diff options
Diffstat (limited to 'array.c')
| -rw-r--r-- | array.c | 7061 |
1 files changed, 4506 insertions, 2555 deletions
@@ -10,22 +10,66 @@ Copyright (C) 2000 Information-technology Promotion Agency, Japan **********************************************************************/ + +#include "debug_counter.h" +#include "id.h" +#include "internal.h" +#include "internal/array.h" +#include "internal/compar.h" +#include "internal/enum.h" +#include "internal/gc.h" +#include "internal/hash.h" +#include "internal/numeric.h" +#include "internal/object.h" +#include "internal/proc.h" +#include "internal/rational.h" +#include "internal/vm.h" +#include "probes.h" #include "ruby/encoding.h" -#include "ruby/util.h" #include "ruby/st.h" -#include "probes.h" -#include "id.h" -#include "debug_counter.h" +#include "ruby/util.h" #include "transient_heap.h" -#include "internal.h" +#include "builtin.h" #if !ARRAY_DEBUG +# undef NDEBUG # define NDEBUG #endif #include "ruby_assert.h" VALUE rb_cArray; +/* Flags of RArray + * + * 1: RARRAY_EMBED_FLAG + * The array is embedded (its contents follow the header, rather than + * being on a separately allocated buffer). + * 2: RARRAY_SHARED_FLAG (equal to ELTS_SHARED) + * The array is shared. The buffer this array points to is owned by + * another array (the shared root). + * if USE_RVARGC + * 3-9: RARRAY_EMBED_LEN + * The length of the array when RARRAY_EMBED_FLAG is set. + * else + * 3-4: RARRAY_EMBED_LEN + * The length of the array when RARRAY_EMBED_FLAG is set. + * endif + * 12: RARRAY_SHARED_ROOT_FLAG + * The array is a shared root that does reference counting. The buffer + * this array points to is owned by this array but may be pointed to + * by other arrays. + * Note: Frozen arrays may be a shared root without this flag being + * set. Frozen arrays do not have reference counting because + * they cannot be modified. Not updating the reference count + * improves copy-on-write performance. Their reference count is + * assumed to be infinity. + * 13: RARRAY_TRANSIENT_FLAG + * The buffer of the array is allocated on the transient heap. + * 14: RARRAY_PTR_IN_USE_FLAG + * The buffer of the array is in use. This is only used during + * debugging. + */ + /* for OPTIMIZED_CMP: */ #define id_cmp idCmp @@ -33,28 +77,13 @@ VALUE rb_cArray; #define ARY_MAX_SIZE (LONG_MAX / (int)sizeof(VALUE)) #define SMALL_ARRAY_LEN 16 +RBIMPL_ATTR_MAYBE_UNUSED() static int should_be_T_ARRAY(VALUE ary) { return RB_TYPE_P(ary, T_ARRAY); } -static int -should_not_be_shared_and_embedded(VALUE ary) -{ - return !FL_TEST((ary), ELTS_SHARED) || !FL_TEST((ary), RARRAY_EMBED_FLAG); -} - -#define ARY_SHARED_P(ary) \ - (assert(should_be_T_ARRAY((VALUE)(ary))), \ - assert(should_not_be_shared_and_embedded((VALUE)ary)), \ - FL_TEST_RAW((ary),ELTS_SHARED)!=0) - -#define ARY_EMBED_P(ary) \ - (assert(should_be_T_ARRAY((VALUE)(ary))), \ - assert(should_not_be_shared_and_embedded((VALUE)ary)), \ - FL_TEST_RAW((ary), RARRAY_EMBED_FLAG) != 0) - #define ARY_HEAP_PTR(a) (assert(!ARY_EMBED_P(a)), RARRAY(a)->as.heap.ptr) #define ARY_HEAP_LEN(a) (assert(!ARY_EMBED_P(a)), RARRAY(a)->as.heap.len) #define ARY_HEAP_CAPA(a) (assert(!ARY_EMBED_P(a)), assert(!ARY_SHARED_ROOT_P(a)), \ @@ -64,11 +93,11 @@ should_not_be_shared_and_embedded(VALUE ary) #define ARY_EMBED_LEN(a) \ (assert(ARY_EMBED_P(a)), \ (long)((RBASIC(a)->flags >> RARRAY_EMBED_LEN_SHIFT) & \ - (RARRAY_EMBED_LEN_MASK >> RARRAY_EMBED_LEN_SHIFT))) + (RARRAY_EMBED_LEN_MASK >> RARRAY_EMBED_LEN_SHIFT))) #define ARY_HEAP_SIZE(a) (assert(!ARY_EMBED_P(a)), assert(ARY_OWNS_HEAP_P(a)), ARY_CAPA(a) * sizeof(VALUE)) #define ARY_OWNS_HEAP_P(a) (assert(should_be_T_ARRAY((VALUE)(a))), \ - !FL_TEST_RAW((a), ELTS_SHARED|RARRAY_EMBED_FLAG)) + !FL_TEST_RAW((a), RARRAY_SHARED_FLAG|RARRAY_EMBED_FLAG)) #define FL_SET_EMBED(a) do { \ assert(!ARY_SHARED_P(a)); \ @@ -80,9 +109,9 @@ should_not_be_shared_and_embedded(VALUE ary) #define FL_UNSET_EMBED(ary) FL_UNSET((ary), RARRAY_EMBED_FLAG|RARRAY_EMBED_LEN_MASK) #define FL_SET_SHARED(ary) do { \ assert(!ARY_EMBED_P(ary)); \ - FL_SET((ary), ELTS_SHARED); \ + FL_SET((ary), RARRAY_SHARED_FLAG); \ } while (0) -#define FL_UNSET_SHARED(ary) FL_UNSET((ary), ELTS_SHARED) +#define FL_UNSET_SHARED(ary) FL_UNSET((ary), RARRAY_SHARED_FLAG) #define ARY_SET_PTR(ary, p) do { \ assert(!ARY_EMBED_P(ary)); \ @@ -92,7 +121,6 @@ should_not_be_shared_and_embedded(VALUE ary) #define ARY_SET_EMBED_LEN(ary, n) do { \ long tmp_n = (n); \ assert(ARY_EMBED_P(ary)); \ - assert(!OBJ_FROZEN(ary)); \ RBASIC(ary)->flags &= ~RARRAY_EMBED_LEN_MASK; \ RBASIC(ary)->flags |= (tmp_n) << RARRAY_EMBED_LEN_SHIFT; \ } while (0) @@ -124,7 +152,7 @@ should_not_be_shared_and_embedded(VALUE ary) } \ } while (0) -#define ARY_CAPA(ary) (ARY_EMBED_P(ary) ? RARRAY_EMBED_LEN_MAX : \ +#define ARY_CAPA(ary) (ARY_EMBED_P(ary) ? ary_embed_capa(ary) : \ ARY_SHARED_ROOT_P(ary) ? RARRAY_LEN(ary) : ARY_HEAP_CAPA(ary)) #define ARY_SET_CAPA(ary, n) do { \ assert(!ARY_EMBED_P(ary)); \ @@ -133,26 +161,25 @@ should_not_be_shared_and_embedded(VALUE ary) RARRAY(ary)->as.heap.aux.capa = (n); \ } while (0) -#define ARY_SHARED_ROOT(ary) (assert(ARY_SHARED_P(ary)), RARRAY(ary)->as.heap.aux.shared_root) #define ARY_SET_SHARED(ary, value) do { \ const VALUE _ary_ = (ary); \ const VALUE _value_ = (value); \ assert(!ARY_EMBED_P(_ary_)); \ assert(ARY_SHARED_P(_ary_)); \ - assert(ARY_SHARED_ROOT_P(_value_)); \ + assert(!OBJ_FROZEN(_ary_)); \ + assert(ARY_SHARED_ROOT_P(_value_) || OBJ_FROZEN(_value_)); \ RB_OBJ_WRITE(_ary_, &RARRAY(_ary_)->as.heap.aux.shared_root, _value_); \ } while (0) -#define RARRAY_SHARED_ROOT_FLAG FL_USER5 -#define ARY_SHARED_ROOT_P(ary) (assert(should_be_T_ARRAY((VALUE)(ary))), \ - FL_TEST_RAW((ary), RARRAY_SHARED_ROOT_FLAG)) -#define ARY_SHARED_ROOT_REFCNT(ary) \ - (assert(ARY_SHARED_ROOT_P(ary)), RARRAY(ary)->as.heap.aux.capa) -#define ARY_SHARED_ROOT_OCCUPIED(ary) (ARY_SHARED_ROOT_REFCNT(ary) == 1) + +#define ARY_SHARED_ROOT_OCCUPIED(ary) (!OBJ_FROZEN(ary) && ARY_SHARED_ROOT_REFCNT(ary) == 1) #define ARY_SET_SHARED_ROOT_REFCNT(ary, value) do { \ assert(ARY_SHARED_ROOT_P(ary)); \ + assert(!OBJ_FROZEN(ary)); \ + assert((value) >= 0); \ RARRAY(ary)->as.heap.aux.capa = (value); \ } while (0) #define FL_SET_SHARED_ROOT(ary) do { \ + assert(!OBJ_FROZEN(ary)); \ assert(!ARY_EMBED_P(ary)); \ assert(!RARRAY_TRANSIENT_P(ary)); \ FL_SET((ary), RARRAY_SHARED_ROOT_FLAG); \ @@ -168,6 +195,65 @@ ARY_SET(VALUE a, long i, VALUE v) } #undef RARRAY_ASET +static long +ary_embed_capa(VALUE ary) +{ +#if USE_RVARGC + size_t size = rb_gc_obj_slot_size(ary) - offsetof(struct RArray, as.ary); + assert(size % sizeof(VALUE) == 0); + return size / sizeof(VALUE); +#else + return RARRAY_EMBED_LEN_MAX; +#endif +} + +static size_t +ary_embed_size(long capa) +{ + return offsetof(struct RArray, as.ary) + (sizeof(VALUE) * capa); +} + +static bool +ary_embeddable_p(long capa) +{ +#if USE_RVARGC + return rb_gc_size_allocatable_p(ary_embed_size(capa)); +#else + return capa <= RARRAY_EMBED_LEN_MAX; +#endif +} + +bool +rb_ary_embeddable_p(VALUE ary) +{ + /* An array cannot be turned embeddable when the array is: + * - Shared root: other objects may point to the buffer of this array + * so we cannot make it embedded. + * - Frozen: this array may also be a shared root without the shared root + * flag. + * - Shared: we don't want to re-embed an array that points to a shared + * root (to save memory). + */ + return !(ARY_SHARED_ROOT_P(ary) || OBJ_FROZEN(ary) || ARY_SHARED_P(ary)); +} + +size_t +rb_ary_size_as_embedded(VALUE ary) +{ + size_t real_size; + + if (ARY_EMBED_P(ary)) { + real_size = ary_embed_size(ARY_EMBED_LEN(ary)); + } + else if (rb_ary_embeddable_p(ary)) { + real_size = ary_embed_size(ARY_HEAP_CAPA(ary)); + } + else { + real_size = sizeof(struct RArray); + } + return real_size; +} + #if ARRAY_DEBUG #define ary_verify(ary) ary_verify_(ary, __FILE__, __LINE__) @@ -177,19 +263,19 @@ ary_verify_(VALUE ary, const char *file, int line) { assert(RB_TYPE_P(ary, T_ARRAY)); - if (FL_TEST(ary, ELTS_SHARED)) { - VALUE root = RARRAY(ary)->as.heap.aux.shared_root; + if (ARY_SHARED_P(ary)) { + VALUE root = ARY_SHARED_ROOT(ary); const VALUE *ptr = ARY_HEAP_PTR(ary); const VALUE *root_ptr = RARRAY_CONST_PTR_TRANSIENT(root); long len = ARY_HEAP_LEN(ary), root_len = RARRAY_LEN(root); - assert(FL_TEST(root, RARRAY_SHARED_ROOT_FLAG)); + assert(ARY_SHARED_ROOT_P(root) || OBJ_FROZEN(root)); assert(root_ptr <= ptr && ptr + len <= root_ptr + root_len); ary_verify(root); } else if (ARY_EMBED_P(ary)) { assert(!RARRAY_TRANSIENT_P(ary)); assert(!ARY_SHARED_P(ary)); - assert(RARRAY_LEN(ary) <= RARRAY_EMBED_LEN_MAX); + assert(RARRAY_LEN(ary) <= ary_embed_capa(ary)); } else { #if 1 @@ -204,9 +290,11 @@ ary_verify_(VALUE ary, const char *file, int line) #endif } +#if USE_TRANSIENT_HEAP if (RARRAY_TRANSIENT_P(ary)) { assert(rb_transient_heap_managed_ptr_p(RARRAY_CONST_PTR_TRANSIENT(ary))); } +#endif rb_transient_heap_verify(); @@ -240,10 +328,10 @@ rb_ary_ptr_use_end(VALUE ary) } void -rb_mem_clear(register VALUE *mem, register long size) +rb_mem_clear(VALUE *mem, long size) { while (size--) { - *mem++ = Qnil; + *mem++ = Qnil; } } @@ -251,7 +339,7 @@ static void ary_mem_clear(VALUE ary, long beg, long size) { RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - rb_mem_clear(ptr + beg, size); + rb_mem_clear(ptr + beg, size); }); } @@ -259,7 +347,7 @@ static inline void memfill(register VALUE *mem, register long size, register VALUE val) { while (size--) { - *mem++ = val; + *mem++ = val; } } @@ -267,8 +355,8 @@ static void ary_memfill(VALUE ary, long beg, long size, VALUE val) { RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - memfill(ptr + beg, size, val); - RB_OBJ_WRITTEN(ary, Qundef, val); + memfill(ptr + beg, size, val); + RB_OBJ_WRITTEN(ary, Qundef, val); }); } @@ -337,14 +425,16 @@ ary_heap_free(VALUE ary) } } -static void +static size_t ary_heap_realloc(VALUE ary, size_t new_capa) { + size_t alloc_capa = new_capa; size_t old_capa = ARY_HEAP_CAPA(ary); if (RARRAY_TRANSIENT_P(ary)) { if (new_capa <= old_capa) { /* do nothing */ + alloc_capa = old_capa; } else { VALUE *new_ptr = rb_transient_heap_alloc(ary, sizeof(VALUE) * new_capa); @@ -362,6 +452,8 @@ ary_heap_realloc(VALUE ary, size_t new_capa) SIZED_REALLOC_N(RARRAY(ary)->as.heap.ptr, VALUE, new_capa, old_capa); } ary_verify(ary); + + return alloc_capa; } #if USE_TRANSIENT_HEAP @@ -369,14 +461,11 @@ static inline void rb_ary_transient_heap_evacuate_(VALUE ary, int transient, int promote) { if (transient) { + assert(!ARY_SHARED_ROOT_P(ary)); + VALUE *new_ptr; const VALUE *old_ptr = ARY_HEAP_PTR(ary); long capa = ARY_HEAP_CAPA(ary); - long len = ARY_HEAP_LEN(ary); - - if (ARY_SHARED_ROOT_P(ary)) { - capa = len; - } assert(ARY_OWNS_HEAP_P(ary)); assert(RARRAY_TRANSIENT_P(ary)); @@ -418,6 +507,27 @@ rb_ary_detransient(VALUE ary) } #endif +void +rb_ary_make_embedded(VALUE ary) +{ + assert(rb_ary_embeddable_p(ary)); + if (!ARY_EMBED_P(ary)) { + const VALUE *buf = ARY_HEAP_PTR(ary); + long len = ARY_HEAP_LEN(ary); + bool was_transient = RARRAY_TRANSIENT_P(ary); + + // FL_SET_EMBED also unsets the transient flag + FL_SET_EMBED(ary); + ARY_SET_EMBED_LEN(ary, len); + + MEMCPY((void *)ARY_EMBED_PTR(ary), (void *)buf, VALUE, len); + + if (!was_transient) { + ary_heap_free_ptr(ary, buf, len * sizeof(VALUE)); + } + } +} + static void ary_resize_capa(VALUE ary, long capacity) { @@ -425,7 +535,8 @@ ary_resize_capa(VALUE ary, long capacity) assert(!OBJ_FROZEN(ary)); assert(!ARY_SHARED_P(ary)); - if (capacity > RARRAY_EMBED_LEN_MAX) { + if (capacity > ary_embed_capa(ary)) { + size_t new_capa = capacity; if (ARY_EMBED_P(ary)) { long len = ARY_EMBED_LEN(ary); VALUE *ptr = ary_heap_alloc(ary, capacity); @@ -436,9 +547,9 @@ ary_resize_capa(VALUE ary, long capacity) ARY_SET_HEAP_LEN(ary, len); } else { - ary_heap_realloc(ary, capacity); + new_capa = ary_heap_realloc(ary, capacity); } - ARY_SET_CAPA(ary, capacity); + ARY_SET_CAPA(ary, new_capa); } else { if (!ARY_EMBED_P(ary)) { @@ -476,10 +587,10 @@ ary_double_capa(VALUE ary, long min) long new_capa = ARY_CAPA(ary) / 2; if (new_capa < ARY_DEFAULT_SIZE) { - new_capa = ARY_DEFAULT_SIZE; + new_capa = ARY_DEFAULT_SIZE; } if (new_capa >= ARY_MAX_SIZE - min) { - new_capa = (ARY_MAX_SIZE - min) / 2; + new_capa = (ARY_MAX_SIZE - min) / 2; } new_capa += min; ary_resize_capa(ary, new_capa); @@ -490,39 +601,40 @@ ary_double_capa(VALUE ary, long min) static void rb_ary_decrement_share(VALUE shared_root) { - if (shared_root) { - long num = ARY_SHARED_ROOT_REFCNT(shared_root) - 1; - if (num == 0) { - rb_ary_free(shared_root); - rb_gc_force_recycle(shared_root); - } - else if (num > 0) { - ARY_SET_SHARED_ROOT_REFCNT(shared_root, num); - } + if (!OBJ_FROZEN(shared_root)) { + long num = ARY_SHARED_ROOT_REFCNT(shared_root); + ARY_SET_SHARED_ROOT_REFCNT(shared_root, num - 1); } } static void rb_ary_unshare(VALUE ary) { - VALUE shared_root = RARRAY(ary)->as.heap.aux.shared_root; + VALUE shared_root = ARY_SHARED_ROOT(ary); rb_ary_decrement_share(shared_root); FL_UNSET_SHARED(ary); } -static inline void -rb_ary_unshare_safe(VALUE ary) +static void +rb_ary_reset(VALUE ary) { - if (ARY_SHARED_P(ary) && !ARY_EMBED_P(ary)) { - rb_ary_unshare(ary); + if (ARY_OWNS_HEAP_P(ary)) { + ary_heap_free(ary); + } + else if (ARY_SHARED_P(ary)) { + rb_ary_unshare(ary); } + + FL_SET_EMBED(ary); + ARY_SET_EMBED_LEN(ary, 0); } static VALUE rb_ary_increment_share(VALUE shared_root) { - long num = ARY_SHARED_ROOT_REFCNT(shared_root); - if (num >= 0) { + if (!OBJ_FROZEN(shared_root)) { + long num = ARY_SHARED_ROOT_REFCNT(shared_root); + assert(num >= 0); ARY_SET_SHARED_ROOT_REFCNT(shared_root, num + 1); } return shared_root; @@ -545,34 +657,33 @@ rb_ary_modify_check(VALUE ary) } void -rb_ary_modify(VALUE ary) +rb_ary_cancel_sharing(VALUE ary) { - rb_ary_modify_check(ary); if (ARY_SHARED_P(ary)) { - long shared_len, len = RARRAY_LEN(ary); + long shared_len, len = RARRAY_LEN(ary); VALUE shared_root = ARY_SHARED_ROOT(ary); ary_verify(shared_root); - if (len <= RARRAY_EMBED_LEN_MAX) { - const VALUE *ptr = ARY_HEAP_PTR(ary); + if (len <= ary_embed_capa(ary)) { + const VALUE *ptr = ARY_HEAP_PTR(ary); FL_UNSET_SHARED(ary); FL_SET_EMBED(ary); - MEMCPY((VALUE *)ARY_EMBED_PTR(ary), ptr, VALUE, len); + MEMCPY((VALUE *)ARY_EMBED_PTR(ary), ptr, VALUE, len); rb_ary_decrement_share(shared_root); ARY_SET_EMBED_LEN(ary, len); } else if (ARY_SHARED_ROOT_OCCUPIED(shared_root) && len > ((shared_len = RARRAY_LEN(shared_root))>>1)) { long shift = RARRAY_CONST_PTR_TRANSIENT(ary) - RARRAY_CONST_PTR_TRANSIENT(shared_root); - FL_UNSET_SHARED(ary); + FL_UNSET_SHARED(ary); ARY_SET_PTR(ary, RARRAY_CONST_PTR_TRANSIENT(shared_root)); - ARY_SET_CAPA(ary, shared_len); + ARY_SET_CAPA(ary, shared_len); RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - MEMMOVE(ptr, ptr+shift, VALUE, len); - }); + MEMMOVE(ptr, ptr+shift, VALUE, len); + }); FL_SET_EMBED(shared_root); rb_ary_decrement_share(shared_root); - } + } else { VALUE *ptr = ary_heap_alloc(ary, len); MEMCPY(ptr, ARY_HEAP_PTR(ary), VALUE, len); @@ -581,11 +692,18 @@ rb_ary_modify(VALUE ary) ARY_SET_PTR(ary, ptr); } - rb_gc_writebarrier_remember(ary); + rb_gc_writebarrier_remember(ary); } ary_verify(ary); } +void +rb_ary_modify(VALUE ary) +{ + rb_ary_modify_check(ary); + rb_ary_cancel_sharing(ary); +} + static VALUE ary_ensure_room_for_push(VALUE ary, long add_len) { @@ -594,40 +712,40 @@ ary_ensure_room_for_push(VALUE ary, long add_len) long capa; if (old_len > ARY_MAX_SIZE - add_len) { - rb_raise(rb_eIndexError, "index %ld too big", new_len); + rb_raise(rb_eIndexError, "index %ld too big", new_len); } if (ARY_SHARED_P(ary)) { - if (new_len > RARRAY_EMBED_LEN_MAX) { + if (new_len > ary_embed_capa(ary)) { VALUE shared_root = ARY_SHARED_ROOT(ary); if (ARY_SHARED_ROOT_OCCUPIED(shared_root)) { if (ARY_HEAP_PTR(ary) - RARRAY_CONST_PTR_TRANSIENT(shared_root) + new_len <= RARRAY_LEN(shared_root)) { - rb_ary_modify_check(ary); + rb_ary_modify_check(ary); ary_verify(ary); ary_verify(shared_root); return shared_root; - } - else { - /* if array is shared, then it is likely it participate in push/shift pattern */ - rb_ary_modify(ary); - capa = ARY_CAPA(ary); - if (new_len > capa - (capa >> 6)) { - ary_double_capa(ary, new_len); - } + } + else { + /* if array is shared, then it is likely it participate in push/shift pattern */ + rb_ary_modify(ary); + capa = ARY_CAPA(ary); + if (new_len > capa - (capa >> 6)) { + ary_double_capa(ary, new_len); + } ary_verify(ary); - return ary; - } - } - } + return ary; + } + } + } ary_verify(ary); rb_ary_modify(ary); } else { - rb_ary_modify_check(ary); + rb_ary_modify_check(ary); } capa = ARY_CAPA(ary); if (new_len > capa) { - ary_double_capa(ary, new_len); + ary_double_capa(ary, new_len); } ary_verify(ary); @@ -636,12 +754,16 @@ ary_ensure_room_for_push(VALUE ary, long add_len) /* * call-seq: - * ary.freeze -> ary + * array.freeze -> self + * + * Freezes +self+; returns +self+: * - * Calls Object#freeze on +ary+ to prevent any further - * modification. A RuntimeError will be raised if a modification - * attempt is made. + * a = [] + * a.frozen? # => false + * a.freeze + * a.frozen? # => true * + * An attempt to modify a frozen \Array raises FrozenError. */ VALUE @@ -661,18 +783,25 @@ VALUE rb_ary_shared_with_p(VALUE ary1, VALUE ary2) { if (!ARY_EMBED_P(ary1) && ARY_SHARED_P(ary1) && - !ARY_EMBED_P(ary2) && ARY_SHARED_P(ary2) && - RARRAY(ary1)->as.heap.aux.shared_root == RARRAY(ary2)->as.heap.aux.shared_root && - RARRAY(ary1)->as.heap.len == RARRAY(ary2)->as.heap.len) { - return Qtrue; + !ARY_EMBED_P(ary2) && ARY_SHARED_P(ary2) && + ARY_SHARED_ROOT(ary1) == ARY_SHARED_ROOT(ary2) && + ARY_HEAP_LEN(ary1) == ARY_HEAP_LEN(ary2)) { + return Qtrue; } return Qfalse; } static VALUE -ary_alloc(VALUE klass) +ary_alloc_embed(VALUE klass, long capa) { - NEWOBJ_OF(ary, struct RArray, klass, T_ARRAY | RARRAY_EMBED_FLAG | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0)); + size_t size = ary_embed_size(capa); + assert(rb_gc_size_allocatable_p(size)); +#if !USE_RVARGC + assert(size <= sizeof(struct RArray)); +#endif + RVARGC_NEWOBJ_OF(ary, struct RArray, klass, + T_ARRAY | RARRAY_EMBED_FLAG | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0), + size); /* Created array is: * FL_SET_EMBED((VALUE)ary); * ARY_SET_EMBED_LEN((VALUE)ary, 0); @@ -681,10 +810,19 @@ ary_alloc(VALUE klass) } static VALUE +ary_alloc_heap(VALUE klass) +{ + RVARGC_NEWOBJ_OF(ary, struct RArray, klass, + T_ARRAY | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0), + sizeof(struct RArray)); + return (VALUE)ary; +} + +static VALUE empty_ary_alloc(VALUE klass) { RUBY_DTRACE_CREATE_HOOK(ARRAY, 0); - return ary_alloc(klass); + return ary_alloc_embed(klass, 0); } static VALUE @@ -693,20 +831,24 @@ ary_new(VALUE klass, long capa) VALUE ary,*ptr; if (capa < 0) { - rb_raise(rb_eArgError, "negative array size (or size too big)"); + rb_raise(rb_eArgError, "negative array size (or size too big)"); } if (capa > ARY_MAX_SIZE) { - rb_raise(rb_eArgError, "array size too big"); + rb_raise(rb_eArgError, "array size too big"); } RUBY_DTRACE_CREATE_HOOK(ARRAY, capa); - ary = ary_alloc(klass); - if (capa > RARRAY_EMBED_LEN_MAX) { + if (ary_embeddable_p(capa)) { + ary = ary_alloc_embed(klass, capa); + } + else { + ary = ary_alloc_heap(klass); + ARY_SET_CAPA(ary, capa); + assert(!ARY_EMBED_P(ary)); + ptr = ary_heap_alloc(ary, capa); - FL_UNSET_EMBED(ary); ARY_SET_PTR(ary, ptr); - ARY_SET_CAPA(ary, capa); ARY_SET_HEAP_LEN(ary, 0); } @@ -722,7 +864,7 @@ rb_ary_new_capa(long capa) VALUE rb_ary_new(void) { - return rb_ary_new2(RARRAY_EMBED_LEN_MAX); + return rb_ary_new_capa(0); } VALUE @@ -736,7 +878,7 @@ VALUE va_start(ar, n); for (i=0; i<n; i++) { - ARY_SET(ary, i, va_arg(ar, VALUE)); + ARY_SET(ary, i, va_arg(ar, VALUE)); } va_end(ar); @@ -751,8 +893,8 @@ rb_ary_tmp_new_from_values(VALUE klass, long n, const VALUE *elts) ary = ary_new(klass, n); if (n > 0 && elts) { - ary_memcpy(ary, 0, n, elts); - ARY_SET_LEN(ary, n); + ary_memcpy(ary, 0, n, elts); + ARY_SET_LEN(ary, n); } return ary; @@ -764,8 +906,79 @@ rb_ary_new_from_values(long n, const VALUE *elts) return rb_ary_tmp_new_from_values(rb_cArray, n, elts); } +static VALUE +ec_ary_alloc_embed(rb_execution_context_t *ec, VALUE klass, long capa) +{ + size_t size = ary_embed_size(capa); + assert(rb_gc_size_allocatable_p(size)); +#if !USE_RVARGC + assert(size <= sizeof(struct RArray)); +#endif + RB_RVARGC_EC_NEWOBJ_OF(ec, ary, struct RArray, klass, + T_ARRAY | RARRAY_EMBED_FLAG | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0), + size); + /* Created array is: + * FL_SET_EMBED((VALUE)ary); + * ARY_SET_EMBED_LEN((VALUE)ary, 0); + */ + return (VALUE)ary; +} + +static VALUE +ec_ary_alloc_heap(rb_execution_context_t *ec, VALUE klass) +{ + RB_RVARGC_EC_NEWOBJ_OF(ec, ary, struct RArray, klass, + T_ARRAY | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0), + sizeof(struct RArray)); + return (VALUE)ary; +} + +static VALUE +ec_ary_new(rb_execution_context_t *ec, VALUE klass, long capa) +{ + VALUE ary,*ptr; + + if (capa < 0) { + rb_raise(rb_eArgError, "negative array size (or size too big)"); + } + if (capa > ARY_MAX_SIZE) { + rb_raise(rb_eArgError, "array size too big"); + } + + RUBY_DTRACE_CREATE_HOOK(ARRAY, capa); + + if (ary_embeddable_p(capa)) { + ary = ec_ary_alloc_embed(ec, klass, capa); + } + else { + ary = ec_ary_alloc_heap(ec, klass); + ARY_SET_CAPA(ary, capa); + assert(!ARY_EMBED_P(ary)); + + ptr = ary_heap_alloc(ary, capa); + ARY_SET_PTR(ary, ptr); + ARY_SET_HEAP_LEN(ary, 0); + } + + return ary; +} + +VALUE +rb_ec_ary_new_from_values(rb_execution_context_t *ec, long n, const VALUE *elts) +{ + VALUE ary; + + ary = ec_ary_new(ec, rb_cArray, n); + if (n > 0 && elts) { + ary_memcpy(ary, 0, n, elts); + ARY_SET_LEN(ary, n); + } + + return ary; +} + VALUE -rb_ary_tmp_new(long capa) +rb_ary_hidden_new(long capa) { VALUE ary = ary_new(0, capa); rb_ary_transient_heap_evacuate(ary, TRUE); @@ -773,12 +986,11 @@ rb_ary_tmp_new(long capa) } VALUE -rb_ary_tmp_new_fill(long capa) +rb_ary_hidden_new_fill(long capa) { - VALUE ary = ary_new(0, capa); + VALUE ary = rb_ary_hidden_new(capa); ary_memfill(ary, 0, capa, Qnil); ARY_SET_LEN(ary, capa); - rb_ary_transient_heap_evacuate(ary, TRUE); return ary; } @@ -786,6 +998,12 @@ void rb_ary_free(VALUE ary) { if (ARY_OWNS_HEAP_P(ary)) { + if (USE_DEBUG_COUNTER && + !ARY_SHARED_ROOT_P(ary) && + ARY_HEAP_CAPA(ary) > RARRAY_LEN(ary)) { + RB_DEBUG_COUNTER_INC(obj_ary_extracapa); + } + if (RARRAY_TRANSIENT_P(ary)) { RB_DEBUG_COUNTER_INC(obj_ary_transient); } @@ -810,63 +1028,69 @@ RUBY_FUNC_EXPORTED size_t rb_ary_memsize(VALUE ary) { if (ARY_OWNS_HEAP_P(ary)) { - return ARY_CAPA(ary) * sizeof(VALUE); + return ARY_CAPA(ary) * sizeof(VALUE); } else { - return 0; + return 0; } } -static inline void -ary_discard(VALUE ary) -{ - rb_ary_free(ary); - RBASIC(ary)->flags |= RARRAY_EMBED_FLAG; - RBASIC(ary)->flags &= ~(RARRAY_EMBED_LEN_MASK | RARRAY_TRANSIENT_FLAG); -} - static VALUE ary_make_shared(VALUE ary) { - assert(!ARY_EMBED_P(ary)); + assert(USE_RVARGC || !ARY_EMBED_P(ary)); ary_verify(ary); if (ARY_SHARED_P(ary)) { return ARY_SHARED_ROOT(ary); } else if (ARY_SHARED_ROOT_P(ary)) { - return ary; + return ary; } else if (OBJ_FROZEN(ary)) { - rb_ary_transient_heap_evacuate(ary, TRUE); - ary_shrink_capa(ary); - FL_SET_SHARED_ROOT(ary); - ARY_SET_SHARED_ROOT_REFCNT(ary, 1); - return ary; + if (!ARY_EMBED_P(ary)) { + rb_ary_transient_heap_evacuate(ary, TRUE); + ary_shrink_capa(ary); + } + return ary; } else { - long capa = ARY_CAPA(ary), len = RARRAY_LEN(ary); - const VALUE *ptr; - NEWOBJ_OF(shared, struct RArray, 0, T_ARRAY | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0)); - rb_ary_transient_heap_evacuate(ary, TRUE); - ptr = ARY_HEAP_PTR(ary); - - FL_UNSET_EMBED(shared); - ARY_SET_LEN((VALUE)shared, capa); - ARY_SET_PTR((VALUE)shared, ptr); - ary_mem_clear((VALUE)shared, len, capa - len); - FL_SET_SHARED_ROOT(shared); - ARY_SET_SHARED_ROOT_REFCNT((VALUE)shared, 1); - FL_SET_SHARED(ary); + + long capa = ARY_CAPA(ary); + long len = RARRAY_LEN(ary); + + /* Shared roots cannot be embedded because the reference count + * (refcnt) is stored in as.heap.aux.capa. */ + VALUE shared = ary_alloc_heap(0); + FL_SET_SHARED_ROOT(shared); + + if (ARY_EMBED_P(ary)) { + /* Cannot use ary_heap_alloc because we don't want to allocate + * on the transient heap. */ + VALUE *ptr = ALLOC_N(VALUE, capa); + ARY_SET_PTR(shared, ptr); + ary_memcpy(shared, 0, len, RARRAY_PTR(ary)); + + FL_UNSET_EMBED(ary); + ARY_SET_HEAP_LEN(ary, len); + ARY_SET_PTR(ary, ptr); + } + else { + ARY_SET_PTR(shared, RARRAY_PTR(ary)); + } + + ARY_SET_LEN(shared, capa); + ary_mem_clear(shared, len, capa - len); + ARY_SET_SHARED_ROOT_REFCNT(shared, 1); + FL_SET_SHARED(ary); RB_DEBUG_COUNTER_INC(obj_ary_shared_create); - ARY_SET_SHARED(ary, (VALUE)shared); - OBJ_FREEZE(shared); + ARY_SET_SHARED(ary, shared); - ary_verify((VALUE)shared); + ary_verify(shared); ary_verify(ary); - return (VALUE)shared; + return shared; } } @@ -875,8 +1099,10 @@ ary_make_substitution(VALUE ary) { long len = RARRAY_LEN(ary); - if (len <= RARRAY_EMBED_LEN_MAX) { - VALUE subst = rb_ary_new2(len); + if (ary_embeddable_p(len)) { + VALUE subst = rb_ary_new_capa(len); + assert(ARY_EMBED_P(subst)); + ary_memcpy(subst, 0, len, RARRAY_CONST_PTR_TRANSIENT(ary)); ARY_SET_EMBED_LEN(subst, len); return subst; @@ -911,23 +1137,24 @@ rb_check_to_array(VALUE ary) return rb_check_convert_type_with_id(ary, T_ARRAY, "Array", idTo_a); } +VALUE +rb_to_array(VALUE ary) +{ + return rb_convert_type_with_id(ary, T_ARRAY, "Array", idTo_a); +} + /* * call-seq: - * Array.try_convert(obj) -> array or nil + * Array.try_convert(object) -> object, new_array, or nil * - * Tries to convert +obj+ into an array, using the +to_ary+ method. Returns - * the converted array or +nil+ if +obj+ cannot be converted. - * This method can be used to check if an argument is an array. + * If +object+ is an \Array object, returns +object+. * - * Array.try_convert([1]) #=> [1] - * Array.try_convert("1") #=> nil + * Otherwise if +object+ responds to <tt>:to_ary</tt>, + * calls <tt>object.to_ary</tt> and returns the result. * - * if tmp = Array.try_convert(arg) - * # the argument is an array - * elsif tmp = String.try_convert(arg) - * # the argument is a string - * end + * Returns +nil+ if +object+ does not respond to <tt>:to_ary</tt> * + * Raises an exception unless <tt>object.to_ary</tt> returns an \Array object. */ static VALUE @@ -936,60 +1163,76 @@ rb_ary_s_try_convert(VALUE dummy, VALUE ary) return rb_check_array_type(ary); } +/* :nodoc: */ +static VALUE +rb_ary_s_new(int argc, VALUE *argv, VALUE klass) +{ + VALUE ary; + + if (klass == rb_cArray) { + long size = 0; + if (argc > 0 && FIXNUM_P(argv[0])) { + size = FIX2LONG(argv[0]); + if (size < 0) size = 0; + } + + ary = ary_new(klass, size); + + rb_obj_call_init_kw(ary, argc, argv, RB_PASS_CALLED_KEYWORDS); + } + else { + ary = rb_class_new_instance_pass_kw(argc, argv, klass); + } + + return ary; +} + /* * call-seq: - * Array.new(size=0, default=nil) - * Array.new(array) - * Array.new(size) {|index| block } - * - * Returns a new array. - * - * In the first form, if no arguments are sent, the new array will be empty. - * When a +size+ and an optional +default+ are sent, an array is created with - * +size+ copies of +default+. Take notice that all elements will reference the - * same object +default+. + * Array.new -> new_empty_array + * Array.new(array) -> new_array + * Array.new(size) -> new_array + * Array.new(size, default_value) -> new_array + * Array.new(size) {|index| ... } -> new_array * - * The second form creates a copy of the array passed as a parameter (the - * array is generated by calling to_ary on the parameter). + * Returns a new \Array. * - * first_array = ["Matz", "Guido"] + * With no block and no arguments, returns a new empty \Array object. * - * second_array = Array.new(first_array) #=> ["Matz", "Guido"] + * With no block and a single \Array argument +array+, + * returns a new \Array formed from +array+: * - * first_array.equal? second_array #=> false + * a = Array.new([:foo, 'bar', 2]) + * a.class # => Array + * a # => [:foo, "bar", 2] * - * In the last form, an array of the given size is created. Each element in - * this array is created by passing the element's index to the given block - * and storing the return value. + * With no block and a single \Integer argument +size+, + * returns a new \Array of the given size + * whose elements are all +nil+: * - * Array.new(3) {|index| index ** 2} - * # => [0, 1, 4] + * a = Array.new(3) + * a # => [nil, nil, nil] * - * == Common gotchas + * With no block and arguments +size+ and +default_value+, + * returns an \Array of the given size; + * each element is that same +default_value+: * - * When sending the second parameter, the same object will be used as the - * value for all the array elements: + * a = Array.new(3, 'x') + * a # => ['x', 'x', 'x'] * - * a = Array.new(2, Hash.new) - * # => [{}, {}] + * With a block and argument +size+, + * returns an \Array of the given size; + * the block is called with each successive integer +index+; + * the element for that +index+ is the return value from the block: * - * a[0]['cat'] = 'feline' - * a # => [{"cat"=>"feline"}, {"cat"=>"feline"}] + * a = Array.new(3) {|index| "Element #{index}" } + * a # => ["Element 0", "Element 1", "Element 2"] * - * a[1]['cat'] = 'Felix' - * a # => [{"cat"=>"Felix"}, {"cat"=>"Felix"}] - * - * Since all the Array elements store the same hash, changes to one of them - * will affect them all. - * - * If multiple copies are what you want, you should use the block - * version which uses the result of that block each time an element - * of the array needs to be initialized: - * - * a = Array.new(2) {Hash.new} - * a[0]['cat'] = 'feline' - * a # => [{"cat"=>"feline"}, {}] + * Raises ArgumentError if +size+ is negative. * + * With a block and no argument, + * or a single argument +0+, + * ignores the block and returns a new empty \Array. */ static VALUE @@ -1000,51 +1243,48 @@ rb_ary_initialize(int argc, VALUE *argv, VALUE ary) rb_ary_modify(ary); if (argc == 0) { - if (ARY_OWNS_HEAP_P(ary) && ARY_HEAP_PTR(ary) != NULL) { - ary_heap_free(ary); - } - rb_ary_unshare_safe(ary); - FL_SET_EMBED(ary); - ARY_SET_EMBED_LEN(ary, 0); - if (rb_block_given_p()) { - rb_warning("given block not used"); - } - return ary; + rb_ary_reset(ary); + assert(ARY_EMBED_P(ary)); + assert(ARY_EMBED_LEN(ary) == 0); + if (rb_block_given_p()) { + rb_warning("given block not used"); + } + return ary; } rb_scan_args(argc, argv, "02", &size, &val); if (argc == 1 && !FIXNUM_P(size)) { - val = rb_check_array_type(size); - if (!NIL_P(val)) { - rb_ary_replace(ary, val); - return ary; - } + val = rb_check_array_type(size); + if (!NIL_P(val)) { + rb_ary_replace(ary, val); + return ary; + } } len = NUM2LONG(size); /* NUM2LONG() may call size.to_int, ary can be frozen, modified, etc */ if (len < 0) { - rb_raise(rb_eArgError, "negative array size"); + rb_raise(rb_eArgError, "negative array size"); } if (len > ARY_MAX_SIZE) { - rb_raise(rb_eArgError, "array size too big"); + rb_raise(rb_eArgError, "array size too big"); } /* recheck after argument conversion */ rb_ary_modify(ary); ary_resize_capa(ary, len); if (rb_block_given_p()) { - long i; + long i; - if (argc == 2) { - rb_warn("block supersedes default value argument"); - } - for (i=0; i<len; i++) { - rb_ary_store(ary, i, rb_yield(LONG2NUM(i))); - ARY_SET_LEN(ary, i + 1); - } + if (argc == 2) { + rb_warn("block supersedes default value argument"); + } + for (i=0; i<len; i++) { + rb_ary_store(ary, i, rb_yield(LONG2NUM(i))); + ARY_SET_LEN(ary, i + 1); + } } else { - ary_memfill(ary, 0, len, val); - ARY_SET_LEN(ary, len); + ary_memfill(ary, 0, len, val); + ARY_SET_LEN(ary, len); } return ary; } @@ -1075,26 +1315,26 @@ rb_ary_store(VALUE ary, long idx, VALUE val) long len = RARRAY_LEN(ary); if (idx < 0) { - idx += len; - if (idx < 0) { - rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", - idx - len, -len); - } + idx += len; + if (idx < 0) { + rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", + idx - len, -len); + } } else if (idx >= ARY_MAX_SIZE) { - rb_raise(rb_eIndexError, "index %ld too big", idx); + rb_raise(rb_eIndexError, "index %ld too big", idx); } rb_ary_modify(ary); if (idx >= ARY_CAPA(ary)) { - ary_double_capa(ary, idx); + ary_double_capa(ary, idx); } if (idx > len) { - ary_mem_clear(ary, len, idx - len + 1); + ary_mem_clear(ary, len, idx - len + 1); } if (idx >= len) { - ARY_SET_LEN(ary, idx + 1); + ARY_SET_LEN(ary, idx + 1); } ARY_SET(ary, idx, val); } @@ -1106,17 +1346,20 @@ ary_make_partial(VALUE ary, VALUE klass, long offset, long len) assert(len >= 0); assert(offset+len <= RARRAY_LEN(ary)); - if (len <= RARRAY_EMBED_LEN_MAX) { - VALUE result = ary_alloc(klass); + const size_t rarray_embed_capa_max = (sizeof(struct RArray) - offsetof(struct RArray, as.ary)) / sizeof(VALUE); + + if ((size_t)len <= rarray_embed_capa_max && ary_embeddable_p(len)) { + VALUE result = ary_alloc_embed(klass, len); ary_memcpy(result, 0, len, RARRAY_CONST_PTR_TRANSIENT(ary) + offset); ARY_SET_EMBED_LEN(result, len); return result; } else { - VALUE shared, result = ary_alloc(klass); - FL_UNSET_EMBED(result); + VALUE shared = ary_make_shared(ary); + + VALUE result = ary_alloc_heap(klass); + assert(!ARY_EMBED_P(result)); - shared = ary_make_shared(ary); ARY_SET_PTR(result, RARRAY_CONST_PTR_TRANSIENT(ary)); ARY_SET_LEN(result, RARRAY_LEN(ary)); rb_ary_set_shared(result, shared); @@ -1131,9 +1374,59 @@ ary_make_partial(VALUE ary, VALUE klass, long offset, long len) } static VALUE +ary_make_partial_step(VALUE ary, VALUE klass, long offset, long len, long step) +{ + assert(offset >= 0); + assert(len >= 0); + assert(offset+len <= RARRAY_LEN(ary)); + assert(step != 0); + + const VALUE *values = RARRAY_CONST_PTR_TRANSIENT(ary); + const long orig_len = len; + + if (step > 0 && step >= len) { + VALUE result = ary_new(klass, 1); + VALUE *ptr = (VALUE *)ARY_EMBED_PTR(result); + RB_OBJ_WRITE(result, ptr, values[offset]); + ARY_SET_EMBED_LEN(result, 1); + return result; + } + else if (step < 0 && step < -len) { + step = -len; + } + + long ustep = (step < 0) ? -step : step; + len = roomof(len, ustep); + + long i; + long j = offset + ((step > 0) ? 