ViewVC logotype

Contents of /code/trunk/HACKING

Parent Directory Parent Directory | Revision Log Revision Log

Revision 87 - (hide annotations) (download)
Sat Feb 24 21:41:21 2007 UTC (8 years, 1 month ago) by nigel
Original Path: code/trunk/doc/Tech.Notes
File size: 14088 byte(s)
Load pcre-6.5 into code/trunk.

1 nigel 41 Technical Notes about PCRE
2     --------------------------
4 nigel 75 Historical note 1
5     -----------------
7 nigel 41 Many years ago I implemented some regular expression functions to an algorithm
8     suggested by Martin Richards. These were not Unix-like in form, and were quite
9     restricted in what they could do by comparison with Perl. The interesting part
10     about the algorithm was that the amount of space required to hold the compiled
11     form of an expression was known in advance. The code to apply an expression did
12 nigel 63 not operate by backtracking, as the original Henry Spencer code and current
13     Perl code does, but instead checked all possibilities simultaneously by keeping
14     a list of current states and checking all of them as it advanced through the
15 nigel 75 subject string. In the terminology of Jeffrey Friedl's book, it was a "DFA
16     algorithm". When the pattern was all used up, all remaining states were
17 nigel 63 possible matches, and the one matching the longest subset of the subject string
18     was chosen. This did not necessarily maximize the individual wild portions of
19     the pattern, as is expected in Unix and Perl-style regular expressions.
20 nigel 41
21 nigel 75 Historical note 2
22     -----------------
24 nigel 41 By contrast, the code originally written by Henry Spencer and subsequently
25     heavily modified for Perl actually compiles the expression twice: once in a
26     dummy mode in order to find out how much store will be needed, and then for
27     real. The execution function operates by backtracking and maximizing (or,
28     optionally, minimizing in Perl) the amount of the subject that matches
29     individual wild portions of the pattern. This is an "NFA algorithm" in Friedl's
30     terminology.
32 nigel 75 OK, here's the real stuff
33     -------------------------
35 nigel 77 For the set of functions that form the "basic" PCRE library (which are
36     unrelated to those mentioned above), I tried at first to invent an algorithm
37     that used an amount of store bounded by a multiple of the number of characters
38     in the pattern, to save on compiling time. However, because of the greater
39     complexity in Perl regular expressions, I couldn't do this. In any case, a
40     first pass through the pattern is needed, for a number of reasons. PCRE works
41     by running a very degenerate first pass to calculate a maximum store size, and
42     then a second pass to do the real compile - which may use a bit less than the
43     predicted amount of store. The idea is that this is going to turn out faster
44     because the first pass is degenerate and the second pass can just store stuff
45     straight into the vector, which it knows is big enough. It does make the
46     compiling functions bigger, of course, but they have got quite big anyway to
47     handle all the Perl stuff.
48 nigel 41
49 nigel 77 Traditional matching function
50     -----------------------------
52     The "traditional", and original, matching function is called pcre_exec(), and
53     it implements an NFA algorithm, similar to the original Henry Spencer algorithm
54     and the way that Perl works. Not surprising, since it is intended to be as
55     compatible with Perl as possible. This is the function most users of PCRE will
56     use most of the time.
58     Supplementary matching function
59     -------------------------------
61     From PCRE 6.0, there is also a supplementary matching function called
62     pcre_dfa_exec(). This implements a DFA matching algorithm that searches
63     simultaneously for all possible matches that start at one point in the subject
64     string. (Going back to my roots: see Historical Note 1 above.) This function
65     intreprets the same compiled pattern data as pcre_exec(); however, not all the
66     facilities are available, and those that are don't always work in quite the
67     same way. See the user documentation for details.
69     Format of compiled patterns
70     ---------------------------
72 nigel 41 The compiled form of a pattern is a vector of bytes, containing items of
73     variable length. The first byte in an item is an opcode, and the length of the
74 nigel 75 item is either implicit in the opcode or contained in the data bytes that
75     follow it.
76 nigel 41
77 nigel 75 In many cases below "two-byte" data values are specified. This is in fact just
78     a default. PCRE can be compiled to use 3-byte or 4-byte values (impairing the
79     performance). This is necessary only when patterns whose compiled length is
80     greater than 64K are going to be processed. In this description, we assume the
81     "normal" compilation options.
83     A list of all the opcodes follows:
85 nigel 41 Opcodes with no following data
86     ------------------------------
88     These items are all just one byte long
90     OP_END end of pattern
91     OP_ANY match any character
92 nigel 75 OP_ANYBYTE match any single byte, even in UTF-8 mode
