problem if <b>pcre_exec()</b> is used, because the entire match has to be rerun
413 each time:
414 <pre>
415 re&#62; /1234|3789/
416 data&#62; ABC123\P\P
417 Partial match: 123
418 data&#62; 1237890
419 0: 3789
420 </pre>
421 Of course, instead of using PCRE_DFA_RESTART, the same technique of re-running
422 the entire match can also be used with <b>pcre_dfa_exec()</b>. Another
423 possibility is to work with two buffers. If a partial match at offset <i>n</i>
424 in the first buffer is followed by "no match" when PCRE_DFA_RESTART is used on
425 the second buffer, you can then try a new match starting at offset <i>n+1</i> in
426 the first buffer.
427 </P>
428 <br><a name="SEC10" href="#TOC1">AUTHOR</a><br>
429 <P>
430 Philip Hazel
431 <br>
432 University Computing Service
433 <br>
434 Cambridge CB2 3QH, England.
435 <br>
436 </P>
437 <br><a name="SEC11" href="#TOC1">REVISION</a><br>
438 <P>
439 Last updated: 26 August 2011
440 <br>
441 Copyright &copy; 1997-2011 University of Cambridge.
442 <br>
443 <p>
444 Return to the <a href="index.html">PCRE index page</a>.
445 </p>


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

  ViewVC Help
Powered by ViewVC 1.1.5