Skip to content


Subversion checkout URL

You can clone with
Download ZIP
100644 269 lines (192 sloc) 9.779 kB
89b9891 @pmichaud Add initial draft of S07-lists.pod, the new Synopsis 7.
pmichaud authored
2 =encoding utf8
4 =head1 TITLE
6 Synopsis 7: Lists and Iteration
8 =head1 AUTHORS
10 Patrick R. Michaud <
12 =head1 VERSION
14 Created: 12 July 2012
16 Last Modified: 12 July 2012
17 Version: 1
19 =head1 Overview
21 Lists and arrays have always been one of Perl's fundamental data types,
22 and Perl 6 is no different. However, lists in Perl 6 have been greatly
23 extended to accommodate lazy lists, infinite lists, lists of mutable
24 and immutable elements, typed lists, flattening behaviors, and so on.
25 So where lists and arrays in Perl 5 tended to be finite sequences of
26 scalar values, in Perl 6 we have additional dimensions of behavior
27 that must be addressed.
29 =head2 The C<List> type
31 The C<List> class is the base class for dealing with other types of
32 lists, including C<Array>. To the programmer, a C<List> is a potentially
33 lazy and infinite sequence of elements. A C<List> is mutable, in that
34 one can manipulate the sequence via operations such as C<push>, C<pop>,
35 C<shift>, C<unshift>, C<splice>, etc. A C<List>'s elements may be
36 either mutable or immutable.
38 C<List> objects are C<Positional>, meaning they can be bound to
39 array variables and support the postfix C<.[]> operator.
41 Lists are also lazy, in that the elements of a C<List> may
42 come from generator functions (called C<Iterator>s) that produce
43 elements on demand.
45 =head2 The C<Array> type
47 An C<Array> is simply a C<List> in which all of the elements are held
48 in scalar containers. This allows assignment to the elements of the
49 array.
51 =head2 The C<Parcel> type
53 The comma operator (C<< infix:<,> >>) creates C<Parcel> objects.
54 These should not be confused with lists; a C<Parcel> represents
55 a raw syntactic sequence of elements. A C<Parcel> is immutable,
56 although the elements of a C<Parcel> may be either mutable or
57 immutable.
59 The name "Parcel" is derived from the phrase "parenthesis cell",
60 since many C<Parcel> objects appear inside of parentheses.
61 However, except for the empty parcel, it's the comma operator
62 that creates C<Parcel> objects.
64 () # empty Parcel
65 (1) # an Int
66 (1,2) # a Parcel with two Ints
67 (1,) # a Parcel with one Int
69 A C<Parcel> is also C<Positional>, and uses flattening context for
70 list operations such as C<.[]> and C<.elems>. See "Flattening contexts"
71 below.
73 =head2 Flattening contexts, the C<Iterable> type
75 C<List> and C<Parcel> objects can have other container objects
76 as elements. In some contexts we want to interpolate the values
77 of container objects into the surrounding C<List> or C<Parcel>,
78 while in other contexts we want any subcontainers to be preserved.
79 Such interpolation is known as "flattening".
81 The C<Iterable> type is performed by container and generator objects
82 that will interpolate their values in flattening contexts. Both
83 C<List> and C<Parcel> are C<Iterable>. C<Iterable> implies support
84 for C<.iterator>, which returns an object capable of producing the
85 elements of the C<Iterable>.
87 Objects held in scalar containers are never interpolated in
88 flattening context, even if the object is C<Iterable>.
90 my @a = 3,4,5;
91 for 1,2,@a { ... } # five iterations
93 my $s = @a;
94 for 1,2,$s { ... } # three iterations
96 Here, both C<$s> and C<@a> refer to the same underlying C<Array> object,
97 but the presence of the scalar container prevents C<$s> from being
98 flattened into the C<for> loop. The C<.list> or C<.flat> method
99 may be used to restore the flattening behavior:
101 for 1,2,$s.list { ... } # five iterations
102 for 1,2,@($s) { ... } # five iterations
104 Conversely, the C<.item> or C<$()> contextualizer can be used to
105 prevent an C<Iterable> from being interpolated:
107 my @b = 1,2,@a; # @b has five elements
108 my @c = 1,2,@a.item; # @c has three elements
109 my @c = 1,2,$(@a); # same thing
111 =head2 Iterators
113 An C<Iterator> is an object that is able to generate or produce
114 elements of a list (called "reification"). Because a single
115 C<Iterator> may be directly or indirectly referenced by
116 multiple container objects simultaneously, C<Iterator>s provide
117 an immutable view of the sequence of elements produced, leaving
118 it up to containers to provide any mutability.
