trunk/doc/Tech.Notes

Technical Notes about PCRE
--------------------------

These are very rough technical notes that record potentially useful information 
about PCRE internals.

Historical note 1
-----------------

Many years ago I implemented some regular expression functions to an algorithm
suggested by Martin Richards. These were not Unix-like in form, and were quite
restricted in what they could do by comparison with Perl. The interesting part
about the algorithm was that the amount of space required to hold the compiled
form of an expression was known in advance. The code to apply an expression did
not operate by backtracking, as the original Henry Spencer code and current
Perl code does, but instead checked all possibilities simultaneously by keeping
a list of current states and checking all of them as it advanced through the
subject string. In the terminology of Jeffrey Friedl's book, it was a "DFA
algorithm". When the pattern was all used up, all remaining states were
possible matches, and the one matching the longest subset of the subject string
was chosen. This did not necessarily maximize the individual wild portions of
the pattern, as is expected in Unix and Perl-style regular expressions.

Historical note 2
-----------------

By contrast, the code originally written by Henry Spencer (which was
subsequently heavily modified for Perl) compiles the expression twice: once in
a dummy mode in order to find out how much store will be needed, and then for
real. (The Perl version probably doesn't do this any more; I'm talking about
the original library.) The execution function operates by backtracking and
maximizing (or, optionally, minimizing in Perl) the amount of the subject that
matches individual wild portions of the pattern. This is an "NFA algorithm" in
Friedl's terminology.

OK, here's the real stuff
-------------------------

For the set of functions that form the "basic" PCRE library (which are
unrelated to those mentioned above), I tried at first to invent an algorithm
that used an amount of store bounded by a multiple of the number of characters
in the pattern, to save on compiling time. However, because of the greater
complexity in Perl regular expressions, I couldn't do this. In any case, a
first pass through the pattern is needed, for a number of reasons. PCRE works
by running a very degenerate first pass to calculate a maximum store size, and
then a second pass to do the real compile - which may use a bit less than the
predicted amount of store. The idea is that this is going to turn out faster
because the first pass is degenerate and the second pass can just store stuff
straight into the vector, which it knows is big enough. It does make the
compiling functions bigger, of course, but they have become quite big anyway to
handle all the Perl stuff.

Traditional matching function
-----------------------------

The "traditional", and original, matching function is called pcre_exec(), and 
it implements an NFA algorithm, similar to the original Henry Spencer algorithm 
and the way that Perl works. Not surprising, since it is intended to be as 
compatible with Perl as possible. This is the function most users of PCRE will 
use most of the time.

Supplementary matching function
-------------------------------

From PCRE 6.0, there is also a supplementary matching function called 
pcre_dfa_exec(). This implements a DFA matching algorithm that searches 
simultaneously for all possible matches that start at one point in the subject 
string. (Going back to my roots: see Historical Note 1 above.) This function 
intreprets the same compiled pattern data as pcre_exec(); however, not all the 
facilities are available, and those that are do not always work in quite the 
same way. See the user documentation for details.

Format of compiled patterns
---------------------------

The compiled form of a pattern is a vector of bytes, containing items of
variable length. The first byte in an item is an opcode, and the length of the
item is either implicit in the opcode or contained in the data bytes that
follow it. 

In many cases below "two-byte" data values are specified. This is in fact just
a default. PCRE can be compiled to use 3-byte or 4-byte values (impairing the
performance). This is necessary only when patterns whose compiled length is
greater than 64K are going to be processed. In this description, we assume the 
"normal" compilation options.

A list of all the opcodes follows:

Opcodes with no following data
------------------------------

These items are all just one byte long

  OP_END                 end of pattern
  OP_ANY                 match any character
  OP_ANYBYTE             match any single byte, even in UTF-8 mode
  OP_SOD                 match start of data: \A
  OP_SOM,                start of match (subject + offset): \G
  OP_CIRC                ^ (start of data, or after \n in multiline)
  OP_NOT_WORD_BOUNDARY   \W
  OP_WORD_BOUNDARY       \w
  OP_NOT_DIGIT           \D
  OP_DIGIT               \d
  OP_NOT_WHITESPACE      \S
  OP_WHITESPACE          \s
  OP_NOT_WORDCHAR        \W
  OP_WORDCHAR            \w
  OP_EODN                match end of data or \n at end: \Z
  OP_EOD                 match end of data: \z
  OP_DOLL                $ (end of data, or before \n in multiline)
  OP_EXTUNI              match an extended Unicode character 
  