93 nigel 41 OP_SOD match start of data: \A
94 nigel 71 OP_SOM, start of match (subject + offset): \G
95 nigel 41 OP_CIRC ^ (start of data, or after \n in multiline)
98     OP_NOT_DIGIT \D
99     OP_DIGIT \d
101     OP_WHITESPACE \s
103     OP_WORDCHAR \w
104     OP_EODN match end of data or \n at end: \Z
105     OP_EOD match end of data: \z
106     OP_DOLL $ (end of data, or before \n in multiline)
107 nigel 75 OP_EXTUNI match an extended Unicode character
109 nigel 41
110     Repeating single characters
111     ---------------------------
113 nigel 75 The common repeats (*, +, ?) when applied to a single character use the
114     following opcodes:
115 nigel 41
116     OP_STAR
117     OP_MINSTAR
118     OP_PLUS
119     OP_MINPLUS
120     OP_QUERY
123 nigel 75 In ASCII mode, these are two-byte items; in UTF-8 mode, the length is variable.
124 nigel 41 Those with "MIN" in their name are the minimizing versions. Each is followed by
125     the character that is to be repeated. Other repeats make use of
127     OP_UPTO
128     OP_MINUPTO
129     OP_EXACT
131     which are followed by a two-byte count (most significant first) and the
132     repeated character. OP_UPTO matches from 0 to the given number. A repeat with a
133     non-zero minimum and a fixed maximum is coded as an OP_EXACT followed by an
134     OP_UPTO (or OP_MINUPTO).
137     Repeating character types
138     -------------------------
140     Repeats of things like \d are done exactly as for single characters, except
141     that instead of a character, the opcode for the type is stored in the data
142     byte. The opcodes are:
155 nigel 75 Match by Unicode property
156     -------------------------
158     OP_PROP and OP_NOTPROP are used for positive and negative matches of a
159     character by testing its Unicode property (the \p and \P escape sequences).
160     Each is followed by a single byte that encodes the desired property value.
162     Repeats of these items use the OP_TYPESTAR etc. set of opcodes, followed by two
163     bytes: OP_PROP or OP_NOTPROP and then the desired property value.
166     Matching literal characters
167 nigel 41 ---------------------------
169 nigel 75 The OP_CHAR opcode is followed by a single character that is to be matched
170     casefully. For caseless matching, OP_CHARNC is used. In UTF-8 mode, the
171     character may be more than one byte long. (Earlier versions of PCRE used
172     multi-character strings, but this was changed to allow some new features to be
173     added.)
174 nigel 41
176     Character classes
177     -----------------
179 nigel 75 If there is only one character, OP_CHAR or OP_CHARNC is used for a positive
180     class, and OP_NOT for a negative one (that is, for something like [^a]).
181     However, in UTF-8 mode, the use of OP_NOT applies only to characters with
182     values < 128, because OP_NOT is confined to single bytes.
183 nigel 41
184 nigel 63 Another set of repeating opcodes (OP_NOTSTAR etc.) are used for a repeated,
185     negated, single-character class. The normal ones (OP_STAR etc.) are used for a
186     repeated positive single-character class.
188 nigel 71 When there's more than one character in a class and all the characters are less
189 nigel 75 than 256, OP_CLASS is used for a positive class, and OP_NCLASS for a negative
190 nigel 71 one. In either case, the opcode is followed by a 32-byte bit map containing a 1
191     bit for every character that is acceptable. The bits are counted from the least
192     significant end of each byte.
193 nigel 41
194 nigel 75 The reason for having both OP_CLASS and OP_NCLASS is so that, in UTF-8 mode,
195     subject characters with values greater than 256 can be handled correctly. For
196 nigel 71 OP_CLASS they don't match, whereas for OP_NCLASS they do.
198 nigel 63 For classes containing characters with values > 255, OP_XCLASS is used. It
199     optionally uses a bit map (if any characters lie within it), followed by a list
200 nigel 75 of pairs and single characters. There is a flag character than indicates
201 nigel 63 whether it's a positive or a negative class.
202 nigel 41
203 nigel 63
204 nigel 41 Back references
205     ---------------
207 nigel 53 OP_REF is followed by two bytes containing the reference number.
208 nigel 41
210     Repeating character classes and back references
211     -----------------------------------------------
213     Single-character classes are handled specially (see above). This applies to
214     OP_CLASS and OP_REF. In both cases, the repeat information follows the base
215     item. The matching code looks at the following opcode to see if it is one of
217     OP_CRSTAR
219     OP_CRPLUS
221     OP_CRQUERY
223     OP_CRRANGE
226     All but the last two are just single-byte items. The others are followed by
227     four bytes of data, comprising the minimum and maximum repeat counts.
230     Brackets and alternation
231     ------------------------
233 nigel 43 A pair of non-capturing (round) brackets is wrapped round each expression at
234 nigel 41 compile time, so alternation always happens in the context of brackets.
235 nigel 53
236 nigel 43 Non-capturing brackets use the opcode OP_BRA, while capturing brackets use
237 nigel 41 OP_BRA+1, OP_BRA+2, etc. [Note for North Americans: "bracket" to some English
238 nigel 43 speakers, including myself, can be round, square, curly, or pointy. Hence this
