/[pcre]/code/trunk/maintain/Tech.Notes
ViewVC logotype

Diff of /code/trunk/maintain/Tech.Notes

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 71 by nigel, Sat Feb 24 21:40:24 2007 UTC revision 91 by nigel, Sat Feb 24 21:41:34 2007 UTC
# Line 1  Line 1 
1  Technical Notes about PCRE  Technical Notes about PCRE
2  --------------------------  --------------------------
3    
4    These are very rough technical notes that record potentially useful information
5    about PCRE internals.
6    
7    Historical note 1
8    -----------------
9    
10  Many years ago I implemented some regular expression functions to an algorithm  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  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  restricted in what they could do by comparison with Perl. The interesting part
# Line 9  form of an expression was known in advan Line 15  form of an expression was known in advan
15  not operate by backtracking, as the original Henry Spencer code and current  not operate by backtracking, as the original Henry Spencer code and current
16  Perl code does, but instead checked all possibilities simultaneously by keeping  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  a list of current states and checking all of them as it advanced through the
18  subject string. (In the terminology of Jeffrey Friedl's book, it was a "DFA  subject string. In the terminology of Jeffrey Friedl's book, it was a "DFA
19  algorithm".) When the pattern was all used up, all remaining states were  algorithm". When the pattern was all used up, all remaining states were
20  possible matches, and the one matching the longest subset of the subject string  possible matches, and the one matching the longest subset of the subject string
21  was chosen. This did not necessarily maximize the individual wild portions of  was chosen. This did not necessarily maximize the individual wild portions of
22  the pattern, as is expected in Unix and Perl-style regular expressions.  the pattern, as is expected in Unix and Perl-style regular expressions.
23    
24  By contrast, the code originally written by Henry Spencer and subsequently  Historical note 2
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,  By contrast, the code originally written by Henry Spencer (which was
28  optionally, minimizing in Perl) the amount of the subject that matches  subsequently heavily modified for Perl) compiles the expression twice: once in
29  individual wild portions of the pattern. This is an "NFA algorithm" in Friedl's  a dummy mode in order to find out how much store will be needed, and then for
30  terminology.  real. (The Perl version probably doesn't do this any more; I'm talking about
31    the original library.) The execution function operates by backtracking and
32  For the set of functions that forms PCRE (which are unrelated to those  maximizing (or, optionally, minimizing in Perl) the amount of the subject that
33  mentioned above), I tried at first to invent an algorithm that used an amount  matches individual wild portions of the pattern. This is an "NFA algorithm" in
34  of store bounded by a multiple of the number of characters in the pattern, to  Friedl's terminology.
35  save on compiling time. However, because of the greater complexity in Perl  
36  regular expressions, I couldn't do this. In any case, a first pass through the  OK, here's the real stuff
37  pattern is needed, for a number of reasons. PCRE works by running a very  -------------------------
38  degenerate first pass to calculate a maximum store size, and then a second pass  
39  to do the real compile - which may use a bit less than the predicted amount of  For the set of functions that form the "basic" PCRE library (which are
40  store. The idea is that this is going to turn out faster because the first pass  unrelated to those mentioned above), I tried at first to invent an algorithm
41  is degenerate and the second pass can just store stuff straight into the  that used an amount of store bounded by a multiple of the number of characters
42  vector. It does make the compiling functions bigger, of course, but they have  in the pattern, to save on compiling time. However, because of the greater
43  got quite big anyway to handle all the Perl stuff.  complexity in Perl regular expressions, I couldn't do this. In any case, a
44    first pass through the pattern is needed, for a number of reasons. PCRE works
45    by running a very degenerate first pass to calculate a maximum store size, and
46    then a second pass to do the real compile - which may use a bit less than the
47    predicted amount of store. The idea is that this is going to turn out faster
48    because the first pass is degenerate and the second pass can just store stuff
49    straight into the vector, which it knows is big enough. It does make the
50    compiling functions bigger, of course, but they have become quite big anyway to
51    handle all the Perl stuff.