Repeating single characters
---------------------------

The common repeats (*, +, ?) when applied to a single character use the
following opcodes:

  OP_STAR
  OP_MINSTAR
  OP_PLUS
  OP_MINPLUS
  OP_QUERY
  OP_MINQUERY

In ASCII mode, these are two-byte items; in UTF-8 mode, the length is variable.
Those with "MIN" in their name are the minimizing versions. Each is followed by
the character that is to be repeated. Other repeats make use of

  OP_UPTO
  OP_MINUPTO
  OP_EXACT

which are followed by a two-byte count (most significant first) and the
repeated character. OP_UPTO matches from 0 to the given number. A repeat with a
non-zero minimum and a fixed maximum is coded as an OP_EXACT followed by an
OP_UPTO (or OP_MINUPTO).


Repeating character types
-------------------------

Repeats of things like \d are done exactly as for single characters, except
that instead of a character, the opcode for the type is stored in the data
byte. The opcodes are:

  OP_TYPESTAR
  OP_TYPEMINSTAR
  OP_TYPEPLUS
  OP_TYPEMINPLUS
  OP_TYPEQUERY
  OP_TYPEMINQUERY
  OP_TYPEUPTO
  OP_TYPEMINUPTO
  OP_TYPEEXACT


Match by Unicode property
-------------------------

OP_PROP and OP_NOTPROP are used for positive and negative matches of a 
character by testing its Unicode property (the \p and \P escape sequences).
Each is followed by two bytes that encode the desired property as a type and a 
value.

Repeats of these items use the OP_TYPESTAR etc. set of opcodes, followed by 
three bytes: OP_PROP or OP_NOTPROP and then the desired property type and 
value.


Matching literal characters
---------------------------

The OP_CHAR opcode is followed by a single character that is to be matched 
casefully. For caseless matching, OP_CHARNC is used. In UTF-8 mode, the 
character may be more than one byte long. (Earlier versions of PCRE used 
multi-character strings, but this was changed to allow some new features to be 
added.)


Character classes
-----------------

If there is only one character, OP_CHAR or OP_CHARNC is used for a positive
class, and OP_NOT for a negative one (that is, for something like [^a]).
However, in UTF-8 mode, the use of OP_NOT applies only to characters with
values < 128, because OP_NOT is confined to single bytes.

Another set of repeating opcodes (OP_NOTSTAR etc.) are used for a repeated,
negated, single-character class. The normal ones (OP_STAR etc.) are used for a
repeated positive single-character class.

When there's more than one character in a class and all the characters are less
than 256, OP_CLASS is used for a positive class, and OP_NCLASS for a negative
one. In either case, the opcode is followed by a 32-byte bit map containing a 1
bit for every character that is acceptable. The bits are counted from the least
significant end of each byte.

The reason for having both OP_CLASS and OP_NCLASS is so that, in UTF-8 mode,
subject characters with values greater than 256 can be handled correctly. For
OP_CLASS they don't match, whereas for OP_NCLASS they do.

For classes containing characters with values > 255, OP_XCLASS is used. It
optionally uses a bit map (if any characters lie within it), followed by a list
of pairs and single characters. There is a flag character than indicates
whether it's a positive or a negative class.


Back references
---------------

OP_REF is followed by two bytes containing the reference number.


Repeating character classes and back references
-----------------------------------------------

Single-character classes are handled specially (see above). This applies to
OP_CLASS and OP_REF. In both cases, the repeat information follows the base
item. The matching code looks at the following opcode to see if it is one of

  OP_CRSTAR
  OP_CRMINSTAR
  OP_CRPLUS
  OP_CRMINPLUS
  OP_CRQUERY
  OP_CRMINQUERY
  OP_CRRANGE
  OP_CRMINRANGE

All but the last two are just single-byte items. The others are followed by
four bytes of data, comprising the minimum and maximum repeat counts.


Brackets and alternation
------------------------

A pair of non-capturing (round) brackets is wrapped round each expression at
compile time, so alternation always happens in the context of brackets.

