ViewVC logotype

Contents of /code/tags/pcre-7.8/HACKING

Parent Directory Parent Directory | Revision Log Revision Log

Revision 375 - (hide annotations) (download)
Fri Sep 5 16:48:55 2008 UTC (6 years, 7 months ago) by ph10
File size: 17900 byte(s)
Tag release 7.8

1 nigel 41 Technical Notes about PCRE
2     --------------------------
4 nigel 91 These are very rough technical notes that record potentially useful information
5     about PCRE internals.
7 nigel 75 Historical note 1
8     -----------------
10 nigel 41 Many years ago I implemented some regular expression functions to an algorithm
11     suggested by Martin Richards. These were not Unix-like in form, and were quite
12     restricted in what they could do by comparison with Perl. The interesting part
13     about the algorithm was that the amount of space required to hold the compiled
14     form of an expression was known in advance. The code to apply an expression did
15 nigel 63 not operate by backtracking, as the original Henry Spencer code and current
16     Perl code does, but instead checked all possibilities simultaneously by keeping
17     a list of current states and checking all of them as it advanced through the
18 nigel 75 subject string. In the terminology of Jeffrey Friedl's book, it was a "DFA
19 nigel 93 algorithm", though it was not a traditional Finite State Machine (FSM). When
20     the pattern was all used up, all remaining states were possible matches, and
21     the one matching the longest subset of the subject string was chosen. This did
22     not necessarily maximize the individual wild portions of the pattern, as is
23     expected in Unix and Perl-style regular expressions.
24 nigel 41
25 nigel 75 Historical note 2
26     -----------------
28 nigel 91 By contrast, the code originally written by Henry Spencer (which was
29     subsequently heavily modified for Perl) compiles the expression twice: once in
30     a dummy mode in order to find out how much store will be needed, and then for
31     real. (The Perl version probably doesn't do this any more; I'm talking about
32     the original library.) The execution function operates by backtracking and
33     maximizing (or, optionally, minimizing in Perl) the amount of the subject that
34     matches individual wild portions of the pattern. This is an "NFA algorithm" in
35     Friedl's terminology.
36 nigel 41
37 nigel 75 OK, here's the real stuff
38     -------------------------
40 nigel 77 For the set of functions that form the "basic" PCRE library (which are
41     unrelated to those mentioned above), I tried at first to invent an algorithm
42     that used an amount of store bounded by a multiple of the number of characters
43     in the pattern, to save on compiling time. However, because of the greater
44     complexity in Perl regular expressions, I couldn't do this. In any case, a
45 nigel 93 first pass through the pattern is helpful for other reasons.
46 nigel 41
47 nigel 93 Computing the memory requirement: how it was
48     --------------------------------------------
50     Up to and including release 6.7, PCRE worked by running a very degenerate first
51     pass to calculate a maximum store size, and then a second pass to do the real
52     compile - which might use a bit less than the predicted amount of memory. The
53     idea was that this would turn out faster than the Henry Spencer code because
54     the first pass is degenerate and the second pass can just store stuff straight
55     into the vector, which it knows is big enough.
57     Computing the memory requirement: how it is
58     -------------------------------------------
60     By the time I was working on a potential 6.8 release, the degenerate first pass
61     had become very complicated and hard to maintain. Indeed one of the early
62     things I did for 6.8 was to fix Yet Another Bug in the memory computation. Then
63     I had a flash of inspiration as to how I could run the real compile function in
64     a "fake" mode that enables it to compute how much memory it would need, while
65     actually only ever using a few hundred bytes of working memory, and without too
66     many tests of the mode that might slow it down. So I re-factored the compiling
67     functions to work this way. This got rid of about 600 lines of source. It
68     should make future maintenance and development easier. As this was such a major
69     change, I never released 6.8, instead upping the number to 7.0 (other quite
70     major changes are also present in the 7.0 release).