52    
53    Traditional matching function
54    -----------------------------
55    
56    The "traditional", and original, matching function is called pcre_exec(), and
57    it implements an NFA algorithm, similar to the original Henry Spencer algorithm
58    and the way that Perl works. Not surprising, since it is intended to be as
59    compatible with Perl as possible. This is the function most users of PCRE will
60    use most of the time.
61    
62    Supplementary matching function
63    -------------------------------
64    
65    From PCRE 6.0, there is also a supplementary matching function called
66    pcre_dfa_exec(). This implements a DFA matching algorithm that searches
67    simultaneously for all possible matches that start at one point in the subject
68    string. (Going back to my roots: see Historical Note 1 above.) This function
69    intreprets the same compiled pattern data as pcre_exec(); however, not all the
70    facilities are available, and those that are do not always work in quite the
71    same way. See the user documentation for details.
72    
73    Format of compiled patterns
74    ---------------------------
75    
76  The compiled form of a pattern is a vector of bytes, containing items of  The compiled form of a pattern is a vector of bytes, containing items of
77  variable length. The first byte in an item is an opcode, and the length of the  variable length. The first byte in an item is an opcode, and the length of the
78  item is either implicit in the opcode or contained in the data bytes which  item is either implicit in the opcode or contained in the data bytes that
79  follow it. A list of all the opcodes follows:  follow it.
80    
81    In many cases below "two-byte" data values are specified. This is in fact just
82    a default. PCRE can be compiled to use 3-byte or 4-byte values (impairing the
83    performance). This is necessary only when patterns whose compiled length is
84    greater than 64K are going to be processed. In this description, we assume the
85    "normal" compilation options.
86    
87    A list of all the opcodes follows:
88    
89  Opcodes with no following data  Opcodes with no following data
90  ------------------------------  ------------------------------
# Line 48  These items are all just one byte long Line 93  These items are all just one byte long
93    
94    OP_END                 end of pattern    OP_END                 end of pattern
95    OP_ANY                 match any character    OP_ANY                 match any character
96    OP_ANYBYTE             match any single byte, even in UTF-8 mode    OP_ANYBYTE             match any single byte, even in UTF-8 mode
97    OP_SOD                 match start of data: \A    OP_SOD                 match start of data: \A
98    OP_SOM,                start of match (subject + offset): \G    OP_SOM,                start of match (subject + offset): \G
99    OP_CIRC                ^ (start of data, or after \n in multiline)    OP_CIRC                ^ (start of data, or after \n in multiline)
# Line 63  These items are all just one byte long Line 108  These items are all just one byte long
108    OP_EODN                match end of data or \n at end: \Z    OP_EODN                match end of data or \n at end: \Z
109    OP_EOD                 match end of data: \z    OP_EOD                 match end of data: \z
110    OP_DOLL                $ (end of data, or before \n in multiline)    OP_DOLL                $ (end of data, or before \n in multiline)
111      OP_EXTUNI              match an extended Unicode character
112    
113    
114  Repeating single characters  Repeating single characters
115  ---------------------------  ---------------------------
116    
117  The common repeats (*, +, ?) when applied to a single character appear as  The common repeats (*, +, ?) when applied to a single character use the
118  two-byte items using the following opcodes:  following opcodes:
119    
120    OP_STAR    OP_STAR
121    OP_MINSTAR    OP_MINSTAR
# Line 78  two-byte items using the following opcod Line 124  two-byte items using the following opcod
124    OP_QUERY    OP_QUERY
125    OP_MINQUERY    OP_MINQUERY
126    
127    In ASCII mode, these are two-byte items; in UTF-8 mode, the length is variable.
128  Those with "MIN" in their name are the minimizing versions. Each is followed by  Those with "MIN" in their name are the minimizing versions. Each is followed by
129  the character that is to be repeated. Other repeats make use of  the character that is to be repeated. Other repeats make use of
130    
# Line 109  byte. The opcodes are: Line 156  byte. The opcodes are:
156    OP_TYPEEXACT    OP_TYPEEXACT
157    
158    
159  Matching a character string  Match by Unicode property
160    -------------------------
161    
162    OP_PROP and OP_NOTPROP are used for positive and negative matches of a
163    character by testing its Unicode property (the \p and \P escape sequences).