Non-capturing brackets use the opcode OP_BRA, while capturing brackets use
OP_BRA+1, OP_BRA+2, etc. [Note for North Americans: "bracket" to some English
speakers, including myself, can be round, square, curly, or pointy. Hence this
usage.]

Originally PCRE was limited to 99 capturing brackets (so as not to use up all
the opcodes). From release 3.5, there is no limit. What happens is that the
first ones, up to EXTRACT_BASIC_MAX are handled with separate opcodes, as
above. If there are more, the opcode is set to EXTRACT_BASIC_MAX+1, and the
first operation in the bracket is OP_BRANUMBER, followed by a 2-byte bracket
number. This opcode is ignored while matching, but is fished out when handling
the bracket itself. (They could have all been done like this, but I was making
minimal changes.)

A bracket opcode is followed by LINK_SIZE bytes which give the offset to the
next alternative OP_ALT or, if there aren't any branches, to the matching
OP_KET opcode. Each OP_ALT is followed by LINK_SIZE bytes giving the offset to
the next one, or to the OP_KET opcode.

OP_KET is used for subpatterns that do not repeat indefinitely, while
OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or
maximally respectively. All three are followed by LINK_SIZE bytes giving (as a
positive number) the offset back to the matching OP_BRA opcode.

If a subpattern is quantified such that it is permitted to match zero times, it
is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte
opcodes which tell the matcher that skipping this subpattern entirely is a
valid branch.

A subpattern with an indefinite maximum repetition is replicated in the
compiled data its minimum number of times (or once with OP_BRAZERO if the
minimum is zero), with the final copy terminating with OP_KETRMIN or OP_KETRMAX
as appropriate.

A subpattern with a bounded maximum repetition is replicated in a nested
fashion up to the maximum number of times, with OP_BRAZERO or OP_BRAMINZERO
before each replication after the minimum, so that, for example, (abc){2,5} is
compiled as (abc)(abc)((abc)((abc)(abc)?)?)?.


Assertions
----------

Forward assertions are just like other subpatterns, but starting with one of
the opcodes OP_ASSERT or OP_ASSERT_NOT. Backward assertions use the opcodes
OP_ASSERTBACK and OP_ASSERTBACK_NOT, and the first opcode inside the assertion
is OP_REVERSE, followed by a two byte count of the number of characters to move
back the pointer in the subject string. When operating in UTF-8 mode, the count
is a character count rather than a byte count. A separate count is present in
each alternative of a lookbehind assertion, allowing them to have different
fixed lengths.


Once-only subpatterns
---------------------

These are also just like other subpatterns, but they start with the opcode
OP_ONCE.


Conditional subpatterns
-----------------------

These are like other subpatterns, but they start with the opcode OP_COND. If
the condition is a back reference, this is stored at the start of the
subpattern using the opcode OP_CREF followed by two bytes containing the
reference number. If the condition is "in recursion" (coded as "(?(R)"), the
same scheme is used, with a "reference number" of 0xffff. Otherwise, a
conditional subpattern always starts with one of the assertions.


Recursion
---------

Recursion either matches the current regex, or some subexpression. The opcode
OP_RECURSE is followed by an value which is the offset to the starting bracket
from the start of the whole pattern. From release 6.5, OP_RECURSE is 
automatically wrapped inside OP_ONCE brackets (because otherwise some patterns 
broke it). OP_RECURSE is also used for "subroutine" calls, even though they 
are not strictly a recursion.


Callout
-------

OP_CALLOUT is followed by one byte of data that holds a callout number in the
range 0 to 254 for manual callouts, or 255 for an automatic callout. In both 
cases there follows a two-byte value giving the offset in the pattern to the
start of the following item, and another two-byte item giving the length of the
next item.


Changing options
----------------

If any of the /i, /m, or /s options are changed within a pattern, an OP_OPT
opcode is compiled, followed by one byte containing the new settings of these
flags. If there are several alternatives, there is an occurrence of OP_OPT at
the start of all those following the first options change, to set appropriate
options for the start of the alternative. Immediately after the end of the
group there is another such item to reset the flags to their previous values. A
change of flag right at the very start of the pattern can be handled entirely
at compile time, and so does not cause anything to be put into the compiled
data.