0 : (orig_len - 1)); + + VALUE result = ary_new(klass, len); + if (ARY_EMBED_P(result)) { + VALUE *ptr = (VALUE *)ARY_EMBED_PTR(result); + for (i = 0; i < len; ++i) { + RB_OBJ_WRITE(result, ptr+i, values[j]); + j += step; + } + ARY_SET_EMBED_LEN(result, len); + } + else { + RARRAY_PTR_USE_TRANSIENT(result, ptr, { + for (i = 0; i < len; ++i) { + RB_OBJ_WRITE(result, ptr+i, values[j]); + j += step; + } + }); + ARY_SET_LEN(result, len); + } + + return result; +} + +static VALUE ary_make_shared_copy(VALUE ary) { - return ary_make_partial(ary, rb_obj_class(ary), 0, RARRAY_LEN(ary)); + return ary_make_partial(ary, rb_cArray, 0, RARRAY_LEN(ary)); } enum ary_take_pos_flags @@ -1158,30 +1451,31 @@ ary_take_first_or_last(int argc, const VALUE *argv, VALUE ary, enum ary_take_pos n = NUM2LONG(argv[0]); len = RARRAY_LEN(ary); if (n > len) { - n = len; + n = len; } else if (n < 0) { - rb_raise(rb_eArgError, "negative array size"); + rb_raise(rb_eArgError, "negative array size"); } if (last) { - offset = len - n; + offset = len - n; } return ary_make_partial(ary, rb_cArray, offset, n); } /* * call-seq: - * ary << obj -> ary + * array << object -> self + * + * Appends +object+ to +self+; returns +self+: * - * Append---Pushes the given object on to the end of this array. This - * expression returns the array itself, so several appends - * may be chained together. + * a = [:foo, 'bar', 2] + * a << :baz # => [:foo, "bar", 2, :baz] * - * a = [ 1, 2 ] - * a << "c" << "d" << [ 3, 4 ] - * #=> [ 1, 2, "c", "d", [ 3, 4 ] ] - * a - * #=> [ 1, 2, "c", "d", [ 3, 4 ] ] + * Appends +object+ as one element, even if it is another \Array: + * + * a = [:foo, 'bar', 2] + * a1 = a << [3, 4] + * a1 # => [:foo, "bar", 2, [3, 4]] * */ @@ -1191,7 +1485,7 @@ rb_ary_push(VALUE ary, VALUE item) long idx = RARRAY_LEN((ary_verify(ary), ary)); VALUE target_ary = ary_ensure_room_for_push(ary, 1); RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - RB_OBJ_WRITE(target_ary, &ptr[idx], item); + RB_OBJ_WRITE(target_ary, &ptr[idx], item); }); ARY_SET_LEN(ary, idx + 1); ary_verify(ary); @@ -1210,19 +1504,24 @@ rb_ary_cat(VALUE ary, const VALUE *argv, long len) /* * call-seq: - * ary.push(obj, ...) -> ary - * ary.append(obj, ...) -> ary - * - * Append --- Pushes the given object(s) on to the end of this array. This - * expression returns the array itself, so several appends - * may be chained together. See also Array#pop for the opposite - * effect. - * - * a = [ "a", "b", "c" ] - * a.push("d", "e", "f") - * #=> ["a", "b", "c", "d", "e", "f"] - * [1, 2, 3].push(4).push(5) - * #=> [1, 2, 3, 4, 5] + * array.push(*objects) -> self + * + * Appends trailing elements. + * + * Appends each argument in +objects+ to +self+; returns +self+: + * + * a = [:foo, 'bar', 2] + * a.push(:baz, :bat) # => [:foo, "bar", 2, :baz, :bat] + * + * Appends each argument as one element, even if it is another \Array: + * + * a = [:foo, 'bar', 2] + * a1 = a.push([:baz, :bat], [:bam, :bad]) + * a1 # => [:foo, "bar", 2, [:baz, :bat], [:bam, :bad]] + * + * Array#append is an alias for Array#push. + * + * Related: #pop, #shift, #unshift. */ static VALUE @@ -1239,10 +1538,10 @@ rb_ary_pop(VALUE ary) n = RARRAY_LEN(ary); if (n == 0) return Qnil; if (ARY_OWNS_HEAP_P(ary) && - n * 3 < ARY_CAPA(ary) && - ARY_CAPA(ary) > ARY_DEFAULT_SIZE) + n * 3 < ARY_CAPA(ary) && + ARY_CAPA(ary) > ARY_DEFAULT_SIZE) { - ary_resize_capa(ary, n * 2); + ary_resize_capa(ary, n * 2); } --n; ARY_SET_LEN(ary, n); @@ -1252,20 +1551,33 @@ rb_ary_pop(VALUE ary) /* * call-seq: - * ary.pop -> obj or nil - * ary.pop(n) -> new_ary + * array.pop -> object or nil + * array.pop(n) -> new_array + * + * Removes and returns trailing elements. + * + * When no argument is given and +self+ is not empty, + * removes and returns the last element: + * + * a = [:foo, 'bar', 2] + * a.pop # => 2 + * a # => [:foo, "bar"] + * + * Returns +nil+ if the array is empty. * - * Removes the last element from +self+ and returns it, or - * +nil+ if the array is empty. + * When a non-negative \Integer argument +n+ is given and is in range, * - * If a number +n+ is given, returns an array of the last +n+ elements - * (or less) just like <code>array.slice!(-n, n)</code> does. See also - * Array#push for the opposite effect. + * removes and returns the last +n+ elements in a new \Array: + * a = [:foo, 'bar', 2] + * a.pop(2) # => ["bar", 2] * - * a = [ "a", "b", "c", "d" ] - * a.pop #=> "d" - * a.pop(2) #=> ["b", "c"] - * a #=> ["a"] + * If +n+ is positive and out of range, + * removes and returns all elements: + * + * a = [:foo, 'bar', 2] + * a.pop(50) # => [:foo, "bar", 2] + * + * Related: #push, #shift, #unshift. */ static VALUE @@ -1274,7 +1586,7 @@ rb_ary_pop_m(int argc, VALUE *argv, VALUE ary) VALUE result; if (argc == 0) { - return rb_ary_pop(ary); + return rb_ary_pop(ary); } rb_ary_modify_check(ary); @@ -1290,55 +1602,49 @@ rb_ary_shift(VALUE ary) VALUE top; long len = RARRAY_LEN(ary); - rb_ary_modify_check(ary); - if (len == 0) return Qnil; - top = RARRAY_AREF(ary, 0); - if (!ARY_SHARED_P(ary)) { - if (len < ARY_DEFAULT_SIZE) { - RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - MEMMOVE(ptr, ptr+1, VALUE, len-1); - }); /* WB: no new reference */ - ARY_INCREASE_LEN(ary, -1); - ary_verify(ary); - return top; - } - assert(!ARY_EMBED_P(ary)); /* ARY_EMBED_LEN_MAX < ARY_DEFAULT_SIZE */ - - ARY_SET(ary, 0, Qnil); - ary_make_shared(ary); - } - else if (ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))) { - RARRAY_PTR_USE_TRANSIENT(ary, ptr, ptr[0] = Qnil); + if (len == 0) { + rb_ary_modify_check(ary); + return Qnil; } - ARY_INCREASE_PTR(ary, 1); /* shift ptr */ - ARY_INCREASE_LEN(ary, -1); - ary_verify(ary); + top = RARRAY_AREF(ary, 0); + + rb_ary_behead(ary, 1); return top; } /* * call-seq: - * ary.shift -> obj or nil - * ary.shift(n) -> new_ary + * array.shift -> object or nil + * array.shift(n) -> new_array + * + * Removes and returns leading elements. + * + * When no argument is given, removes and returns the first element: + * + * a = [:foo, 'bar', 2] + * a.shift # => :foo + * a # => ['bar', 2] + * + * Returns +nil+ if +self+ is empty. * - * Removes the first element of +self+ and returns it (shifting all - * other elements down by one). Returns +nil+ if the array - * is empty. + * When positive \Integer argument +n+ is given, removes the first +n+ elements; + * returns those elements in a new \Array: * - * If a number +n+ is given, returns an array of the first +n+ elements - * (or less) just like <code>array.slice!(0, n)</code> does. With +ary+ - * containing only the remainder elements, not including what was shifted to - * +new_ary+. See also Array#unshift for the opposite effect. + * a = [:foo, 'bar', 2] + * a.shift(2) # => [:foo, 'bar'] + * a # => [2] * - * args = [ "-m", "-q", "filename" ] - * args.shift #=> "-m" - * args #=> ["-q", "filename"] + * If +n+ is as large as or larger than <tt>self.length</tt>, + * removes all elements; returns those elements in a new \Array: * - * args = [ "-m", "-q", "filename" ] - * args.shift(2) #=> ["-m", "-q"] - * args #=> ["filename"] + * a = [:foo, 'bar', 2] + * a.shift(3) # => [:foo, 'bar', 2] + * + * If +n+ is zero, returns a new empty \Array; +self+ is unmodified. + * + * Related: #push, #pop, #unshift. */ static VALUE @@ -1348,7 +1654,7 @@ rb_ary_shift_m(int argc, VALUE *argv, VALUE ary) long n; if (argc == 0) { - return rb_ary_shift(ary); + return rb_ary_shift(ary); } rb_ary_modify_check(ary); @@ -1362,109 +1668,133 @@ rb_ary_shift_m(int argc, VALUE *argv, VALUE ary) MJIT_FUNC_EXPORTED VALUE rb_ary_behead(VALUE ary, long n) { - if (n<=0) return ary; + if (n <= 0) { + return ary; + } rb_ary_modify_check(ary); - if (ARY_SHARED_P(ary)) { - if (ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))) { - setup_occupied_shared: - ary_mem_clear(ary, 0, n); - } - ARY_INCREASE_PTR(ary, n); - } - else { - if (RARRAY_LEN(ary) < ARY_DEFAULT_SIZE) { + + if (!ARY_SHARED_P(ary)) { + if (ARY_EMBED_P(ary) || RARRAY_LEN(ary) < ARY_DEFAULT_SIZE) { RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - MEMMOVE(ptr, ptr+n, VALUE, RARRAY_LEN(ary)-n); - }); /* WB: no new reference */ - } - else { - ary_make_shared(ary); - goto setup_occupied_shared; - } + MEMMOVE(ptr, ptr + n, VALUE, RARRAY_LEN(ary) - n); + }); /* WB: no new reference */ + ARY_INCREASE_LEN(ary, -n); + ary_verify(ary); + return ary; + } + + ary_mem_clear(ary, 0, n); + ary_make_shared(ary); + } + else if (ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))) { + ary_mem_clear(ary, 0, n); } - ARY_INCREASE_LEN(ary, -n); + ARY_INCREASE_PTR(ary, n); + ARY_INCREASE_LEN(ary, -n); ary_verify(ary); + return ary; } static VALUE -ary_ensure_room_for_unshift(VALUE ary, int argc) +make_room_for_unshift(VALUE ary, const VALUE *head, VALUE *sharedp, int argc, long capa, long len) +{ + if (head - sharedp < argc) { + long room = capa - len - argc; + + room -= room >> 4; + MEMMOVE((VALUE *)sharedp + argc + room, head, VALUE, len); + head = sharedp + argc + room; + } + ARY_SET_PTR(ary, head - argc); + assert(ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))); + + ary_verify(ary); + return ARY_SHARED_ROOT(ary); +} + +static VALUE +ary_modify_for_unshift(VALUE ary, int argc) { long len = RARRAY_LEN(ary); long new_len = len + argc; long capa; const VALUE *head, *sharedp; - if (len > ARY_MAX_SIZE - argc) { - rb_raise(rb_eIndexError, "index %ld too big", new_len); - } - - if (ARY_SHARED_P(ary)) { - VALUE shared_root = ARY_SHARED_ROOT(ary); - capa = RARRAY_LEN(shared_root); - if (ARY_SHARED_ROOT_OCCUPIED(shared_root) && capa > new_len) { - rb_ary_modify_check(ary); - head = RARRAY_CONST_PTR_TRANSIENT(ary); - sharedp = RARRAY_CONST_PTR_TRANSIENT(shared_root); - goto makeroom_if_need; - } - } - rb_ary_modify(ary); capa = ARY_CAPA(ary); if (capa - (capa >> 6) <= new_len) { - ary_double_capa(ary, new_len); + ary_double_capa(ary, new_len); } /* use shared array for big "queues" */ - if (new_len > ARY_DEFAULT_SIZE * 4) { + if (new_len > ARY_DEFAULT_SIZE * 4 && !ARY_EMBED_P(ary)) { ary_verify(ary); /* make a room for unshifted items */ - capa = ARY_CAPA(ary); - ary_make_shared(ary); + capa = ARY_CAPA(ary); + ary_make_shared(ary); head = sharedp = RARRAY_CONST_PTR_TRANSIENT(ary); - goto makeroom; - makeroom_if_need: - if (head - sharedp < argc) { - long room; - makeroom: - room = capa - new_len; - room -= room >> 4; - MEMMOVE((VALUE *)sharedp + argc + room, head, VALUE, len); - head = sharedp + argc + room; - } - ARY_SET_PTR(ary, head - argc); - assert(ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))); - - ary_verify(ary); - return ARY_SHARED_ROOT(ary); + return