]> git.proxmox.com Git - ceph.git/blob - ceph/src/boost/libs/metaparse/doc/manual.qbk
bump version to 12.2.2-pve1
[ceph.git] / ceph / src / boost / libs / metaparse / doc / manual.qbk
1 [#manual]
2 [section User manual]
3
4 [section What is a parser]
5
6 See the [link parser parser] section of the [link reference reference] for the
7 explanation of what a parser is.
8
9 [section The input of the parsers]
10
11 Parsers take a [link string `string`] as input, which represents a string
12 for template metaprograms. For example the string `"Hello World!"` can be
13 defined the following way:
14
15 string<'H','e','l','l','o',' ','W','o','r','l','d','!'>
16
17 This syntax makes the input of the parsers difficult to read. Metaparse works
18 with compilers using C++98, but the input of the parsers has to be defined the
19 way it is described above.
20
21 Based on `constexpr`, a feature provided by C++11, Metaparse provides a macro,
22 [link BOOST_METAPARSE_STRING `BOOST_METAPARSE_STRING`] for defining strings:
23
24 BOOST_METAPARSE_STRING("Hello World!")
25
26 This defines a [link string `string`] as well, however, it is easier to
27 read. The maximum length of the string that can be defined this way is limited,
28 however, this limit is configurable. It is specified by the
29 `BOOST_METAPARSE_LIMIT_STRING_SIZE` macro.
30
31 [endsect]
32
33 [section Source positions]
34
35 A source position is described using a compile-time data structure. The
36 following functions can be used to query it:
37
38 * [link get_col `get_col`]
39 * [link get_line `get_line`]
40
41 The beginning of the input is [link start `start`] which requires
42 `<boost/metaparse/start.hpp>` to be included.
43
44 [endsect]
45
46 [section Error handling]
47
48 An error is described using a compile-time data structure. It contains
49 information about the source position where the error was detected and some
50 [link parsing_error_message description] about the error.
51 [link debug_parsing_error `debug_parsing_error`] can be used to display the
52 error message. Metaparse provides the
53 [link BOOST_METAPARSE_DEFINE_ERROR `BOOST_METAPARSE_DEFINE_ERROR`] macro for
54 defining simple [link parsing_error_message parsing error message]s.
55
56 [endsect]
57
58 [section Some examples of simple parsers]
59
60 * A parser that parses nothing and always succeeds is
61 [link return_ `return_`].
62 * A parser that always fails is [link fail `fail`].
63 * A parser that parses one character and returns the parsed character as the
64 result is [link one_char `one_char`].
65
66 [endsect]
67
68 [section Combining parsers]
69
70 Complex parsers can be built by combining simple parsers. The parser library
71 contains a number of parser combinators that build new parsers from already
72 existing ones.
73
74 For example
75 [link accept_when `accept_when`]`<Parser, Predicate, RejectErrorMsg>` is a
76 parser. It uses `Parser` to parse the input. When `Parser` rejects the input,
77 the combinator returns the error `Parser` failed with. When `Parser` is
78 successful, the combinator validates the result using `Predicate`. If the
79 predicate returns true, the combinator accepts the input, otherwise it generates
80 an error with the message `RejectErrorMsg`.
81
82 Having [link accept_when `accept_when`], [link one_char `one_char`] can be
83 used to build parsers that accept only digit characters, only whitespaces, etc.
84 For example [link digit `digit`] accepts only digit characters:
85
86 typedef
87 boost::metaparse::accept_when<
88 boost::metaparse::one_char,
89 boost::metaparse::util::is_digit,
90 boost::metaparse::errors::digit_expected
91 >
92 digit;
93
94 [endsect]
95
96 [section Sequence]
97
98 The result of a successful parsing is some value and the remaining string that
99 was not parsed. The remaining string can be processed by another parser. The
100 parser library provides a parser combinator, [link sequence `sequence`],
101 that takes a number of parsers as arguments and builds a new parser from them
102 that:
103
104 * Parses the input using the first parser
105 * If parsing succeeds, it parses the remaining string with the second parser
106 * It continues applying the parsers in order as long as they succeed
107 * If all of them succeed, it returns the list of results
108 * If any of the parsers fails, the combinator fails as well and returns the
109 error the first failing parser returned with
110
111 [endsect]
112
113 [#repetition]
114 [section Repetition]
115
116 It is a common thing to parse a list of things of unknown length. As an example
117 let's start with something simple: the text is a list of numbers. For example:
118
119 11 13 3 21
120
121 We want the result of parsing to be the sum of these values. Metaparse provides
122 the [link int_ `int_`] parser we can use to parse one of these numbers.