72     A side effect of this work is that the previous limit of 200 on the nesting
73     depth of parentheses was removed. However, there is a downside: pcre_compile()
74     runs more slowly than before (30% or more, depending on the pattern) because it
75     is doing a full analysis of the pattern. My hope is that this is not a big
76     issue.
78 nigel 77 Traditional matching function
79     -----------------------------
81     The "traditional", and original, matching function is called pcre_exec(), and
82     it implements an NFA algorithm, similar to the original Henry Spencer algorithm
83     and the way that Perl works. Not surprising, since it is intended to be as
84     compatible with Perl as possible. This is the function most users of PCRE will
85     use most of the time.
87     Supplementary matching function
88     -------------------------------
90     From PCRE 6.0, there is also a supplementary matching function called
91     pcre_dfa_exec(). This implements a DFA matching algorithm that searches
92     simultaneously for all possible matches that start at one point in the subject
93     string. (Going back to my roots: see Historical Note 1 above.) This function
94     intreprets the same compiled pattern data as pcre_exec(); however, not all the
95 nigel 91 facilities are available, and those that are do not always work in quite the
96 nigel 77 same way. See the user documentation for details.
98 nigel 93 The algorithm that is used for pcre_dfa_exec() is not a traditional FSM,
99     because it may have a number of states active at one time. More work would be
100     needed at compile time to produce a traditional FSM where only one state is
101     ever active at once. I believe some other regex matchers work this way.
104 nigel 77 Format of compiled patterns
105     ---------------------------
107 nigel 41 The compiled form of a pattern is a vector of bytes, containing items of
108     variable length. The first byte in an item is an opcode, and the length of the
109 nigel 75 item is either implicit in the opcode or contained in the data bytes that
110     follow it.
111 nigel 41
112 ph10 212 In many cases below LINK_SIZE data values are specified for offsets within the
113     compiled pattern. The default value for LINK_SIZE is 2, but PCRE can be
114 nigel 93 compiled to use 3-byte or 4-byte values for these offsets (impairing the
115 nigel 75 performance). This is necessary only when patterns whose compiled length is
116 nigel 93 greater than 64K are going to be processed. In this description, we assume the
117 ph10 212 "normal" compilation options. Data values that are counts (e.g. for
118     quantifiers) are always just two bytes long.
119 nigel 75
120 ph10 212 A list of the opcodes follows:
121 nigel 75
122 nigel 41 Opcodes with no following data
123     ------------------------------
125     These items are all just one byte long
127     OP_END end of pattern
128 ph10 342 OP_ANY match any one character other than newline
129     OP_ALLANY match any one character, including newline
130 nigel 75 OP_ANYBYTE match any single byte, even in UTF-8 mode
131 nigel 41 OP_SOD match start of data: \A
132 nigel 71 OP_SOM, start of match (subject + offset): \G
133 ph10 181 OP_SET_SOM, set start of match (\K)
134 nigel 41 OP_CIRC ^ (start of data, or after \n in multiline)
137     OP_NOT_DIGIT \D
138     OP_DIGIT \d
139 ph10 181 OP_NOT_HSPACE \H
140     OP_HSPACE \h
141 nigel 41 OP_NOT_WHITESPACE \S
142     OP_WHITESPACE \s
143 ph10 181 OP_NOT_VSPACE \V
144     OP_VSPACE \v
145 nigel 41 OP_NOT_WORDCHAR \W
146     OP_WORDCHAR \w
147     OP_EODN match end of data or \n at end: \Z
148     OP_EOD match end of data: \z
149     OP_DOLL $ (end of data, or before \n in multiline)
150 nigel 75 OP_EXTUNI match an extended Unicode character
151 nigel 93 OP_ANYNL match any Unicode newline sequence
152 nigel 75
153 ph10 212 OP_ACCEPT )