164    Each is followed by two bytes that encode the desired property as a type and a
165    value.
166    
167    Repeats of these items use the OP_TYPESTAR etc. set of opcodes, followed by
168    three bytes: OP_PROP or OP_NOTPROP and then the desired property type and
169    value.
170    
171    
172    Matching literal characters
173  ---------------------------  ---------------------------
174    
175  The OP_CHARS opcode is followed by a one-byte count and then that number of  The OP_CHAR opcode is followed by a single character that is to be matched
176  characters. If there are more than 255 characters in sequence, successive  casefully. For caseless matching, OP_CHARNC is used. In UTF-8 mode, the
177  instances of OP_CHARS are used.  character may be more than one byte long. (Earlier versions of PCRE used
178    multi-character strings, but this was changed to allow some new features to be
179    added.)
180    
181    
182  Character classes  Character classes
183  -----------------  -----------------
184    
185  If there is only one character, OP_CHARS is used for a positive class,  If there is only one character, OP_CHAR or OP_CHARNC is used for a positive
186  and OP_NOT for a negative one (that is, for something like [^a]). However, in  class, and OP_NOT for a negative one (that is, for something like [^a]).
187  UTF-8 mode, this applies only to characters with values < 128, because OP_NOT  However, in UTF-8 mode, the use of OP_NOT applies only to characters with
188  is confined to single bytes.  values < 128, because OP_NOT is confined to single bytes.
189    
190  Another set of repeating opcodes (OP_NOTSTAR etc.) are used for a repeated,  Another set of repeating opcodes (OP_NOTSTAR etc.) are used for a repeated,
191  negated, single-character class. The normal ones (OP_STAR etc.) are used for a  negated, single-character class. The normal ones (OP_STAR etc.) are used for a
192  repeated positive single-character class.  repeated positive single-character class.
193    
194  When there's more than one character in a class and all the characters are less  When there's more than one character in a class and all the characters are less
195  than 256, OP_CLASS is used for a positive class, and OP_NCLASS for a negative  than 256, OP_CLASS is used for a positive class, and OP_NCLASS for a negative
196  one. In either case, the opcode is followed by a 32-byte bit map containing a 1  one. In either case, the opcode is followed by a 32-byte bit map containing a 1
197  bit for every character that is acceptable. The bits are counted from the least  bit for every character that is acceptable. The bits are counted from the least
198  significant end of each byte.  significant end of each byte.
199    
200  The reason for having both OP_CLASS and OP_NCLASS is so that, in UTF-8 mode,  The reason for having both OP_CLASS and OP_NCLASS is so that, in UTF-8 mode,
201  subject characters with values greater than 256 can be handled correctly. For  subject characters with values greater than 256 can be handled correctly. For
202  OP_CLASS they don't match, whereas for OP_NCLASS they do.  OP_CLASS they don't match, whereas for OP_NCLASS they do.
203    
204  For classes containing characters with values > 255, OP_XCLASS is used. It  For classes containing characters with values > 255, OP_XCLASS is used. It
205  optionally uses a bit map (if any characters lie within it), followed by a list  optionally uses a bit map (if any characters lie within it), followed by a list
206  of pairs and single characters. There is a flag character than indicates  of pairs and single characters. There is a flag character than indicates
207  whether it's a positive or a negative class.  whether it's a positive or a negative class.
208    
209    
# Line 191  number. This opcode is ignored while mat Line 253  number. This opcode is ignored while mat
253  the bracket itself. (They could have all been done like this, but I was making  the bracket itself. (They could have all been done like this, but I was making
254  minimal changes.)  minimal changes.)