Philip Hazel
June 2006
1	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
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	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	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
20	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
22	the pattern, as is expected in Unix and Perl-style regular expressions.
23
24	Historical note 2
25	-----------------
26
27	By contrast, the code originally written by Henry Spencer (which was
28	subsequently heavily modified for Perl) compiles the expression twice: once in
29	a dummy mode in order to find out how much store will be needed, and then for
30	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	maximizing (or, optionally, minimizing in Perl) the amount of the subject that
33	matches individual wild portions of the pattern. This is an "NFA algorithm" in
34	Friedl's terminology.
35
36	OK, here's the real stuff
37	-------------------------
38
39	For the set of functions that form the "basic" PCRE library (which are
40	unrelated to those mentioned above), I tried at first to invent an algorithm
41	that used an amount of store bounded by a multiple of the number of characters
42	in the pattern, to save on compiling time. However, because of the greater
43	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
77	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 that
79	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
90	------------------------------
91
92	These items are all just one byte long
93
94	OP_END end of pattern
95	OP_ANY match any character
96	OP_ANYBYTE match any single byte, even in UTF-8 mode
97	OP_SOD match start of data: \A
98	OP_SOM, start of match (subject + offset): \G
99	OP_CIRC ^ (start of data, or after \n in multiline)
100	OP_NOT_WORD_BOUNDARY \W
101	OP_WORD_BOUNDARY \w
102	OP_NOT_DIGIT \D
103	OP_DIGIT \d
104	OP_NOT_WHITESPACE \S
105	OP_WHITESPACE \s
106	OP_NOT_WORDCHAR \W
107	OP_WORDCHAR \w
108	OP_EODN match end of data or \n at end: \Z
109	OP_EOD match end of data: \z
110	OP_DOLL $ (end of data, or before \n in multiline)
111	OP_EXTUNI match an extended Unicode character
112
113
114	Repeating single characters
115	---------------------------
116
117	The common repeats (*, +, ?) when applied to a single character use the
118	following opcodes:
119
120	OP_STAR
121	OP_MINSTAR
122	OP_PLUS
123	OP_MINPLUS
124	OP_QUERY
125	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
129	the character that is to be repeated. Other repeats make use of
130
131	OP_UPTO
132	OP_MINUPTO
133	OP_EXACT
134
135	which are followed by a two-byte count (most significant first) and the
136	repeated character. OP_UPTO matches from 0 to the given number. A repeat with a
137	non-zero minimum and a fixed maximum is coded as an OP_EXACT followed by an
138	OP_UPTO (or OP_MINUPTO).
139
140
141	Repeating character types
142	-------------------------
143
144	Repeats of things like \d are done exactly as for single characters, except
145	that instead of a character, the opcode for the type is stored in the data
146	byte. The opcodes are:
147
148	OP_TYPESTAR
149	OP_TYPEMINSTAR
150	OP_TYPEPLUS
151	OP_TYPEMINPLUS
152	OP_TYPEQUERY
153	OP_TYPEMINQUERY
154	OP_TYPEUPTO
155	OP_TYPEMINUPTO
156	OP_TYPEEXACT
157
158
159	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_CHAR opcode is followed by a single character that is to be matched
176	casefully. For caseless matching, OP_CHARNC is used. In UTF-8 mode, the
177	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
183	-----------------
184
185	If there is only one character, OP_CHAR or OP_CHARNC is used for a positive
186	class, and OP_NOT for a negative one (that is, for something like [^a]).
187	However, in UTF-8 mode, the use of OP_NOT applies only to characters with
188	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,
191	negated, single-character class. The normal ones (OP_STAR etc.) are used for a
192	repeated positive single-character class.
193
194	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
196	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
198	significant end of each byte.
199
200	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
202	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
205	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
207	whether it's a positive or a negative class.
208
209
210	Back references
211	---------------
212
213	OP_REF is followed by two bytes containing the reference number.
214
215
216	Repeating character classes and back references
217	-----------------------------------------------
218
219	Single-character classes are handled specially (see above). This applies to
220	OP_CLASS and OP_REF. In both cases, the repeat information follows the base
221	item. The matching code looks at the following opcode to see if it is one of
222
223	OP_CRSTAR
224	OP_CRMINSTAR
225	OP_CRPLUS