120 =head3 The C<.reify> method
122 The C<.reify($n)> method requests an C<Iterator> to return a C<Parcel>
123 consisting of at least C<$n> reified elements, followed by any additional
124 iterators needed for the remaining elements in the sequence. For example:
126 my $r = 10..50;
127 say $r.iterator.reify(5).perl; # (10, 11, 12, 13, 14, 15..50)
129 Iterators are allowed to return more or fewer elements than requested
130 by C<$n>; it's up to the caller to properly handle any shortage or
131 excess. (In general an iterator is expected to always reify at
132 least one element, however.) If C<$n> is C<*>, then the iterator
133 is asked to generate elements according to whatever is most natural
134 for the thing being iterated. For a C<Range> this might be a substantial
135 number (or even all) of the elements of the range; for a filehandle
136 it might read a single record; for a C<List> it may return only the
137 non-lazy portion of the list.
139 Once C<.reify> is called on an iterator, the iterator is expected
140 to return the same results for all subsequent calls to C<.reify>.
141 This is true even if the C<$n> argument is different from one call
142 to the next. The general pattern is often something like:
144 class SomeIter is Iterator {
145 has $!reified;
147 method reify($n = 1) {
148 unless $!reified.defined {
149 # generate return Parcel and bind to $!reified
150 }
151 $!reified;
152 }
153 }
155 =head3 The C<.infinite> method
157 Because lists in Perl 6 can be lazy, they can also be infinite.
158 Each C<Iterator> must provide an C<.infinite> method, which returns
159 a value indicating the knowable finiteness of the iteration:
161 .infinite Meaning
162 ----------------------------------------------
163 True iteration is known to be infinite
164 False iteration is known to be finite
165 Mu finiteness isn't currently known
167 As an example, C<Range> iterators can generally know finiteness simply
168 by looking at the endpoint of the C<Range>. The iterator for the
169 C<< infix:<...> >> sequence operator treats any sequence ending in
170 C<*> as being "known infinite", all other C<...> sequences have unknown
171 finiteness. In the general case it's not possible for loop iterators
172 to definitively know if the sequence will be finite or infinite.
173 (Conjecture: There will be a syntax or mechanism for the programmer
174 to indicate that such sequences are to be treated as known finite
175 or known infinite.)
177 =head2 Levels of laziness
179 Laziness is one of the primary virtues of a Perl programmer; it is
180 also one of the virtues of Perl 6. However, it's also possible to
181 succumb to false laziness, so there are times when Perl 6 chooses
182 to be eager instead of lazy.
184 Perl 6 defines four levels of laziness:
186 =over
188 =item Strictly Lazy
190 Does not evaluate anything unless explicitly required by the caller,
191 including not traversing non-lazy objects. This behavior generally
192 occurs only by a pragma or by explicit lazy primitives.
194 =item Mostly Lazy
196 Try to obtain available elements without causing eager evaluation
197 of other lazy objects. However, the implementation is allowed to
198 do batching for efficiency. The programmer must not rely on
199 circular side effects when the implementation is working ahead
200 like this, or the results will be indeterminate. (However, this
201 does not rule out pure I<definitional> circularity; earlier array
202 values may be used in the definition of subsequent values.)
204 =item Mostly Eager
206 Obtain all leading items that are not known to be infinite.
207 The remainder of the items are left for lazy evaluation.
208 As with the "mostly lazy" case, the programmer must not depend
209 on circular side effects in any situation where the implementation
210 is allowed to work ahead. Note that there are no language-defined
211 limits on the amount of conjectural evaluation allowed, up to the
212 size of available memory; however, an implementation may allow
213 the arbitrary limitation of workahead by use of pragmas.
215 In any case there should be no profound semantic differences
216 between "mostly lazy" and "mostly eager". These levels are
217 primarily just hints to the implementation as to whether you are
218 likely to want to use all the values in the list. Nothing
219 transactional should be allowed to depend on the difference.
221 =item Strictly Eager
223 Obtain all items, failing immediately with an appropriate message
224 if asked to evaluate any data structures known to be infinite.
225 (Data structures that are effectively infinite but not provably
226 so will likely exhaust memory instead.)
228 =back
230 =head2 The laziness level of some common operations
232 =over
234 =item List assignment; my @a = @something;
236 In order to provide p5-like behavior in list assignment, it is performed
237 at the Mostly Eager level. This means that if you do
239 my @a = foo();
241 it will eagerly evaluate the return value from C<foo()> to place
242 elements into C<@a>, stopping only when encountering something
243 that is "known infinite" or upon reaching the end of the sequence.
245 =item C<for>, C<map>, C<grep>, etc.
247 The C<for> loop and looping routines such as C<map>, C<grep>, etc. are
248 Mostly Lazy by default, but will be eager when evaluated in sink context.
250 =item Slurpy parameters
252 Slurpy parameters are Mostly Lazy.
254 =item Feed operators
256 The feed operator is strictly lazy, meaning that no operation
257 should be performed before the user requests any element. That's how
259 my @a <== grep { ... } <== map { ... } <== grep { ... } <== 1, 2, 3
261 is completely lazy, even if 1,2,3 is a fairly small known compact
262 list.
264 =back
266 =head2 Common list operations
Something went wrong with that request. Please try again.