255    
256  A bracket opcode is followed by two bytes which give the offset to the next  A bracket opcode is followed by LINK_SIZE bytes which give the offset to the
257  alternative OP_ALT or, if there aren't any branches, to the matching KET  next alternative OP_ALT or, if there aren't any branches, to the matching
258  opcode. Each OP_ALT is followed by two bytes giving the offset to the next one,  OP_KET opcode. Each OP_ALT is followed by LINK_SIZE bytes giving the offset to
259  or to the KET opcode.  the next one, or to the OP_KET opcode.
260    
261  OP_KET is used for subpatterns that do not repeat indefinitely, while  OP_KET is used for subpatterns that do not repeat indefinitely, while
262  OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or  OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or
263  maximally respectively. All three are followed by two bytes giving (as a  maximally respectively. All three are followed by LINK_SIZE bytes giving (as a
264  positive number) the offset back to the matching BRA opcode.  positive number) the offset back to the matching OP_BRA opcode.
265    
266  If a subpattern is quantified such that it is permitted to match zero times, it  If a subpattern is quantified such that it is permitted to match zero times, it
267  is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte  is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte
# Line 207  opcodes which tell the matcher that skip Line 269  opcodes which tell the matcher that skip
269  valid branch.  valid branch.
270    
271  A subpattern with an indefinite maximum repetition is replicated in the  A subpattern with an indefinite maximum repetition is replicated in the
272  compiled data its minimum number of times (or once with a BRAZERO if the  compiled data its minimum number of times (or once with OP_BRAZERO if the
273  minimum is zero), with the final copy terminating with a KETRMIN or KETRMAX as  minimum is zero), with the final copy terminating with OP_KETRMIN or OP_KETRMAX
274  appropriate.  as appropriate.
275    
276  A subpattern with a bounded maximum repetition is replicated in a nested  A subpattern with a bounded maximum repetition is replicated in a nested
277  fashion up to the maximum number of times, with BRAZERO or BRAMINZERO before  fashion up to the maximum number of times, with OP_BRAZERO or OP_BRAMINZERO
278  each replication after the minimum, so that, for example, (abc){2,5} is  before each replication after the minimum, so that, for example, (abc){2,5} is
279  compiled as (abc)(abc)((abc)((abc)(abc)?)?)?. The 99 and 200 bracket limits do  compiled as (abc)(abc)((abc)((abc)(abc)?)?)?.
 not apply to these internally generated brackets.  
280    
281    
282  Assertions  Assertions
# Line 254  Recursion Line 315  Recursion
315    
316  Recursion either matches the current regex, or some subexpression. The opcode  Recursion either matches the current regex, or some subexpression. The opcode
317  OP_RECURSE is followed by an value which is the offset to the starting bracket  OP_RECURSE is followed by an value which is the offset to the starting bracket
318  from the start of the whole pattern.  from the start of the whole pattern. From release 6.5, OP_RECURSE is
319    automatically wrapped inside OP_ONCE brackets (because otherwise some patterns
320    broke it). OP_RECURSE is also used for "subroutine" calls, even though they
321    are not strictly a recursion.
322    
323    
324  Callout  Callout
325  -------  -------
326    
327  OP_CALLOUT is followed by one byte of data that holds a callout number in the  OP_CALLOUT is followed by one byte of data that holds a callout number in the
328  range 0 to 255.  range 0 to 254 for manual callouts, or 255 for an automatic callout. In both
329    cases there follows a two-byte value giving the offset in the pattern to the
330    start of the following item, and another two-byte item giving the length of the
331    next item.
332    
333    
334  Changing options  Changing options
# Line 278  at compile time, and so does not cause a Line 345  at compile time, and so does not cause a
345  data.  data.
346    
347  Philip Hazel  Philip Hazel
348  August 2003  June 2006

Legend:
Removed from v.71  
changed lines
  Added in v.91

  ViewVC Help
Powered by ViewVC 1.1.5