| 1 | <html> |
|---|
| 2 | <head> |
|---|
| 3 | <title>sequences.html</title> |
|---|
| 4 | <link rel="stylesheet" type="text/css" href="../styles.css"> |
|---|
| 5 | </head> |
|---|
| 6 | <body> |
|---|
| 7 | <h4> |
|---|
| 8 | Sequences |
|---|
| 9 | </h4> |
|---|
| 10 | <div> |
|---|
| 11 | A <i>sequence</i> (abbreviated to <i>seq</i>) is a group of adjacent parenthesized elements. For example, |
|---|
| 12 | </div> |
|---|
| 13 | <div class="code"> |
|---|
| 14 | (<i>a</i>)(<i>b</i>)(<i>c</i>) |
|---|
| 15 | </div> |
|---|
| 16 | <div> |
|---|
| 17 | ...is a <i>seq</i> of <i>3</i> elements--<i>a</i>, <i>b</i>, and <i>c</i>. |
|---|
| 18 | </div> |
|---|
| 19 | <div> |
|---|
| 20 | <i>Sequences</i> are data structures that merge the properties of both <i>lists</i> and |
|---|
| 21 | <i>tuples</i> with the exception that a <i>seq</i> cannot be empty. |
|---|
| 22 | Therefore, an "empty" <i>seq</i> is considered a special case scenario that |
|---|
| 23 | must be handled separately in C++. |
|---|
| 24 | </div> |
|---|
| 25 | <div class="code"> |
|---|
| 26 | <pre> |
|---|
| 27 | #define SEQ (x)(y)(z) |
|---|
| 28 | #define REVERSE(s, state, elem) (elem) state |
|---|
| 29 | // append to head ^ |
|---|
| 30 | |
|---|
| 31 | BOOST_PP_SEQ_FOLD_LEFT(REVERSE, BOOST_PP_EMPTY, SEQ)() |
|---|
| 32 | // #1 #2 |
|---|
| 33 | // 1) placeholder for "empty" seq |
|---|
| 34 | // 2) remove placeholder |
|---|
| 35 | |
|---|
| 36 | #define SEQ_B (1)(2)(3) |
|---|
| 37 | #define INC(s, state, elem) state (BOOST_PP_INC(elem)) |
|---|
| 38 | // append to tail ^ |
|---|
| 39 | |
|---|
| 40 | BOOST_PP_SEQ_FOLD_RIGHT(INC, BOOST_PP_SEQ_NIL, SEQ) |
|---|
| 41 | // ^ |
|---|
| 42 | // special placeholder that will be "eaten" |
|---|
| 43 | // by appending to the tail |
|---|
| 44 | </pre> |
|---|
| 45 | </div> |
|---|
| 46 | <div> |
|---|
| 47 | <i>Sequences</i> are extremely efficient. Element access speed approaches |
|---|
| 48 | random access--even with <i>seqs</i> of up to <i>256</i> elements. This |
|---|
| 49 | is because element access (among other things) is implemented iteratively |
|---|
| 50 | rather than recursively. Therefore, elements can be accessed at extremely |
|---|
| 51 | high indices even on preprocessors with low maximum expansion depths. |
|---|
| 52 | </div> |
|---|
| 53 | <div> |
|---|
| 54 | Elements of a <i>seq</i> can be extracted with <b>BOOST_PP_SEQ_ELEM</b>. |
|---|
| 55 | </div> |
|---|
| 56 | <h4> |
|---|
| 57 | Primitives |
|---|
| 58 | </h4> |
|---|
| 59 | <ul> |
|---|
| 60 | <li> |
|---|
| 61 | <a href="../ref/seq_elem.html">BOOST_PP_SEQ_ELEM</a></li> |
|---|
| 62 | </ul> |
|---|
| 63 | </body> |
|---|
| 64 | </html> |
|---|