make_room_for_unshift(ary, head, (void *)sharedp, argc, capa, len); } else { - /* sliding items */ + /* sliding items */ RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - MEMMOVE(ptr + argc, ptr, VALUE, len); - }); + MEMMOVE(ptr + argc, ptr, VALUE, len); + }); ary_verify(ary); - return ary; + return ary; + } +} + +static VALUE +ary_ensure_room_for_unshift(VALUE ary, int argc) +{ + long len = RARRAY_LEN(ary); + long new_len = len + argc; + + if (len > ARY_MAX_SIZE - argc) { + rb_raise(rb_eIndexError, "index %ld too big", new_len); + } + else if (! ARY_SHARED_P(ary)) { + return ary_modify_for_unshift(ary, argc); + } + else { + VALUE shared_root = ARY_SHARED_ROOT(ary); + long capa = RARRAY_LEN(shared_root); + + if (! ARY_SHARED_ROOT_OCCUPIED(shared_root)) { + return ary_modify_for_unshift(ary, argc); + } + else if (new_len > capa) { + return ary_modify_for_unshift(ary, argc); + } + else { + const VALUE * head = RARRAY_CONST_PTR_TRANSIENT(ary); + void *sharedp = (void *)RARRAY_CONST_PTR_TRANSIENT(shared_root); + + rb_ary_modify_check(ary); + return make_room_for_unshift(ary, head, sharedp, argc, capa, len); + } } } /* * call-seq: - * ary.unshift(obj, ...) -> ary - * ary.prepend(obj, ...) -> ary + * array.unshift(*objects) -> self + * + * Prepends the given +objects+ to +self+: * - * Prepends objects to the front of +self+, moving other elements upwards. - * See also Array#shift for the opposite effect. + * a = [:foo, 'bar', 2] + * a.unshift(:bam, :bat) # => [:bam, :bat, :foo, "bar", 2] * - * a = [ "b", "c", "d" ] - * a.unshift("a") #=> ["a", "b", "c", "d"] - * a.unshift(1, 2) #=> [ 1, 2, "a", "b", "c", "d"] + * Array#prepend is an alias for Array#unshift. + * + * Related: #push, #pop, #shift. */ static VALUE @@ -1474,8 +1804,8 @@ rb_ary_unshift_m(int argc, VALUE *argv, VALUE ary) VALUE target_ary; if (argc == 0) { - rb_ary_modify_check(ary); - return ary; + rb_ary_modify_check(ary); + return ary; } target_ary = ary_ensure_room_for_unshift(ary, argc); @@ -1497,7 +1827,7 @@ rb_ary_elt(VALUE ary, long offset) long len = RARRAY_LEN(ary); if (len == 0) return Qnil; if (offset < 0 || len <= offset) { - return Qnil; + return Qnil; } return RARRAY_AREF(ary, offset); } @@ -1509,7 +1839,7 @@ rb_ary_entry(VALUE ary, long offset) } VALUE -rb_ary_subseq(VALUE ary, long beg, long len) +rb_ary_subseq_step(VALUE ary, long beg, long len, long step) { VALUE klass; long alen = RARRAY_LEN(ary); @@ -1518,48 +1848,129 @@ rb_ary_subseq(VALUE ary, long beg, long len) if (beg < 0 || len < 0) return Qnil; if (alen < len || alen < beg + len) { - len = alen - beg; + len = alen - beg; } - klass = rb_obj_class(ary); + klass = rb_cArray; if (len == 0) return ary_new(klass, 0); + if (step == 0) + rb_raise(rb_eArgError, "slice step cannot be zero"); + if (step == 1) + return ary_make_partial(ary, klass, beg, len); + else + return ary_make_partial_step(ary, klass, beg, len, step); +} - return ary_make_partial(ary, klass, beg, len); +VALUE +rb_ary_subseq(VALUE ary, long beg, long len) +{ + return rb_ary_subseq_step(ary, beg, len, 1); } +static VALUE rb_ary_aref2(VALUE ary, VALUE b, VALUE e); + /* * call-seq: - * ary[index] -> obj or nil - * ary[start, length] -> new_ary or nil - * ary[range] -> new_ary or nil - * ary.slice(index) -> obj or nil - * ary.slice(start, length) -> new_ary or nil - * ary.slice(range) -> new_ary or nil - * - * Element Reference --- Returns the element at +index+, or returns a - * subarray starting at the +start+ index and continuing for +length+ - * elements, or returns a subarray specified by +range+ of indices. - * - * Negative indices count backward from the end of the array (-1 is the last - * element). For +start+ and +range+ cases the starting index is just before - * an element. Additionally, an empty array is returned when the starting - * index for an element range is at the end of the array. - * - * Returns +nil+ if the index (or starting index) are out of range. - * - * a = [ "a", "b", "c", "d", "e" ] - * a[2] + a[0] + a[1] #=> "cab" - * a[6] #=> nil - * a[1, 2] #=> [ "b", "c" ] - * a[1..3] #=> [ "b", "c", "d" ] - * a[4..7] #=> [ "e" ] - * a[6..10] #=> nil - * a[-3, 3] #=> [ "c", "d", "e" ] - * # special cases - * a[5] #=> nil - * a[6, 1] #=> nil - * a[5, 1] #=> [] - * a[5..10] #=> [] + * array[index] -> object or nil + * array[start, length] -> object or nil + * array[range] -> object or nil + * array[aseq] -> object or nil + * array.slice(index) -> object or nil + * array.slice(start, length) -> object or nil + * array.slice(range) -> object or nil + * array.slice(aseq) -> object or nil + * + * Returns elements from +self+; does not modify +self+. + * + * When a single \Integer argument +index+ is given, returns the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a[0] # => :foo + * a[2] # => 2 + * a # => [:foo, "bar", 2] + * + * If +index+ is negative, counts relative to the end of +self+: + * + * a = [:foo, 'bar', 2] + * a[-1] # => 2 + * a[-2] # => "bar" + * + * If +index+ is out of range, returns +nil+. + * + * When two \Integer arguments +start+ and +length+ are given, + * returns a new \Array of size +length+ containing successive elements beginning at offset +start+: + * + * a = [:foo, 'bar', 2] + * a[0, 2] # => [:foo, "bar"] + * a[1, 2] # => ["bar", 2] + * + * If <tt>start + length</tt> is greater than <tt>self.length</tt>, + * returns all elements from offset +start+ to the end: + * + * a = [:foo, 'bar', 2] + * a[0, 4] # => [:foo, "bar", 2] + * a[1, 3] # => ["bar", 2] + * a[2, 2] # => [2] + * + * If <tt>start == self.size</tt> and <tt>length >= 0</tt>, + * returns a new empty \Array. + * + * If +length+ is negative, returns +nil+. * + * When a single \Range argument +range+ is given, + * treats <tt>range.min</tt> as +start+ above + * and <tt>range.size</tt> as +length+ above: + * + * a = [:foo, 'bar', 2] + * a[0..1] # => [:foo, "bar"] + * a[1..2] # => ["bar", 2] + * + * Special case: If <tt>range.start == a.size</tt>, returns a new empty \Array. + * + * If <tt>range.end</tt> is negative, calculates the end index from the end: + * + * a = [:foo, 'bar', 2] + * a[0..-1] # => [:foo, "bar", 2] + * a[0..-2] # => [:foo, "bar"] + * a[0..-3] # => [:foo] + * + * If <tt>range.start</tt> is negative, calculates the start index from the end: + * + * a = [:foo, 'bar', 2] + * a[-1..2] # => [2] + * a[-2..2] # => ["bar", 2] + * a[-3..2] # => [:foo, "bar", 2] + * + * If <tt>range.start</tt> is larger than the array size, returns +nil+. + * + * a = [:foo, 'bar', 2] + * a[4..1] # => nil + * a[4..0] # => nil + * a[4..-1] # => nil + * + * When a single Enumerator::ArithmeticSequence argument +aseq+ is given, + * returns an \Array of elements corresponding to the indexes produced by + * the sequence. + * + * a = ['--', 'data1', '--', 'data2', '--', 'data3'] + * a[(1..).step(2)] # => ["data1", "data2", "data3"] + * + * Unlike slicing with range, if the start or the end of the arithmetic sequence + * is larger than array size, throws RangeError. + * + * a = ['--', 'data1', '--', 'data2', '--', 'data3'] + * a[(1..11).step(2)] + * # RangeError (((1..11).step(2)) out of range) + * a[(7..).step(2)] + * # RangeError (((7..).step(2)) out of range) + * + * If given a single argument, and its type is not one of the listed, tries to + * convert it to Integer, and raises if it is impossible: + * + * a = [:foo, 'bar', 2] + * # Raises TypeError (no implicit conversion of Symbol into Integer): + * a[:foo] + * + * Array#slice is an alias for Array#[]. */ VALUE @@ -1567,18 +1978,18 @@ rb_ary_aref(int argc, const VALUE *argv, VALUE ary) { rb_check_arity(argc, 1, 2); if (argc == 2) { - return rb_ary_aref2(ary, argv[0], argv[1]); + return rb_ary_aref2(ary, argv[0], argv[1]); } return rb_ary_aref1(ary, argv[0]); } -VALUE +static VALUE rb_ary_aref2(VALUE ary, VALUE b, VALUE e) { long beg = NUM2LONG(b); long len = NUM2LONG(e); if (beg < 0) { - beg += RARRAY_LEN(ary); + beg += RARRAY_LEN(ary); } return rb_ary_subseq(ary, beg, len); } @@ -1586,35 +1997,34 @@ rb_ary_aref2(VALUE ary, VALUE b, VALUE e) MJIT_FUNC_EXPORTED VALUE rb_ary_aref1(VALUE ary, VALUE arg) { - long beg, len; + long beg, len, step; /* special case - speeding up */ if (FIXNUM_P(arg)) { - return rb_ary_entry(ary, FIX2LONG(arg)); + return rb_ary_entry(ary, FIX2LONG(arg)); } - /* check if idx is Range */ - switch (rb_range_beg_len(arg, &beg, &len, RARRAY_LEN(ary), 0)) { + /* check if idx is Range or ArithmeticSequence */ + switch (rb_arithmetic_sequence_beg_len_step(arg, &beg, &len, &step, RARRAY_LEN(ary), 0)) { case Qfalse: - break; + break; case Qnil: - return Qnil; + return Qnil; default: - return rb_ary_subseq(ary, beg, len); + return rb_ary_subseq_step(ary, beg, len, step); } + return rb_ary_entry(ary, NUM2LONG(arg)); } /* * call-seq: - * ary.at(index) -> obj or nil + * array.at(index) -> object * - * Returns the element at +index+. A negative index counts from the end of - * +self+. Returns +nil+ if the index is out of range. See also - * Array#[]. + * Returns the element at \Integer offset +index+; does not modify +self+. + * a = [:foo, 'bar', 2] + * a.at(0) # => :foo + * a.at(2) # => 2 * - * a = [ "a", "b", "c", "d", "e" ] - * a.at(0) #=> "a" - * a.at(-1) #=> "e" */ VALUE @@ -1625,81 +2035,131 @@ rb_ary_at(VALUE ary, VALUE pos) /* * call-seq: - * ary.first -> obj or nil - * ary.first(n) -> new_ary + * array.first -> object or nil + * array.first(n) -> new_array * - * Returns the first element, or the first +n+ elements, of the array. - * If the array is empty, the first form returns +nil+, and the - * second form returns an empty array. See also Array#last for - * the opposite effect. + * Returns elements from +self+; does not modify +self+. * - * a = [ "q", "r", "s", "t" ] - * a.first #=> "q" - * a.first(2) #=> ["q", "r"] + * When no argument is given, returns the first element: + * + * a = [:foo, 'bar', 2] + * a.first # => :foo + * a # => [:foo, "bar", 2] + * + * If +self+ is empty, returns +nil+. + * + * When non-negative \Integer argument +n+ is given, + * returns the first +n+ elements in a new \Array: + * + * a = [:foo, 'bar', 2] + * a.first(2) # => [:foo, "bar"] + * + * If <tt>n >= array.size</tt>, returns all elements: + * + * a = [:foo, 'bar', 2] + * a.first(50) # => [:foo, "bar", 2] + * + * If <tt>n == 0</tt> returns an new empty \Array: + * + * a = [:foo, 'bar', 2] + * a.first(0) # [] + * + * Related: #last. */ - static VALUE rb_ary_first(int argc, VALUE *argv, VALUE ary) { if (argc == 0) { - if (RARRAY_LEN(ary) == 0) return Qnil; - return RARRAY_AREF(ary, 0); + if (RARRAY_LEN(ary) == 0) return Qnil; + return RARRAY_AREF(ary, 0); } else { - return ary_take_first_or_last(argc, argv, ary, ARY_TAKE_FIRST); + return ary_take_first_or_last(argc, argv, ary, ARY_TAKE_FIRST); } } /* * call-seq: - * ary.last -> obj or nil - * ary.last(n) -> new_ary + * array.last -> object or nil + * array.last(n) -> new_array + * + * Returns elements from +self+; +self+ is not modified. * - * Returns the last element(s) of +self+. If the array is empty, - * the first form returns +nil+. + * When no argument is given, returns the last element: * - * See also Array#first for the opposite effect. + * a = [:foo, 'bar', 2] + * a.last # => 2 + * a # => [:foo, "bar", 2] * - * a = [ "w", "x", "y", "z" ] - * a.last #=> "z" - * a.last(2) #=> ["y", "z"] + * If +self+ is empty, returns +nil+. + * + * When non-negative \Integer argument +n+ is given, + * returns the last +n+ elements in a new \Array: + * + * a = [:foo, 'bar', 2] + * a.last(2) # => ["bar", 2] + * + * If <tt>n >= array.size</tt>, returns all elements: + * + * a = [:foo, 'bar', 2] + * a.last(50) # => [:foo, "bar", 2] + * + * If <tt>n == 0</tt>, returns an new empty \Array: + * + * a = [:foo, 'bar', 2] + * a.last(0) # [] + * + * Related: #first. */ VALUE rb_ary_last(int argc, const VALUE *argv, VALUE ary) { if (argc == 0) { - long len = RARRAY_LEN(ary); - if (len == 0) return Qnil; - return RARRAY_AREF(ary, len-1); + long len = RARRAY_LEN(ary); + if (len == 0) return Qnil; + return RARRAY_AREF(ary, len-1); } else { - return ary_take_first_or_last(argc, argv, ary, ARY_TAKE_LAST); + return ary_take_first_or_last(argc, argv, ary, ARY_TAKE_LAST); } } /* * call-seq: - * ary.fetch(index) -> obj - * ary.fetch(index, default) -> obj - * ary.fetch(index) {|index| block} -> obj - * - * Tries to return the element at position +index+, but throws an IndexError - * exception if the referenced +index+ lies outside of the array bounds. This - * error can be prevented by supplying a second argument, which will act as a - * +default+ value. - * - * Alternatively, if a block is given it will only be executed when an - * invalid +index+ is referenced. - * - * Negative values of +index+ count from the end of the array. - * - * a = [ 11, 22, 33, 44 ] - * a.fetch(1) #=> 22 - * a.fetch(-1) #=> 44 - * a.fetch(4, 'cat') #=> "cat" - * a.fetch(100) {|i| puts "#{i} is out of bounds"} - * #=> "100 is out of bounds" + * array.fetch(index) -> element + * array.fetch(index, default_value) -> element + * array.fetch(index) {|index| ... } -> element + * + * Returns the element at offset +index+. + * + * With the single \Integer argument +index+, + * returns the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a.fetch(1) # => "bar" + * + * If +index+ is negative, counts from the end of the array: + * + * a = [:foo, 'bar', 2] + * a.fetch(-1) # => 2 + * a.fetch(-2) # => "bar" + * + * With arguments +index+ and +default_value+, + * returns the element at offset +index+ if index is in range, + * otherwise returns +default_value+: + * + * a = [:foo, 'bar', 2] + * a.fetch(1, nil) # => "bar" + * + * With argument +index+ and a block, + * returns the element at offset +index+ if index is in range + * (and the block is not called); otherwise calls the block with index and returns its return value: + * + * a = [:foo, 'bar', 2] + * a.fetch(1) {|index| raise 'Cannot happen' } # => "bar" + * a.fetch(50) {|index| "Value for #{index}" } # => "Value for 50" + * */ static VALUE @@ -1712,48 +2172,60 @@ rb_ary_fetch(int argc, VALUE *argv, VALUE ary) rb_scan_args(argc, argv, "11", &pos, &ifnone); block_given = rb_block_given_p(); if (block_given && argc == 2) { - rb_warn("block supersedes default value argument"); + rb_warn("block supersedes default value argument"); } idx = NUM2LONG(pos); if (idx < 0) { - idx += RARRAY_LEN(ary); + idx += RARRAY_LEN(ary); } if (idx < 0 || RARRAY_LEN(ary) <= idx) { - if (block_given) return rb_yield(pos); - if (argc == 1) { - rb_raise(rb_eIndexError, "index %ld outside of array bounds: %ld...%ld", - idx - (idx < 0 ? RARRAY_LEN(ary) : 0), -RARRAY_LEN(ary), RARRAY_LEN(ary)); - } - return ifnone; + if (block_given) return rb_yield(pos); + if (argc == 1) { + rb_raise(rb_eIndexError, "index %ld outside of array bounds: %ld...%ld", + idx - (idx < 0 ? RARRAY_LEN(ary) : 0), -RARRAY_LEN(ary), RARRAY_LEN(ary)); + } + return ifnone; } return RARRAY_AREF(ary, idx); } /* * call-seq: - * ary.find_index(obj) -> int or nil - * ary.find_index {|item| block} -> int or nil - * ary.find_index -> Enumerator - * ary.index(obj) -> int or nil - * ary.index {|item| block} -> int or nil - * ary.index -> Enumerator + * array.index(object) -> integer or nil + * array.index {|element| ... } -> integer or nil + * array.index -> new_enumerator + * + * Returns the index of a specified element. + * + * When argument +object+ is given but no block, + * returns the index of the first element +element+ + * for which <tt>object == element</tt>: * - * Returns the _index_ of the first object in +ary+ such that the object is - * <code>==</code> to +obj+. + * a = [:foo, 'bar', 2, 'bar'] + * a.index('bar') # => 1 * - * If a block is given instead of an argument, returns the _index_ of the - * first object for which the block returns +true+. Returns +nil+ if no - * match is found. + * Returns +nil+ if no such element found. * - * See also Array#rindex. + * When both argument +object+ and a block are given, + * calls the block with each successive element; + * returns the index of the first element for which the block returns a truthy value: * - * An Enumerator is returned if neither a block nor argument is given. + * a = [:foo, 'bar', 2, 'bar'] + * a.index {|element| element == 'bar' } # => 1 * - * a = [ "a", "b", "c" ] - * a.index("b") #=> 1 - * a.index("z") #=> nil - * a.index {|x| x == "b"} #=> 1 + * Returns +nil+ if the block never returns a truthy value. + * + * When neither an argument nor a block is given, returns a new Enumerator: + * + * a = [:foo, 'bar', 2] + * e = a.index + * e # => #<Enumerator: [:foo, "bar", 2]:index> + * e.each {|element| element == 'bar' } # => 1 + * + * Array#find_index is an alias for Array#index. + * + * Related: #rindex. */ static VALUE @@ -1763,49 +2235,58 @@ rb_ary_index(int argc, VALUE *argv, VALUE ary) long i; if (argc == 0) { - RETURN_ENUMERATOR(ary, 0, 0); - for (i=0; i<RARRAY_LEN(ary); i++) { - if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { - return LONG2NUM(i); - } - } - return Qnil; + RETURN_ENUMERATOR(ary, 0, 0); + for (i=0; i<RARRAY_LEN(ary); i++) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { + return LONG2NUM(i); + } + } + return Qnil; } rb_check_arity(argc, 0, 1); val = argv[0]; if (rb_block_given_p()) - rb_warn("given block not used"); + rb_warn("given block not used"); for (i=0; i<RARRAY_LEN(ary); i++) { - VALUE e = RARRAY_AREF(ary, i); - if (rb_equal(e, val)) { - return LONG2NUM(i); - } + VALUE e = RARRAY_AREF(ary, i); + if (rb_equal(e, val)) { + return LONG2NUM(i); + } } return Qnil; } /* * call-seq: - * ary.rindex(obj) -> int or nil - * ary.rindex {|item| block} -> int or nil - * ary.rindex -> Enumerator + * array.rindex(object) -> integer or nil + * array.rindex {|element| ... } -> integer or nil + * array.rindex -> new_enumerator * - * Returns the _index_ of the last object in +self+ <code>==</code> to +obj+. + * Returns the index of the last element for which <tt>object == element</tt>. * - * If a block is given instead of an argument, returns the _index_ of the - * first object for which the block returns +true+, starting from the last - * object. + * When argument +object+ is given but no block, returns the index of the last such element found: * - * Returns +nil+ if no match is found. + * a = [:foo, 'bar', 2, 'bar'] + * a.rindex('bar') # => 3 * - * See also Array#index. + * Returns +nil+ if no such object found. * - * If neither block nor argument is given, an Enumerator is returned instead. + * When a block is given but no argument, calls the block with each successive element; + * returns the index of the last element for which the block returns a truthy value: * - * a = [ "a", "b", "b", "b", "c" ] - * a.rindex("b") #=> 3 - * a.rindex("z") #=> nil - * a.rindex {|x| x == "b"} #=> 3 + * a = [:foo, 'bar', 2, 'bar'] + * a.rindex {|element| element == 'bar' } # => 3 + * + * Returns +nil+ if the block never returns a truthy value. + * + * When neither an argument nor a block is given, returns a new \Enumerator: + * + * a = [:foo, 'bar', 2, 'bar'] + * e = a.rindex + * e # => #<Enumerator: [:foo, "bar", 2, "bar"]:rindex> + * e.each {|element| element == 'bar' } # => 3 + * + * Related: #index. */ static VALUE @@ -1815,25 +2296,25 @@ rb_ary_rindex(int argc, VALUE *argv, VALUE ary) long i = RARRAY_LEN(ary), len; if (argc == 0) { - RETURN_ENUMERATOR(ary, 0, 0); - while (i--) { - if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) - return LONG2NUM(i); - if (i > (len = RARRAY_LEN(ary))) { - i = len; - } - } - return Qnil; + RETURN_ENUMERATOR(ary, 0, 0); + while (i--) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) + return LONG2NUM(i); + if (i > (len = RARRAY_LEN(ary))) { + i = len; + } + } + return Qnil; } rb_check_arity(argc, 0, 1); val = argv[0]; if (rb_block_given_p()) - rb_warn("given block not used"); + rb_warn("given block not used"); while (i--) { - VALUE e = RARRAY_AREF(ary, i); - if (rb_equal(e, val)) { - return LONG2NUM(i); - } + VALUE e = RARRAY_AREF(ary, i); + if (rb_equal(e, val)) { + return LONG2NUM(i); + } if (i > RARRAY_LEN(ary)) { break; } @@ -1859,54 +2340,54 @@ rb_ary_splice(VALUE ary, long beg, long len, const VALUE *rptr, long rlen) if (len < 0) rb_raise(rb_eIndexError, "negative length (%ld)", len); olen = RARRAY_LEN(ary); if (beg < 0) { - beg += olen; - if (beg < 0) { - rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", - beg - olen, -olen); - } + beg += olen; + if (beg < 0) { + rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", + beg - olen, -olen); + } } if (olen < len || olen < beg + len) { - len = olen - beg; + len = olen - beg; } { const VALUE *optr = RARRAY_CONST_PTR_TRANSIENT(ary); - rofs = (rptr >= optr && rptr < optr + olen) ? rptr - optr : -1; + rofs = (rptr >= optr && rptr < optr + olen) ? rptr - optr : -1; } if (beg >= olen) { - VALUE target_ary; - if (beg > ARY_MAX_SIZE - rlen) { - rb_raise(rb_eIndexError, "index %ld too big", beg); - } - target_ary = ary_ensure_room_for_push(ary, rlen-len); /* len is 0 or negative */ - len = beg + rlen; - ary_mem_clear(ary, olen, beg - olen); - if (rlen > 0) { + VALUE target_ary; + if (beg > ARY_MAX_SIZE - rlen) { + rb_raise(rb_eIndexError, "index %ld too big", beg); + } + target_ary = ary_ensure_room_for_push(ary, rlen-len); /* len is 0 or negative */ + len = beg + rlen; + ary_mem_clear(ary, olen, beg - olen); + if (rlen > 0) { if (rofs != -1) rptr = RARRAY_CONST_PTR_TRANSIENT(ary) + rofs; - ary_memcpy0(ary, beg, rlen, rptr, target_ary); - } - ARY_SET_LEN(ary, len); + ary_memcpy0(ary, beg, rlen, rptr, target_ary); + } + ARY_SET_LEN(ary, len); } else { - long alen; - - if (olen - len > ARY_MAX_SIZE - rlen) { - rb_raise(rb_eIndexError, "index %ld too big", olen + rlen - len); - } - rb_ary_modify(ary); - alen = olen + rlen - len; - if (alen >= ARY_CAPA(ary)) { - ary_double_capa(ary, alen); - } - - if (len != rlen) { + long alen; + + if (olen - len > ARY_MAX_SIZE - rlen) { + rb_raise(rb_eIndexError, "index %ld too big", olen + rlen - len); + } + rb_ary_modify(ary); + alen = olen + rlen - len; + if (alen >= ARY_CAPA(ary)) { + ary_double_capa(ary, alen); + } + + if (len != rlen) { RARRAY_PTR_USE_TRANSIENT(ary, ptr, MEMMOVE(ptr + beg + rlen, ptr + beg + len, VALUE, olen - (beg + len))); - ARY_SET_LEN(ary, alen); - } - if (rlen > 0) { + ARY_SET_LEN(ary, alen); + } + if (rlen > 0) { if (rofs != -1) rptr = RARRAY_CONST_PTR_TRANSIENT(ary) + rofs; /* give up wb-protected ary */ RB_OBJ_WB_UNPROTECT_FOR(ARRAY, ary); @@ -1916,7 +2397,7 @@ rb_ary_splice(VALUE