154     OP_COMMIT )
155     OP_FAIL ) These are Perl 5.10's "backtracking
156     OP_PRUNE ) control verbs".
157     OP_SKIP )
158     OP_THEN )
160 nigel 41
161     Repeating single characters
162     ---------------------------
164 nigel 75 The common repeats (*, +, ?) when applied to a single character use the
165     following opcodes:
166 nigel 41
167     OP_STAR
168     OP_MINSTAR
169 nigel 93 OP_POSSTAR
170 nigel 41 OP_PLUS
171     OP_MINPLUS
172 nigel 93 OP_POSPLUS
173 nigel 41 OP_QUERY
175 nigel 93 OP_POSQUERY
176 nigel 41
177 nigel 75 In ASCII mode, these are two-byte items; in UTF-8 mode, the length is variable.
178 nigel 93 Those with "MIN" in their name are the minimizing versions. Those with "POS" in
179     their names are possessive versions. Each is followed by the character that is
180     to be repeated. Other repeats make use of
181 nigel 41
182     OP_UPTO
183     OP_MINUPTO
184 nigel 93 OP_POSUPTO
185 nigel 41 OP_EXACT
187     which are followed by a two-byte count (most significant first) and the
188     repeated character. OP_UPTO matches from 0 to the given number. A repeat with a
189     non-zero minimum and a fixed maximum is coded as an OP_EXACT followed by an
190 nigel 93 OP_UPTO (or OP_MINUPTO or OPT_POSUPTO).
191 nigel 41
193     Repeating character types
194     -------------------------
196     Repeats of things like \d are done exactly as for single characters, except
197     that instead of a character, the opcode for the type is stored in the data
198     byte. The opcodes are:
202 nigel 93 OP_TYPEPOSSTAR
203 nigel 41 OP_TYPEPLUS
205 nigel 93 OP_TYPEPOSPLUS
206 nigel 41 OP_TYPEQUERY
208 nigel 93 OP_TYPEPOSQUERY
209 nigel 41 OP_TYPEUPTO
211 nigel 93 OP_TYPEPOSUPTO
212 nigel 41 OP_TYPEEXACT
215 nigel 75 Match by Unicode property
216     -------------------------
218     OP_PROP and OP_NOTPROP are used for positive and negative matches of a
219     character by testing its Unicode property (the \p and \P escape sequences).
220 nigel 91 Each is followed by two bytes that encode the desired property as a type and a
221     value.
222 nigel 75
223 nigel 91 Repeats of these items use the OP_TYPESTAR etc. set of opcodes, followed by
224     three bytes: OP_PROP or OP_NOTPROP and then the desired property type and
225     value.
226 nigel 75
228     Matching literal characters
229 nigel 41 ---------------------------
231 nigel 75 The OP_CHAR opcode is followed by a single character that is to be matched
232     casefully. For caseless matching, OP_CHARNC is used. In UTF-8 mode, the
233     character may be more than one byte long. (Earlier versions of PCRE used
234     multi-character strings, but this was changed to allow some new features to be
235     added.)
236 nigel 41
238     Character classes
239     -----------------
241 nigel 75 If there is only one character, OP_CHAR or OP_CHARNC is used for a positive
242     class, and OP_NOT for a negative one (that is, for something like [^a]).
243     However, in UTF-8 mode, the use of OP_NOT applies only to characters with
244     values < 128, because OP_NOT is confined to single bytes.
245 nigel 41
246 nigel 63 Another set of repeating opcodes (OP_NOTSTAR etc.) are used for a repeated,
247     negated, single-character class. The normal ones (OP_STAR etc.) are used for a
248     repeated positive single-character class.
250 nigel 71 When there's more than one character in a class and all the characters are less
251 nigel 75 than 256, OP_CLASS is used for a positive class, and OP_NCLASS for a negative
252 nigel 71 one. In either case, the opcode is followed by a 32-byte bit map containing a 1
253     bit for every character that is acceptable. The bits are counted from the least
254     significant end of each byte.