123 Metaparse provides the [link token `token`] combinator to consume the
124 whitespaces after the number. So the following parser parses one number and the
125 whitespaces after it:
126
127 using int_token = token<int_>;
128
129 The result of parsing is a boxed integer value: the value of the parsed number.
130 For example parsing
131 [link BOOST_METAPARSE_STRING `BOOST_METAPARSE_STRING`]`("13 ")` gives
132 `boost::mpl::int_<13>` as the result.
133
134 Our example input is a list of numbers. Each number can be parsed by
135 `int_token`:
136
137 [$images/metaparse/repeated_diag0.png [width 70%]]
138
139 This diagram shows how the repeated application of `int_token` can parse the
140 example input. Metaparse provides the [link repeated `repeated`] parser to
141 easily implement this. The result of parsing is a typelist: the list of the
142 individual numbers.
143
144 [$images/metaparse/repeated_diag1.png [width 70%]]
145
146 This diagram shows how [link repeated `repeated`]`<int_token>` works. It uses
147 the `int_token` parser repeatedly and builds a `boost::mpl::vector` from the
148 results it provides.
149
150 But we need the sum of these, so we need to summarise the result. We can do this
151 by wrapping our parser, [link repeated `repeated`]`<int_token>` with
152 [link transform `transform`]. That gives us the opportunity to specify a
153 function transforming this typelist to some other value - the sum of the
154 elements in our case. Initially let's ignore how to summarise the elements in
155 the vector. Let's assume that it can be implemented by a lambda expression and
156 use `boost::mpl::lambda<...>::type` representing that lambda expression. Here is
157 an example using [link transform `transform`] and this lambda expression:
158
159 using sum_parser =
160 transform<
161 repeated<int_token>,
162 boost::mpl::lambda<...>::type
163 >;
164
165 The [link transform `transform`]`<>` parser combinator wraps the
166 [link repeated `repeated`]`<int_token>` to build the parser we need. Here is a
167 diagram showing how it works:
168
169 [$images/metaparse/repeated_diag2.png [width 70%]]
170
171 As the diagram shows, the
172 [link transform `transform`]`<`[link repeated `repeated`]`<int_token>, ...>`
173 parser parses the input using [link repeated `repeated`]`<int_token>` and then
174 does some processing on the result of parsing.
175
176 Let's implement the missing lambda expression that tells
177 [link transform `transform`] how to change the result coming from
178 [link repeated `repeated`]`<int_token>`. We can summarise the numbers in a
179 typelist by using Boost.MPL's `fold` or `accumulate`. Here is an example doing
180 that:
181
182 using sum_op = mpl::lambda<mpl::plus<mpl::_1, mpl::_2>>::type;
183
184 using sum_parser =
185 transform<
186 repeated<int_token>,
187 mpl::lambda<
188 mpl::fold<mpl::_1, mpl::int_<0>, sum_op>
189 >::type
190 >;
191
192 Here is an extended version of the above diagram showing what happens here:
193
194 [$images/metaparse/repeated_diag3.png [width 70%]]
195
196 This example parses the input, builds the list of numbers and then loops over it
197 and summarises the values. It starts with the second argument of `fold`,
198 `int_<0>` and adds every item of the list of numbers (which is the result of
199 the parser [link repeated `repeated`]`<int_token>`) one by one.
200
201 [note
202 Note that [link transform `transform`] wraps another parser,
203 [link repeated `repeated`]`<int_token>` here. It parses the input with that
204 parser, gets the result of that parsing and changes that result.
205 [link transform `transform`] itself will be a parser returning that updated
206 result.
207 ]
208
209 [#introducing-foldl]
210 [section Introducing foldl]
211
212 It works, however, this is rather inefficient: it has a loop parsing the
213 integers one by one, building a typelist and then it loops over this typelist to
214 summarise the result. Using template metaprograms in your applications can have
215 a serious impact on the compiler's memory usage and the speed of the
216 compilation, therefore I recommend being careful with these things.