226	OP_CRMINPLUS
227	OP_CRQUERY
228	OP_CRMINQUERY
229	OP_CRRANGE
230	OP_CRMINRANGE
231
232	All but the last two are just single-byte items. The others are followed by
233	four bytes of data, comprising the minimum and maximum repeat counts.
234
235
236	Brackets and alternation
237	------------------------
238
239	A pair of non-capturing (round) brackets is wrapped round each expression at
240	compile time, so alternation always happens in the context of brackets.
241
242	Non-capturing brackets use the opcode OP_BRA, while capturing brackets use
243	OP_BRA+1, OP_BRA+2, etc. [Note for North Americans: "bracket" to some English
244	speakers, including myself, can be round, square, curly, or pointy. Hence this
245	usage.]
246
247	Originally PCRE was limited to 99 capturing brackets (so as not to use up all
248	the opcodes). From release 3.5, there is no limit. What happens is that the
249	first ones, up to EXTRACT_BASIC_MAX are handled with separate opcodes, as
250	above. If there are more, the opcode is set to EXTRACT_BASIC_MAX+1, and the
251	first operation in the bracket is OP_BRANUMBER, followed by a 2-byte bracket
252	number. This opcode is ignored while matching, but is fished out when handling
253	the bracket itself. (They could have all been done like this, but I was making
254	minimal changes.)
255
256	A bracket opcode is followed by LINK_SIZE bytes which give the offset to the
257	next alternative OP_ALT or, if there aren't any branches, to the matching
258	OP_KET opcode. Each OP_ALT is followed by LINK_SIZE bytes giving the offset to
259	the next one, or to the OP_KET opcode.
260
261	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
263	maximally respectively. All three are followed by LINK_SIZE bytes giving (as a
264	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
267	is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte
268	opcodes which tell the matcher that skipping this subpattern entirely is a
269	valid branch.
270
271	A subpattern with an indefinite maximum repetition is replicated in the
272	compiled data its minimum number of times (or once with OP_BRAZERO if the
273	minimum is zero), with the final copy terminating with OP_KETRMIN or OP_KETRMAX
274	as appropriate.
275
276	A subpattern with a bounded maximum repetition is replicated in a nested
277	fashion up to the maximum number of times, with OP_BRAZERO or OP_BRAMINZERO
278	before each replication after the minimum, so that, for example, (abc){2,5} is
279	compiled as (abc)(abc)((abc)((abc)(abc)?)?)?.
280
281
282	Assertions
283	----------
284
285	Forward assertions are just like other subpatterns, but starting with one of
286	the opcodes OP_ASSERT or OP_ASSERT_NOT. Backward assertions use the opcodes
287	OP_ASSERTBACK and OP_ASSERTBACK_NOT, and the first opcode inside the assertion
288	is OP_REVERSE, followed by a two byte count of the number of characters to move
289	back the pointer in the subject string. When operating in UTF-8 mode, the count
290	is a character count rather than a byte count. A separate count is present in
291	each alternative of a lookbehind assertion, allowing them to have different
292	fixed lengths.
293
294
295	Once-only subpatterns
296	---------------------
297
298	These are also just like other subpatterns, but they start with the opcode
299	OP_ONCE.
300
301
302	Conditional subpatterns
303	-----------------------
304
305	These are like other subpatterns, but they start with the opcode OP_COND. If
306	the condition is a back reference, this is stored at the start of the
307	subpattern using the opcode OP_CREF followed by two bytes containing the
308	reference number. If the condition is "in recursion" (coded as "(?(R)"), the
309	same scheme is used, with a "reference number" of 0xffff. Otherwise, a
310	conditional subpattern always starts with one of the assertions.
311
312
313	Recursion
314	---------
315
316	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
318	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
325	-------
326
327	OP_CALLOUT is followed by one byte of data that holds a callout number in the
328	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
335	----------------
336
337	If any of the /i, /m, or /s options are changed within a pattern, an OP_OPT
338	opcode is compiled, followed by one byte containing the new settings of these
339	flags. If there are several alternatives, there is an occurrence of OP_OPT at
340	the start of all those following the first options change, to set appropriate
341	options for the start of the alternative. Immediately after the end of the
342	group there is another such item to reset the flags to their previous values. A
343	change of flag right at the very start of the pattern can be handled entirely
344	at compile time, and so does not cause anything to be put into the compiled
345	data.
346
347	Philip Hazel
348	June 2006