255 nigel 41
256 nigel 75 The reason for having both OP_CLASS and OP_NCLASS is so that, in UTF-8 mode,
257     subject characters with values greater than 256 can be handled correctly. For
258 nigel 71 OP_CLASS they don't match, whereas for OP_NCLASS they do.
260 nigel 63 For classes containing characters with values > 255, OP_XCLASS is used. It
261     optionally uses a bit map (if any characters lie within it), followed by a list
262 nigel 75 of pairs and single characters. There is a flag character than indicates
263 nigel 63 whether it's a positive or a negative class.
264 nigel 41
265 nigel 63
266 nigel 41 Back references
267     ---------------
269 nigel 53 OP_REF is followed by two bytes containing the reference number.
270 nigel 41
272     Repeating character classes and back references
273     -----------------------------------------------
275 nigel 93 Single-character classes are handled specially (see above). This section
276     applies to OP_CLASS and OP_REF. In both cases, the repeat information follows
277     the base item. The matching code looks at the following opcode to see if it is
278     one of
279 nigel 41
280     OP_CRSTAR
282     OP_CRPLUS
284     OP_CRQUERY
286     OP_CRRANGE
289     All but the last two are just single-byte items. The others are followed by
290 nigel 93 four bytes of data, comprising the minimum and maximum repeat counts. There are
291     no special possessive opcodes for these repeats; a possessive repeat is
292     compiled into an atomic group.
293 nigel 41
295     Brackets and alternation
296     ------------------------
298 nigel 43 A pair of non-capturing (round) brackets is wrapped round each expression at
299 nigel 41 compile time, so alternation always happens in the context of brackets.
300 nigel 53
301 nigel 93 [Note for North Americans: "bracket" to some English speakers, including
302     myself, can be round, square, curly, or pointy. Hence this usage.]
303 nigel 41
304 nigel 93 Non-capturing brackets use the opcode OP_BRA. Originally PCRE was limited to 99
305     capturing brackets and it used a different opcode for each one. From release
306     3.5, the limit was removed by putting the bracket number into the data for
307     higher-numbered brackets. From release 7.0 all capturing brackets are handled
308     this way, using the single opcode OP_CBRA.
309 nigel 53
310 nigel 77 A bracket opcode is followed by LINK_SIZE bytes which give the offset to the
311     next alternative OP_ALT or, if there aren't any branches, to the matching
312     OP_KET opcode. Each OP_ALT is followed by LINK_SIZE bytes giving the offset to
313 nigel 93 the next one, or to the OP_KET opcode. For capturing brackets, the bracket
314     number immediately follows the offset, always as a 2-byte item.
315 nigel 41
316     OP_KET is used for subpatterns that do not repeat indefinitely, while
317     OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or
318 nigel 77 maximally respectively. All three are followed by LINK_SIZE bytes giving (as a
319 nigel 93 positive number) the offset back to the matching bracket opcode.
320 nigel 41
321     If a subpattern is quantified such that it is permitted to match zero times, it
322 ph10 335 is preceded by one of OP_BRAZERO, OP_BRAMINZERO, or OP_SKIPZERO. These are
323     single-byte opcodes that tell the matcher that skipping the following
324     subpattern entirely is a valid branch. In the case of the first two, not
325     skipping the pattern is also valid (greedy and non-greedy). The third is used
326     when a pattern has the quantifier {0,0}. It cannot be entirely discarded,
327     because it may be called as a subroutine from elsewhere in the regex.
328 nigel 41
329     A subpattern with an indefinite maximum repetition is replicated in the
330 nigel 75 compiled data its minimum number of times (or once with OP_BRAZERO if the
331     minimum is zero), with the final copy terminating with OP_KETRMIN or OP_KETRMAX
332     as appropriate.
333 nigel 41
334     A subpattern with a bounded maximum repetition is replicated in a nested
335 nigel 75 fashion up to the maximum number of times, with OP_BRAZERO or OP_BRAMINZERO
336     before each replication after the minimum, so that, for example, (abc){2,5} is
337 nigel 93 compiled as (abc)(abc)((abc)((abc)(abc)?)?)?, except that each bracketed group
338     has the same number.