ary, long beg, long len, const VALUE *rptr, long rlen) */ RARRAY_PTR_USE_TRANSIENT(ary, ptr, MEMMOVE(ptr + beg, rptr, VALUE, rlen)); - } + } } } @@ -1927,22 +2408,14 @@ rb_ary_set_len(VALUE ary, long len) rb_ary_modify_check(ary); if (ARY_SHARED_P(ary)) { - rb_raise(rb_eRuntimeError, "can't set length of shared "); + rb_raise(rb_eRuntimeError, "can't set length of shared "); } if (len > (capa = (long)ARY_CAPA(ary))) { - rb_bug("probable buffer overflow: %ld for %ld", len, capa); + rb_bug("probable buffer overflow: %ld for %ld", len, capa); } ARY_SET_LEN(ary, len); } -/*! - * expands or shrinks \a ary to \a len elements. - * expanded region will be filled with Qnil. - * \param ary an array - * \param len new size - * \return \a ary - * \post the size of \a ary is \a len. - */ VALUE rb_ary_resize(VALUE ary, long len) { @@ -1952,118 +2425,231 @@ rb_ary_resize(VALUE ary, long len) olen = RARRAY_LEN(ary); if (len == olen) return ary; if (len > ARY_MAX_SIZE) { - rb_raise(rb_eIndexError, "index %ld too big", len); + rb_raise(rb_eIndexError, "index %ld too big", len); } if (len > olen) { - if (len >= ARY_CAPA(ary)) { - ary_double_capa(ary, len); - } - ary_mem_clear(ary, olen, len - olen); - ARY_SET_LEN(ary, len); + if (len >= ARY_CAPA(ary)) { + ary_double_capa(ary, len); + } + ary_mem_clear(ary, olen, len - olen); + ARY_SET_LEN(ary, len); } else if (ARY_EMBED_P(ary)) { ARY_SET_EMBED_LEN(ary, len); } - else if (len <= RARRAY_EMBED_LEN_MAX) { - VALUE tmp[RARRAY_EMBED_LEN_MAX]; - MEMCPY(tmp, ARY_HEAP_PTR(ary), VALUE, len); - ary_discard(ary); - MEMCPY((VALUE *)ARY_EMBED_PTR(ary), tmp, VALUE, len); /* WB: no new reference */ + else if (len <= ary_embed_capa(ary)) { + const VALUE *ptr = ARY_HEAP_PTR(ary); + long ptr_capa = ARY_HEAP_SIZE(ary); + bool is_malloc_ptr = !ARY_SHARED_P(ary) && !RARRAY_TRANSIENT_P(ary); + + FL_UNSET(ary, RARRAY_TRANSIENT_FLAG); + FL_SET_EMBED(ary); + + MEMCPY((VALUE *)ARY_EMBED_PTR(ary), ptr, VALUE, len); /* WB: no new reference */ ARY_SET_EMBED_LEN(ary, len); + + if (is_malloc_ptr) ruby_sized_xfree((void *)ptr, ptr_capa); } else { - if (olen > len + ARY_DEFAULT_SIZE) { - ary_heap_realloc(ary, len); - ARY_SET_CAPA(ary, len); - } - ARY_SET_HEAP_LEN(ary, len); + if (olen > len + ARY_DEFAULT_SIZE) { + size_t new_capa = ary_heap_realloc(ary, len); + ARY_SET_CAPA(ary, new_capa); + } + ARY_SET_HEAP_LEN(ary, len); } ary_verify(ary); return ary; } +static VALUE +ary_aset_by_rb_ary_store(VALUE ary, long key, VALUE val) +{ + rb_ary_store(ary, key, val); + return val; +} + +static VALUE +ary_aset_by_rb_ary_splice(VALUE ary, long beg, long len, VALUE val) +{ + VALUE rpl = rb_ary_to_ary(val); + rb_ary_splice(ary, beg, len, RARRAY_CONST_PTR_TRANSIENT(rpl), RARRAY_LEN(rpl)); + RB_GC_GUARD(rpl); + return val; +} + /* * call-seq: - * ary[index] = obj -> obj - * ary[start, length] = obj or other_ary or nil -> obj or other_ary or nil - * ary[range] = obj or other_ary or nil -> obj or other_ary or nil - * - * Element Assignment --- Sets the element at +index+, or replaces a subarray - * from the +start+ index for +length+ elements, or replaces a subarray - * specified by the +range+ of indices. - * - * If indices are greater than the current capacity of the array, the array - * grows automatically. Elements are inserted into the array at +start+ if - * +length+ is zero. - * - * Negative indices will count backward from the end of the array. For - * +start+ and +range+ cases the starting index is just before an element. - * - * An IndexError is raised if a negative index points past the beginning of - * the array. - * - * See also Array#push, and Array#unshift. - * - * a = Array.new - * a[4] = "4"; #=> [nil, nil, nil, nil, "4"] - * a[0, 3] = [ 'a', 'b', 'c' ] #=> ["a", "b", "c", nil, "4"] - * a[1..2] = [ 1, 2 ] #=> ["a", 1, 2, nil, "4"] - * a[0, 2] = "?" #=> ["?", 2, nil, "4"] - * a[0..2] = "A" #=> ["A", "4"] - * a[-1] = "Z" #=> ["A", "Z"] - * a[1..-1] = nil #=> ["A", nil] - * a[1..-1] = [] #=> ["A"] - * a[0, 0] = [ 1, 2 ] #=> [1, 2, "A"] - * a[3, 0] = "B" #=> [1, 2, "A", "B"] + * array[index] = object -> object + * array[start, length] = object -> object + * array[range] = object -> object + * + * Assigns elements in +self+; returns the given +object+. + * + * When \Integer argument +index+ is given, assigns +object+ to an element in +self+. + * + * If +index+ is non-negative, assigns +object+ the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a[0] = 'foo' # => "foo" + * a # => ["foo", "bar", 2] + * + * If +index+ is greater than <tt>self.length</tt>, extends the array: + * + * a = [:foo, 'bar', 2] + * a[7] = 'foo' # => "foo" + * a # => [:foo, "bar", 2, nil, nil, nil, nil, "foo"] + * + * If +index+ is negative, counts backwards from the end of the array: + * + * a = [:foo, 'bar', 2] + * a[-1] = 'two' # => "two" + * a # => [:foo, "bar", "two"] + * + * When \Integer arguments +start+ and +length+ are given and +object+ is not an \Array, + * removes <tt>length - 1</tt> elements beginning at offset +start+, + * and assigns +object+ at offset +start+: + * + * a = [:foo, 'bar', 2] + * a[0, 2] = 'foo' # => "foo" + * a # => ["foo", 2] + * + * If +start+ is negative, counts backwards from the end of the array: + * + * a = [:foo, 'bar', 2] + * a[-2, 2] = 'foo' # => "foo" + * a # => [:foo, "foo"] + * + * If +start+ is non-negative and outside the array (<tt> >= self.size</tt>), + * extends the array with +nil+, assigns +object+ at offset +start+, + * and ignores +length+: + * + * a = [:foo, 'bar', 2] + * a[6, 50] = 'foo' # => "foo" + * a # => [:foo, "bar", 2, nil, nil, nil, "foo"] + * + * If +length+ is zero, shifts elements at and following offset +start+ + * and assigns +object+ at offset +start+: + * + * a = [:foo, 'bar', 2] + * a[1, 0] = 'foo' # => "foo" + * a # => [:foo, "foo", "bar", 2] + * + * If +length+ is too large for the existing array, does not extend the array: + * + * a = [:foo, 'bar', 2] + * a[1, 5] = 'foo' # => "foo" + * a # => [:foo, "foo"] + * + * When \Range argument +range+ is given and +object+ is an \Array, + * removes <tt>length - 1</tt> elements beginning at offset +start+, + * and assigns +object+ at offset +start+: + * + * a = [:foo, 'bar', 2] + * a[0..1] = 'foo' # => "foo" + * a # => ["foo", 2] + * + * if <tt>range.begin</tt> is negative, counts backwards from the end of the array: + * + * a = [:foo, 'bar', 2] + * a[-2..2] = 'foo' # => "foo" + * a # => [:foo, "foo"] + * + * If the array length is less than <tt>range.begin</tt>, + * assigns +object+ at offset <tt>range.begin</tt>, and ignores +length+: + * + * a = [:foo, 'bar', 2] + * a[6..50] = 'foo' # => "foo" + * a # => [:foo, "bar", 2, nil, nil, nil, "foo"] + * + * If <tt>range.end</tt> is zero, shifts elements at and following offset +start+ + * and assigns +object+ at offset +start+: + * + * a = [:foo, 'bar', 2] + * a[1..0] = 'foo' # => "foo" + * a # => [:foo, "foo", "bar", 2] + * + * If <tt>range.end</tt> is negative, assigns +object+ at offset +start+, + * retains <tt>range.end.abs -1</tt> elements past that, and removes those beyond: + * + * a = [:foo, 'bar', 2] + * a[1..-1] = 'foo' # => "foo" + * a # => [:foo, "foo"] + * a = [:foo, 'bar', 2] + * a[1..-2] = 'foo' # => "foo" + * a # => [:foo, "foo", 2] + * a = [:foo, 'bar', 2] + * a[1..-3] = 'foo' # => "foo" + * a # => [:foo, "foo", "bar", 2] + * a = [:foo, 'bar', 2] + * + * If <tt>range.end</tt> is too large for the existing array, + * replaces array elements, but does not extend the array with +nil+ values: + * + * a = [:foo, 'bar', 2] + * a[1..5] = 'foo' # => "foo" + * a # => [:foo, "foo"] + * */ static VALUE rb_ary_aset(int argc, VALUE *argv, VALUE ary) { long offset, beg, len; - VALUE rpl; + rb_check_arity(argc, 2, 3); + rb_ary_modify_check(ary); if (argc == 3) { - rb_ary_modify_check(ary); - beg = NUM2LONG(argv[0]); - len = NUM2LONG(argv[1]); - goto range; + beg = NUM2LONG(argv[0]); + len = NUM2LONG(argv[1]); + return ary_aset_by_rb_ary_splice(ary, beg, len, argv[2]); } - rb_check_arity(argc, 2, 2); - rb_ary_modify_check(ary); if (FIXNUM_P(argv[0])) { - offset = FIX2LONG(argv[0]); - goto fixnum; + offset = FIX2LONG(argv[0]); + return ary_aset_by_rb_ary_store(ary, offset, argv[1]); } if (rb_range_beg_len(argv[0], &beg, &len, RARRAY_LEN(ary), 1)) { - /* check if idx is Range */ - range: - rpl = rb_ary_to_ary(argv[argc-1]); - rb_ary_splice(ary, beg, len, RARRAY_CONST_PTR_TRANSIENT(rpl), RARRAY_LEN(rpl)); - RB_GC_GUARD(rpl); - return argv[argc-1]; + /* check if idx is Range */ + return ary_aset_by_rb_ary_splice(ary, beg, len, argv[1]); } offset = NUM2LONG(argv[0]); -fixnum: - rb_ary_store(ary, offset, argv[1]); - return argv[1]; + return ary_aset_by_rb_ary_store(ary, offset, argv[1]); } /* * call-seq: - * ary.insert(index, obj...) -> ary + * array.insert(index, *objects) -> self + * + * Inserts given +objects+ before or after the element at \Integer index +offset+; + * returns +self+. + * + * When +index+ is non-negative, inserts all given +objects+ + * before the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a.insert(1, :bat, :bam) # => [:foo, :bat, :bam, "bar", 2] + * + * Extends the array if +index+ is beyond the array (<tt>index >= self.size</tt>): + * + * a = [:foo, 'bar', 2] + * a.insert(5, :bat, :bam) + * a # => [:foo, "bar", 2, nil, nil, :bat, :bam] + * + * Does nothing if no objects given: * - * Inserts the given values before the element with the given +index+. + * a = [:foo, 'bar', 2] + * a.insert(1) + * a.insert(50) + * a.insert(-50) + * a # => [:foo, "bar", 2] * - * Negative indices count backwards from the end of the array, where +-1+ is - * the last element. If a negative index is used, the given values will be - * inserted after that element, so using an index of +-1+ will insert the - * values at the end of the array. + * When +index+ is negative, inserts all given +objects+ + * _after_ the element at offset <tt>index+self.size</tt>: + * + * a = [:foo, 'bar', 2] + * a.insert(-2, :bat, :bam) + * a # => [:foo, "bar", :bat, :bam, 2] * - * a = %w{ a b c d } - * a.insert(2, 99) #=> ["a", "b", 99, "c", "d"] - * a.insert(-2, 1, 2, 3) #=> ["a", "b", 99, "c", 1, 2, 3, "d"] */ static VALUE @@ -2076,15 +2662,15 @@ rb_ary_insert(int argc, VALUE *argv, VALUE ary) pos = NUM2LONG(argv[0]); if (argc == 1) return ary; if (pos == -1) { - pos = RARRAY_LEN(ary); + pos = RARRAY_LEN(ary); } else if (pos < 0) { - long minpos = -RARRAY_LEN(ary) - 1; - if (pos < minpos) { - rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", - pos, minpos); - } - pos++; + long minpos = -RARRAY_LEN(ary) - 1; + if (pos < minpos) { + rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", + pos, minpos); + } + pos++; } rb_ary_splice(ary, pos, 0, argv + 1, argc - 1); return ary; @@ -2101,20 +2687,47 @@ ary_enum_length(VALUE ary, VALUE args, VALUE eobj) /* * call-seq: - * ary.each {|item| block} -> ary - * ary.each -> Enumerator + * array.each {|element| ... } -> self + * array.each -> Enumerator + * + * Iterates over array elements. + * + * When a block given, passes each successive array element to the block; + * returns +self+: + * + * a = [:foo, 'bar', 2] + * a.each {|element| puts "#{element.class} #{element}" } + * + * Output: + * + * Symbol foo + * String bar + * Integer 2 + * + * Allows the array to be modified during iteration: + * + * a = [:foo, 'bar', 2] + * a.each {|element| puts element; a.clear if element.to_s.start_with?