239     usage.]
240 nigel 41
241 nigel 53 Originally PCRE was limited to 99 capturing brackets (so as not to use up all
242     the opcodes). From release 3.5, there is no limit. What happens is that the
243     first ones, up to EXTRACT_BASIC_MAX are handled with separate opcodes, as
244     above. If there are more, the opcode is set to EXTRACT_BASIC_MAX+1, and the
245     first operation in the bracket is OP_BRANUMBER, followed by a 2-byte bracket
246     number. This opcode is ignored while matching, but is fished out when handling
247     the bracket itself. (They could have all been done like this, but I was making
248     minimal changes.)
250 nigel 77 A bracket opcode is followed by LINK_SIZE bytes which give the offset to the
251     next alternative OP_ALT or, if there aren't any branches, to the matching
252     OP_KET opcode. Each OP_ALT is followed by LINK_SIZE bytes giving the offset to
253     the next one, or to the OP_KET opcode.
254 nigel 41
255     OP_KET is used for subpatterns that do not repeat indefinitely, while
256     OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or
257 nigel 77 maximally respectively. All three are followed by LINK_SIZE bytes giving (as a
258 nigel 75 positive number) the offset back to the matching OP_BRA opcode.
259 nigel 41
260     If a subpattern is quantified such that it is permitted to match zero times, it
261     is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte
262     opcodes which tell the matcher that skipping this subpattern entirely is a
263     valid branch.
265     A subpattern with an indefinite maximum repetition is replicated in the
266 nigel 75 compiled data its minimum number of times (or once with OP_BRAZERO if the
267     minimum is zero), with the final copy terminating with OP_KETRMIN or OP_KETRMAX
268     as appropriate.
269 nigel 41
270     A subpattern with a bounded maximum repetition is replicated in a nested
271 nigel 75 fashion up to the maximum number of times, with OP_BRAZERO or OP_BRAMINZERO
272     before each replication after the minimum, so that, for example, (abc){2,5} is
273     compiled as (abc)(abc)((abc)((abc)(abc)?)?)?.
274 nigel 41
276     Assertions
277     ----------
279     Forward assertions are just like other subpatterns, but starting with one of
280     the opcodes OP_ASSERT or OP_ASSERT_NOT. Backward assertions use the opcodes
281     OP_ASSERTBACK and OP_ASSERTBACK_NOT, and the first opcode inside the assertion
282     is OP_REVERSE, followed by a two byte count of the number of characters to move
283 nigel 49 back the pointer in the subject string. When operating in UTF-8 mode, the count
284     is a character count rather than a byte count. A separate count is present in
285     each alternative of a lookbehind assertion, allowing them to have different
286     fixed lengths.
287 nigel 41
289     Once-only subpatterns
290     ---------------------
292     These are also just like other subpatterns, but they start with the opcode
293     OP_ONCE.
296     Conditional subpatterns
297     -----------------------
299     These are like other subpatterns, but they start with the opcode OP_COND. If
300     the condition is a back reference, this is stored at the start of the
301 nigel 53 subpattern using the opcode OP_CREF followed by two bytes containing the
302 nigel 63 reference number. If the condition is "in recursion" (coded as "(?(R)"), the
303     same scheme is used, with a "reference number" of 0xffff. Otherwise, a
304     conditional subpattern always starts with one of the assertions.
305 nigel 41
307 nigel 71 Recursion
308     ---------
310     Recursion either matches the current regex, or some subexpression. The opcode
311     OP_RECURSE is followed by an value which is the offset to the starting bracket
312 nigel 87 from the start of the whole pattern. From release 6.5, OP_RECURSE is
313     automatically wrapped inside OP_ONCE brackets (because otherwise some patterns
314     broke it). OP_RECURSE is also used for "subroutine" calls, even though they
315     are not strictly a recursion.
316 nigel 71
318     Callout
319     -------
321 nigel 75 OP_CALLOUT is followed by one byte of data that holds a callout number in the
322     range 0 to 254 for manual callouts, or 255 for an automatic callout. In both
323     cases there follows a two-byte value giving the offset in the pattern to the
324     start of the following item, and another two-byte item giving the length of the
325     next item.
326 nigel 71
328 nigel 41 Changing options
329     ----------------
331 nigel 63 If any of the /i, /m, or /s options are changed within a pattern, an OP_OPT
332     opcode is compiled, followed by one byte containing the new settings of these
333     flags. If there are several alternatives, there is an occurrence of OP_OPT at
334     the start of all those following the first options change, to set appropriate
335     options for the start of the alternative. Immediately after the end of the
336     group there is another such item to reset the flags to their previous values. A
337     change of flag right at the very start of the pattern can be handled entirely
338     at compile time, and so does not cause anything to be put into the compiled
339     data.
340 nigel 41
341     Philip Hazel
342 nigel 87 January 2006

ViewVC Help
Powered by ViewVC 1.1.12