217
218 Metaparse offers more efficient ways of achieving the same result. You don't
219 need two loops: you can merge them together and add every number to your summary
220 right after parsing it. Metaparse offers the [link foldl `foldl`] for this.
221
222 With [link foldl `foldl`] you specify:
223
224 * the parser to parse the individual elements of the list
225 (which is `int_token` in our example)
226 * the initial value used for folding (which is `int_<0>` in our example)
227 * the forward operation merging the sub-result we have so far and the value
228 coming from the last application of the parser (this was `sum_op` in our
229 example)
230
231 Our parser can be implemented this way:
232
233 using better_sum_parser = foldl<int_token, mpl::int_<0>, sum_op>;
234
235 As you can see the implementation of the parser is more compact.
236 Here is a diagram showing what happens when you use this parser to parse some
237 input:
238
239 [$images/metaparse/foldl_diag1.png [width 70%]]
240
241 As you can see, not only the implementation of the parser is more compact, but
242 it achieves the same result by doing less as well. It parses the input by
243 applying `int_token` repeatedly, just like the previous solution. But it
244 produces the final result without building a typelist as an internal step. Here
245 is how it works internally:
246
247 [$images/metaparse/foldl_diag2.png [width 70%]]
248
249 It summarises the results of the repeated `int_token` application using
250 `sum_op`. This implementation is more efficient. It accepts an empty string as a
251 valid input: the sum of it is `0`. It may be good for you, in which case you are
252 done. If you don't wan to accept it, you can use [link foldl1 `foldl1`] instead
253 of [link foldl `foldl`]. This is the same, but it rejects empty input.
254 (Metaparse offers [link repeated1 `repeated1`] as well if you choose the first
255 approach and would like to reject empty string)
256
257 [endsect]
258
259 [#introducing-foldr]
260 [section Introducing foldr]
261
262 [note
263 Note that if you are reading this manual for the first time, you probably want
264 to skip this section and proceed with
265 [link introducing-foldl_start_with_parser Introducing foldl_start_with_parser]
266 ]
267
268 You might have noticed that Metaparse offers [link foldr `foldr`] as well. The
269 difference between [link foldl `foldl`] and [link foldr `foldr`] is the
270 direction in which the results are summarised. (`l` stands for ['from the Left]
271 and `r` stands for ['from the Right]) Here is a diagram showing how
272 `better_sum_parser` works if it is implemented using [link foldr `foldr`]:
273
274 [$images/metaparse/foldr_diag1.png [width 70%]]
275
276 As you can see this is very similar to using [link foldl `foldl`], but the
277 results coming out of the individual applications of `int_token` are summarised
278 in a right-to-left order. As `sum_op` is addition, it does not affect the end
279 result, but in other cases it might.
280
281 [note
282 Note that the implementation of [link foldl `foldl`] is more efficient than
283 [link foldr `foldr`]. Prefer [link foldl `foldl`] whenever possible.
284 ]
285
286 As you might expect it, Metaparse offers [link foldr1 `foldr1`] as well, which
287 folds from the right and rejects empty input.
288
289 [endsect]
290
291 [#introducing-foldl_start_with_parser]
292 [section Introducing foldl_start_with_parser]
293
294 Let's change the grammar of our little language. Instead of a list of numbers,
295 let's expect numbers separated by a `+` symbol. Our example input becomes the
296 following:
297
298 BOOST_METAPARSE_STRING("11 + 13 + 3 + 21")
299
300 Parsing it with [link foldl `foldl`] or [link repeated `repeated`] is difficult:
301 there has to be a `+` symbol before every element ['except] the first one. None
302 of the already introduced repetition constructs offer a way of treating the
303 first element in a different way.
304
305 If we forget about the first number for a moment, the rest of the input is
306 `"+ 13 + 3 + 21"`. This can easily be parsed by [link foldl `foldl`] (or
307 [link repeated `repeated`]):
308
309 using plus_token = token<lit_c<'+'>>;
310 using plus_int = last_of<plus_token, int_token>;
311
312 using sum_parser2 = foldl<plus_int, int_<0>, sum_op>;
313
314 It uses `plus_int`, that is [link last_of `last_of`]`<plus_token, int_token>`
315 as the parser that is used repeatedly to get the numbers. It does the following:
316
317 * Uses `plus_token` to parse the `+` symbol and any whitespace that might follow
318 it.