339 nigel 41
340 nigel 93 When a repeated subpattern has an unbounded upper limit, it is checked to see
341     whether it could match an empty string. If this is the case, the opcode in the
342     final replication is changed to OP_SBRA or OP_SCBRA. This tells the matcher
343     that it needs to check for matching an empty string when it hits OP_KETRMIN or
344     OP_KETRMAX, and if so, to break the loop.
345 nigel 41
346 nigel 93
347 nigel 41 Assertions
348     ----------
350     Forward assertions are just like other subpatterns, but starting with one of
351     the opcodes OP_ASSERT or OP_ASSERT_NOT. Backward assertions use the opcodes
352     OP_ASSERTBACK and OP_ASSERTBACK_NOT, and the first opcode inside the assertion
353     is OP_REVERSE, followed by a two byte count of the number of characters to move
354 nigel 49 back the pointer in the subject string. When operating in UTF-8 mode, the count
355     is a character count rather than a byte count. A separate count is present in
356     each alternative of a lookbehind assertion, allowing them to have different
357     fixed lengths.
358 nigel 41
360 nigel 93 Once-only (atomic) subpatterns
361     ------------------------------
362 nigel 41
363     These are also just like other subpatterns, but they start with the opcode
364 nigel 93 OP_ONCE. The check for matching an empty string in an unbounded repeat is
365     handled entirely at runtime, so there is just this one opcode.
366 nigel 41
368     Conditional subpatterns
369     -----------------------
371 nigel 93 These are like other subpatterns, but they start with the opcode OP_COND, or
372     OP_SCOND for one that might match an empty string in an unbounded repeat. If
373 nigel 41 the condition is a back reference, this is stored at the start of the
374 nigel 53 subpattern using the opcode OP_CREF followed by two bytes containing the
375 nigel 93 reference number. If the condition is "in recursion" (coded as "(?(R)"), or "in
376     recursion of group x" (coded as "(?(Rx)"), the group number is stored at the
377     start of the subpattern using the opcode OP_RREF, and a value of zero for "the
378     whole pattern". For a DEFINE condition, just the single byte OP_DEF is used (it
379     has no associated data). Otherwise, a conditional subpattern always starts with
380     one of the assertions.
381 nigel 41
383 nigel 71 Recursion
384     ---------
386     Recursion either matches the current regex, or some subexpression. The opcode
387     OP_RECURSE is followed by an value which is the offset to the starting bracket
388 nigel 87 from the start of the whole pattern. From release 6.5, OP_RECURSE is
389     automatically wrapped inside OP_ONCE brackets (because otherwise some patterns
390     broke it). OP_RECURSE is also used for "subroutine" calls, even though they
391     are not strictly a recursion.
392 nigel 71
394     Callout
395     -------
397 nigel 75 OP_CALLOUT is followed by one byte of data that holds a callout number in the
398     range 0 to 254 for manual callouts, or 255 for an automatic callout. In both
399     cases there follows a two-byte value giving the offset in the pattern to the
400     start of the following item, and another two-byte item giving the length of the
401     next item.
402 nigel 71
404 nigel 41 Changing options
405     ----------------
407 nigel 63 If any of the /i, /m, or /s options are changed within a pattern, an OP_OPT
408     opcode is compiled, followed by one byte containing the new settings of these
409     flags. If there are several alternatives, there is an occurrence of OP_OPT at
410     the start of all those following the first options change, to set appropriate
411     options for the start of the alternative. Immediately after the end of the
412     group there is another such item to reset the flags to their previous values. A
413     change of flag right at the very start of the pattern can be handled entirely
414     at compile time, and so does not cause anything to be put into the compiled
415     data.
416 nigel 41
417     Philip Hazel
418 ph10 335 April 2008

ViewVC Help
Powered by ViewVC 1.1.12