319 * Uses then `int_token` to parse the number
320 * Combines the above two with [link last_of `last_of`] to use both parsers in
321 order and keep only the result of using the second one (the result of parsing
322 the `+` symbol is thrown away - we don't care about it).
323
324 This way [link last_of `last_of`]`<plus_token, int_token>` returns the value of
325 the number as the result of parsing, just like our previous parser, `int_token`
326 did. Because of this, it can be used as a drop-in replacement of `int_token` in
327 the previous example and we get a parser for our updated language. Or at least
328 for all number except the first one.
329
330 This [link foldl `foldl`] can not parse the first element, because it expects a
331 `+` symbol before every number. You might think of making the `+` symbol
332 optional in the above approach - don't do that. It makes the parser accept
333 `"11 + 13 3 21"` as well as the `+` symbol is now optional ['everywhere].
334
335 What you could do is parsing the first element with `int_token`, the rest of
336 the elements with the above [link foldl `foldl`]-based solution and add the
337 result of the two. This is left as an exercise to the reader.
338
339 Metaparse offers [link foldl_start_with_parser `foldl_start_with_parser`] to
340 implement this. [link foldl_start_with_parser `foldl_start_with_parser`] is the
341 same as [link foldl `foldl`]. The difference is that instead of an initial value
342 to combine the list elements with it takes an ['initial parser]:
343
344 using plus_token = token<lit_c<'+'>>;
345 using plus_int = last_of<plus_token, int_token>;
346
347 using sum_parser3 = foldl_start_with_parser<plus_int, int_token, sum_op>;
348
349 [link foldl_start_with_parser `foldl_start_with_parser`] starts with applying
350 that initial parser and uses the result it returns as the initial value for
351 folding. It does the same as [link foldl `foldl`] after that. The following
352 diagram shows how it can be used to parse a list of numbers separated by `+`
353 symbols:
354
355 [$images/metaparse/foldl_start_with_parser_diag1.png [width 70%]]
356
357 As the diagram shows, it start parsing the list of numbers with `int_token`,
358 uses its value as the starting value for folding (earlier approaches were using
359 the value `int_<0>` as this starting value). Then it parses all elements of the
360 list by using `plus_int` multiple times.
361
362 [endsect]
363
364 [#introducing-foldr_start_with_parser]
365 [section Introducing foldr_start_with_parser]
366
367 [note
368 Note that if you are reading this manual for the first time, you probably want
369 to skip this section and try creating some parsers using
370 [link foldl_start_with_parser `foldl_start_with_parser`] instead.
371 ]
372
373 [@foldl_start_with_parser.hpp `foldl_start_with_parser`] has its
374 ['from the right] pair,
375 [link foldr_start_with_parser `foldr_start_with_parser`]. It uses the same
376 elements as [link foldl_start_with_parser `foldl_start_with_parser`] but in a
377 different order. Here is a parser for our example language implemented with
378 [link foldr_start_with_parser `foldr_start_with_parser`]:
379
380 using plus_token = token<lit_c<'+'>>;
381 using int_plus = first_of<int_token, plus_token>;
382
383 using sum_parser4 = foldr_start_with_parser<int_plus, int_token, sum_op>;
384
385 Note that it uses `int_plus` instead of `plus_int`. This is because the parser
386 the initial value for folding comes from is used after `int_plus` has parsed the
387 input as many times as it could. It might sound strange for the first time, but
388 the following diagram should help you understand how it works:
389
390 [$images/metaparse/foldr_start_with_parser_diag1.png [width 70%]]
391
392 As you can see, it starts with the parser that is applied repeatedly on the
393 input, thus instead of parsing `plus_token int_token` repeatedly, we need to
394 parse `int_token plus_token` repeatedly. The last number is not followed by `+`,
395 thus `int_plus` fails to parse it and it stops the iteration.
396 [link foldr_start_with_parser `foldr_start_with_parser`] then uses the other
397 parser, `int_token` to parse the input. It succeeds and the result it returns is
398 used as the starting value for folding from the right.
399
400 [note
401 Note that as the above description also suggests, the implementation of
402 [link foldl_start_with_parser `foldl_start_with_parser`] is more efficient
403 than [link foldr_start_with_parser `foldr_start_with_parser`]. Prefer
404 [link foldl_start_with_parser `foldl_start_with_parser`] whenever possible.
405 ]
406
407 [endsect]
408
409 [#introducing-foldl_reject_incomplete_start_with_parser]
410 [section Introducing foldl_reject_incomplete_start_with_parser]
411
412 Using a parser built with
413 [link foldl_start_with_parser `foldl_start_with_parser`] we can parse the input
414 when the input is correct. However, it is not always the case. Consider the
415 following input for example:
416
417 BOOST_METAPARSE_STRING("11 + 13 + 3 + 21 +")
418
419 This is an invalid expression. However, if we parse it using the
420 [link foldl_start_with_parser `foldl_start_with_parser`]-based parser presented
421 earlier (`sum_parser3`), it accepts the input and the result is `48`. This is
422 because [link foldl_start_with_parser `foldl_start_with_parser`] parses the
423 input ['as long as it can]. It parses the first`int_token` (`11`) and then it
424 starts parsing the `plus_int` elements (`+ 13`, `+ 3`, `+ 21`). After parsing
425 all of these, it tries to parse the remaining `" +"` input using `plus_int`
426 which fails and therefore
427 [link foldl_start_with_parser `foldl_start_with_parser`] stops after `+ 21`.
428
429 The problem is that the parser parses the longest sub-expression starting from
430 the beginning, that represents a valid expression. The rest is ignored. The
431 parser can be wrapped by [link entire_input `entire_input`] to make sure to
432 reject expressions with invalid extra characters at the end, however, that
433 won't make the error message useful. ([link entire_input `entire_input`] can
434 only tell the author of the invalid expression that after `+ 21` is something
435 wrong).
436
437 Metaparse provides
438 [link foldl_reject_incomplete_start_with_parser `foldl_reject_incomplete_start_with_parser`],
439 which does the same as [link foldl_start_with_parser `foldl_start_with_parser`],
440 except that once no further repetitions are found, it checks ['where] the
441 repeated parser (in our example `plus_int`) fails. When it can make any progress
442 (eg. it finds a `+` symbol), then
443 [link foldl_reject_incomplete_start_with_parser `foldl_reject_incomplete_start_with_parser`]
444 assumes, that the expression's author intended to make the repetition longer,
445 but made a mistake and propagates the error message coming from that last broken
446 expression.
447
448 [$images/metaparse/foldl_reject_incomplete_start_with_parser_diag1.png [width 70%]]
449
450 The above diagram shows how
451 [link foldl_reject_incomplete_start_with_parser `foldl_reject_incomplete_start_with_parser`]
452 parses the example invalid input and how it fails. This can be used for better
453 error reporting from the parsers.
454
455 Other folding parsers also have their `f` version. (eg.
456 [link foldr_reject_incomplete `foldr_reject_incomplete`],
457 [link foldl_reject_incomplete1 `foldl_reject_incomplete1`], etc).
458
459 [endsect]
460
461 [#finding-the-right-folding-parser-combinator]
462 [section Finding the right folding parser combinator]
463
464 As you might have noticed, there are a lot of different folding parser
465 combinators. To help you find the right one, the following naming convention is
466 used:
467
468 [$images/metaparse/folds.png [width 70%]]
469
470 [note
471 Note that there is no `foldr_reject_incomplete_start_with_parser`. The `p`
472 version of the right-folding parsers applies the special parser, whose result
473 is the initial value, after the repeated elements. Therefore, when the parser
474 parsing one repeated element fails, `foldr_start_with_parser` would apply that
475 special final parser instead of checking how the repeated element's parser
476 failed.
477 ]
478
479 [endsect]
480
481 [endsect]
482
483 [#result_types]
484 [section What can be built from a compile-time string?]
485
486 Parsers built using Metaparse are template metaprograms parsing text (or code)
487 at compile-time. Here is a list of things that can be the "result" of parsing:
488
489 * A ['type]. An example for this is a parser parsing a `printf` format string
490 and returning the typelist (eg. `boost::mpl::vector`) of the expected
491 arguments.
492 * A ['constant value]. An example for this is the result of a calculator
493 language. See the [link getting_started Getting Started] section for further
494 details.
495 * A ['runtime object]. A static runtime object can be generated that might be
496 used at runtime. An example for this is parsing regular expressions at
497 compile-time and building `boost::xpressive::sregex` objects. See the
498 `regex` example of Metaparse for an example.
499 * A C++ ['function], which might be called at runtime. A C++ function can be
500 generated that can be called at runtime. It is good for generating native
501 (and optimised) code from EDSLs. See the `compile_to_native_code` example of
502 Metaparse as an example for this.
503 * A [link metafunction_class ['template metafunction class]]. The result of
504 parsing might be a type, which is a
505 [link metafunction_class template metafunction class]. This is good for
506 building an EDSL for template metaprogramming. See the `meta_hs` example of
507 Metaparse as an example for this.
508
509 [endsect]
510
511 [section Grammars]
512
513 Metaparse provides a way to define grammars in a syntax that resembles EBNF. The
514 [link grammar `grammar`] template can be used to define a grammar. It can be
515 used the following way:
516
517 grammar<BOOST_METAPARSE_STRING("plus_exp")>
518 ::import<BOOST_METAPARSE_STRING("int_token"), token<int_>>::type
519
520 ::rule<BOOST_METAPARSE_STRING("ws ::= (' ' | '\n' | '\r' | '\t')*")>::type
521 ::rule<BOOST_METAPARSE_STRING("plus_token ::= '+' ws"), front<_1>>::type
522 ::rule<BOOST_METAPARSE_STRING("plus_exp ::= int_token (plus_token int_token)*"), plus_action>::type
523
524 The code above defines a parser from a grammar definition. The start symbol of
525 the grammar is `plus_exp`. The lines beginning with `::rule` define rules.
526 Rules optionally have a semantic action, which is a metafunction class that
527 transforms the result of parsing after the rule has been applied.
528 Existing parsers can be bound to names and be used in the rules by importing
529 them. Lines beginning with `::import` bind existing parsers to names.
530
531 The result of a grammar definition is a parser which can be given to other
532 parser combinators or be used directly. Given that grammars can import existing
533 parsers and build new ones, they are parser combinators as well.
534
535 [endsect]
536
537 [endsect]
538
539 [section Parsing based on `constexpr`]
540
541 Metaparse is based on template metaprogramming, however, C++11 provides
542 `constexpr`, which can be used for parsing at compile-time as well. While
543 implementing parsers based on `constexpr` is easier for a C++ developer, since
544 its syntax resembles the regular syntax of the language, the result of parsing
545 has to be a `constexpr` value. Parsers based on template metaprogramming can
546 build types as the result of parsing. These types may be boxed `constexpr`
547 values but can be metafunction classes, classes with static functions which can
548 be called at runtime, etc.
549
550 When a parser built with Metaparse needs a sub-parser for processing a part of
551 the input text and generating a `constexpr` value as the result of parsing, one
552 can implement the sub-parser based on `constexpr` functions. Metaparse
553 can be integrated with them and lift their results into C++ template
554 metaprogramming. An example demonstrating this feature can be found among the
555 examples (`constexpr_parser`). This capability makes it possible to integrate
556 Metaparse with parsing libraries based on `constexpr`.
557
558 [endsect]
559
560 [section What types of grammars can be used?]
561
562 It is possible to write parsers for ['context free grammars] using Metaparse.
563 However, this is not the most general category of grammars that can be used. As
564 Metaparse is a highly extendable framework, it is not clear what should be
565 considered to be the limit of Metaparse itself. For example Metaparse provides
566 the [link accept_when `accept_when`] [link parser_combinator parser combinator].
567 It can be used to provide arbitrary predicates for enabled/disabling a specific
568 rule. One can go as far as providing the Turing machine (as a
569 [link metafunction metafunction]) of the entire grammar as a predicate, so one
570 can build parsers for ['unrestricted grammars] that can be parsed using a Turing
571 machine. Note that such a parser would not be considered to be a parser built
572 with Metaparse, however, it is not clear how far a solution might go and still
573 be considered using Metaparse.
574
575 Metaparse assumes that the parsers are ['deterministic], as they have only "one"
576 result. It is of course possible to write parsers and combinators that return a
577 set (or list or some other container) of results as that "one" result, but that
578 can be considered building a new parser library. There is no clear boundary for
579 Metaparse.
580
581 Metaparse supports building ['top-down parsers] and ['left-recursion] is not
582 supported as it would lead to infinite recursion. ['Right-recursion] is
583 supported, however, in most cases the
584 [link repetition iterative parser combinators] provide better alternatives.
585
586 [endsect]
587
588 [endsect]
589