('b') } + * + * Output: * - * Calls the given block once for each element in +self+, passing that element - * as a parameter. Returns the array itself. + * foo + * bar * - * If no block is given, an Enumerator is returned. + * When no block given, returns a new \Enumerator: + * a = [:foo, 'bar', 2] * - * a = [ "a", "b", "c" ] - * a.each {|x| print x, " -- " } + * e = a.each + * e # => #<Enumerator: [:foo, "bar", 2]:each> + * a1 = e.each {|element| puts "#{element.class} #{element}" } * - * produces: + * Output: * - * a -- b -- c -- + * Symbol foo + * String bar + * Integer 2 + * + * Related: #each_index, #reverse_each. */ VALUE @@ -2124,27 +2737,54 @@ rb_ary_each(VALUE ary) ary_verify(ary); RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); for (i=0; i<RARRAY_LEN(ary); i++) { - rb_yield(RARRAY_AREF(ary, i)); + rb_yield(RARRAY_AREF(ary, i)); } return ary; } /* * call-seq: - * ary.each_index {|index| block} -> ary - * ary.each_index -> Enumerator + * array.each_index {|index| ... } -> self + * array.each_index -> Enumerator + * + * Iterates over array indexes. + * + * When a block given, passes each successive array index to the block; + * returns +self+: + * + * a = [:foo, 'bar', 2] + * a.each_index {|index| puts "#{index} #{a[index]}" } + * + * Output: + * + * 0 foo + * 1 bar + * 2 2 * - * Same as Array#each, but passes the +index+ of the element instead of the - * element itself. + * Allows the array to be modified during iteration: * - * An Enumerator is returned if no block is given. + * a = [:foo, 'bar', 2] + * a.each_index {|index| puts index; a.clear if index > 0 } * - * a = [ "a", "b", "c" ] - * a.each_index {|x| print x, " -- " } + * Output: * - * produces: + * 0 + * 1 * - * 0 -- 1 -- 2 -- + * When no block given, returns a new \Enumerator: + * + * a = [:foo, 'bar', 2] + * e = a.each_index + * e # => #<Enumerator: [:foo, "bar", 2]:each_index> + * a1 = e.each {|index| puts "#{index} #{a[index]}"} + * + * Output: + * + * 0 foo + * 1 bar + * 2 2 + * + * Related: #each, #reverse_each. */ static VALUE @@ -2154,24 +2794,54 @@ rb_ary_each_index(VALUE ary) RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); for (i=0; i<RARRAY_LEN(ary); i++) { - rb_yield(LONG2NUM(i)); + rb_yield(LONG2NUM(i)); } return ary; } /* * call-seq: - * ary.reverse_each {|item| block} -> ary - * ary.reverse_each -> Enumerator + * array.reverse_each {|element| ... } -> self + * array.reverse_each -> Enumerator + * + * Iterates backwards over array elements. * - * Same as Array#each, but traverses +self+ in reverse order. + * When a block given, passes, in reverse order, each element to the block; + * returns +self+: * - * a = [ "a", "b", "c" ] - * a.reverse_each {|x| print x, " " } + * a = [:foo, 'bar', 2] + * a.reverse_each {|element| puts "#{element.class} #{element}" } * - * produces: + * Output: * - * c b a + * Integer 2 + * String bar + * Symbol foo + * + * Allows the array to be modified during iteration: + * + * a = [:foo, 'bar', 2] + * a.reverse_each {|element| puts element; a.clear if element.to_s.start_with?('b') } + * + * Output: + * + * 2 + * bar + * + * When no block given, returns a new \Enumerator: + * + * a = [:foo, 'bar', 2] + * e = a.reverse_each + * e # => #<Enumerator: [:foo, "bar", 2]:reverse_each> + * a1 = e.each {|element| puts "#{element.class} #{element}" } + * + * Output: + * + * Integer 2 + * String bar + * Symbol foo + * + * Related: #each, #each_index. */ static VALUE @@ -2182,24 +2852,21 @@ rb_ary_reverse_each(VALUE ary) RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); len = RARRAY_LEN(ary); while (len--) { - long nlen; - rb_yield(RARRAY_AREF(ary, len)); - nlen = RARRAY_LEN(ary); - if (nlen < len) { - len = nlen; - } + long nlen; + rb_yield(RARRAY_AREF(ary, len)); + nlen = RARRAY_LEN(ary); + if (nlen < len) { + len = nlen; + } } return ary; } /* * call-seq: - * ary.length -> int - * - * Returns the number of elements in +self+. May be zero. + * array.length -> an_integer * - * [ 1, 2, 3, 4, 5 ].length #=> 5 - * [].length #=> 0 + * Returns the count of elements in +self+. */ static VALUE @@ -2211,19 +2878,16 @@ rb_ary_length(VALUE ary) /* * call-seq: - * ary.empty? -> true or false + * array.empty? -> true or false * - * Returns +true+ if +self+ contains no elements. - * - * [].empty? #=> true + * Returns +true+ if the count of elements in +self+ is zero, + * +false+ otherwise. */ static VALUE rb_ary_empty_p(VALUE ary) { - if (RARRAY_LEN(ary) == 0) - return Qtrue; - return Qfalse; + return RBOOL(RARRAY_LEN(ary) == 0); } VALUE @@ -2259,15 +2923,15 @@ recursive_join(VALUE obj, VALUE argp, int recur) int *first = (int *)arg[3]; if (recur) { - rb_raise(rb_eArgError, "recursive array join"); + rb_raise(rb_eArgError, "recursive array join"); } else { - ary_join_1(obj, ary, sep, 0, result, first); + ary_join_1(obj, ary, sep, 0, result, first); } return Qnil; } -static void +static long ary_join_0(VALUE ary, VALUE sep, long max, VALUE result) { long i; @@ -2275,11 +2939,40 @@ ary_join_0(VALUE ary, VALUE sep, long max, VALUE result) if (max > 0) rb_enc_copy(result, RARRAY_AREF(ary, 0)); for (i=0; i<max; i++) { - val = RARRAY_AREF(ary, i); - if (i > 0 && !NIL_P(sep)) - rb_str_buf_append(result, sep); - rb_str_buf_append(result, val); - if (OBJ_TAINTED(val)) OBJ_TAINT(result); + val = RARRAY_AREF(ary, i); + if (!RB_TYPE_P(val, T_STRING)) break; + if (i > 0 && !NIL_P(sep)) + rb_str_buf_append(result, sep); + rb_str_buf_append(result, val); + } + return i; +} + +static void +ary_join_1_str(VALUE dst, VALUE src, int *first) +{ + rb_str_buf_append(dst, src); + if (*first) { + rb_enc_copy(dst, src); + *first = FALSE; + } +} + +static void +ary_join_1_ary(VALUE obj, VALUE ary, VALUE sep, VALUE result, VALUE val, int *first) +{ + if (val == ary) { + rb_raise(rb_eArgError, "recursive array join"); + } + else { + VALUE args[4]; + + *first = FALSE; + args[0] = val; + args[1] = sep; + args[2] = result; + args[3] = (VALUE)first; + rb_exec_recursive(recursive_join, obj, (VALUE)args); } } @@ -2289,50 +2982,25 @@ ary_join_1(VALUE obj, VALUE ary, VALUE sep, long i, VALUE result, int *first) VALUE val, tmp; for (; i<RARRAY_LEN(ary); i++) { - if (i > 0 && !NIL_P(sep)) - rb_str_buf_append(result, sep); - - val = RARRAY_AREF(ary, i); - if (RB_TYPE_P(val, T_STRING)) { - str_join: - rb_str_buf_append(result, val); - if (*first) { - rb_enc_copy(result, val); - *first = FALSE; - } - } - else if (RB_TYPE_P(val, T_ARRAY)) { - obj = val; - ary_join: - if (val == ary) { - rb_raise(rb_eArgError, "recursive array join"); - } - else { - VALUE args[4]; - - *first = FALSE; - args[0] = val; - args[1] = sep; - args[2] = result; - args[3] = (VALUE)first; - rb_exec_recursive(recursive_join, obj, (VALUE)args); - } - } - else { - tmp = rb_check_string_type(val); - if (!NIL_P(tmp)) { - val = tmp; - goto str_join; - } - tmp = rb_check_array_type(val); - if (!NIL_P(tmp)) { - obj = val; - val = tmp; - goto ary_join; - } - val = rb_obj_as_string(val); - goto str_join; - } + if (i > 0 && !NIL_P(sep)) + rb_str_buf_append(result, sep); + + val = RARRAY_AREF(ary, i); + if (RB_TYPE_P(val, T_STRING)) { + ary_join_1_str(result, val, first); + } + else if (RB_TYPE_P(val, T_ARRAY)) { + ary_join_1_ary(val, ary, sep, result, val, first); + } + else if (!NIL_P(tmp = rb_check_string_type(val))) { + ary_join_1_str(result, tmp, first); + } + else if (!NIL_P(tmp = rb_check_array_type(val))) { + ary_join_1_ary(val, ary, sep, result, tmp, first); + } + else { + ary_join_1_str(result, rb_obj_as_string(val), first); + } } } @@ -2340,38 +3008,36 @@ VALUE rb_ary_join(VALUE ary, VALUE sep) { long len = 1, i; - int taint = FALSE; VALUE val, tmp, result; if (RARRAY_LEN(ary) == 0) return rb_usascii_str_new(0, 0); - if (OBJ_TAINTED(ary)) taint = TRUE; if (!NIL_P(sep)) { - StringValue(sep); - len += RSTRING_LEN(sep) * (RARRAY_LEN(ary) - 1); + StringValue(sep); + len += RSTRING_LEN(sep) * (RARRAY_LEN(ary) - 1); } for (i=0; i<RARRAY_LEN(ary); i++) { - val = RARRAY_AREF(ary, i); - tmp = rb_check_string_type(val); - - if (NIL_P(tmp) || tmp != val) { - int first; - result = rb_str_buf_new(len + (RARRAY_LEN(ary)-i)*10); - rb_enc_associate(result, rb_usascii_encoding()); - if (taint) OBJ_TAINT(result); - ary_join_0(ary, sep, i, result); - first = i == 0; - ary_join_1(ary, ary, sep, i, result, &first); - return result; - } + val = RARRAY_AREF(ary, i); + tmp = rb_check_string_type(val); + + if (NIL_P(tmp) || tmp != val) { + int first; + long n = RARRAY_LEN(ary); + if (i > n) i = n; + result = rb_str_buf_new(len + (n-i)*10); + rb_enc_associate(result, rb_usascii_encoding()); + i = ary_join_0(ary, sep, i, result); + first = i == 0; + ary_join_1(ary, ary, sep, i, result, &first); + return result; + } - len += RSTRING_LEN(tmp); + len += RSTRING_LEN(tmp); } result = rb_str_new(0, len); rb_str_set_len(result, 0); - if (taint) OBJ_TAINT(result); ary_join_0(ary, sep, RARRAY_LEN(ary), result); return result; @@ -2379,22 +3045,32 @@ rb_ary_join(VALUE ary, VALUE sep) /* * call-seq: - * ary.join(separator=$,) -> str + * array.join ->new_string + * array.join(separator = $,) -> new_string + * + * Returns the new \String formed by joining the array elements after conversion. + * For each element +element+: * - * Returns a string created by converting each element of the array to - * a string, separated by the given +separator+. - * If the +separator+ is +nil+, it uses current <code>$,</code>. - * If both the +separator+ and <code>$,</code> are +nil+, - * it uses an empty string. + * - Uses <tt>element.to_s</tt> if +element+ is not a <tt>kind_of?(Array)</tt>. + * - Uses recursive <tt>element.join(separator)</tt> if +element+ is a <tt>kind_of?(Array)</tt>. * - * [ "a", "b", "c" ].join #=> "abc" - * [ "a", "b", "c" ].join("-") #=> "a-b-c" + * With no argument, joins using the output field separator, <tt>$,</tt>: * - * For nested arrays, join is applied recursively: + * a = [:foo, 'bar', 2] + * $, # => nil + * a.join # => "foobar2" + * + * With \string argument +separator+, joins using that separator: + * + * a = [:foo, 'bar', 2] + * a.join("\n") # => "foo\nbar\n2" + * + * Joins recursively for nested Arrays: + * + * a = [:foo, [:bar, [:baz, :bat]]] + * a.join # => "foobarbazbat" * - * [ "a", [1, 2, [:x, :y]], "b" ].join("-") #=> "a-1-2-x-y-b" */ - static VALUE rb_ary_join_m(int argc, VALUE *argv, VALUE ary) { @@ -2403,7 +3079,7 @@ rb_ary_join_m(int argc, VALUE *argv, VALUE ary) if (rb_check_arity(argc, 0, 1) == 0 || NIL_P(sep = argv[0])) { sep = rb_output_fs; if (!NIL_P(sep)) { - rb_warn("$, is set to non-nil value"); + rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "$, is set to non-nil value"); } } @@ -2413,32 +3089,32 @@ rb_ary_join_m(int argc, VALUE *argv, VALUE ary) static VALUE inspect_ary(VALUE ary, VALUE dummy, int recur) { - int tainted = OBJ_TAINTED(ary); long i; VALUE s, str; if (recur) return rb_usascii_str_new_cstr("[...]"); str = rb_str_buf_new2("["); for (i=0; i<RARRAY_LEN(ary); i++) { - s = rb_inspect(RARRAY_AREF(ary, i)); - if (OBJ_TAINTED(s)) tainted = TRUE; - if (i > 0) rb_str_buf_cat2(str, ", "); - else rb_enc_copy(str, s); - rb_str_buf_append(str, s); + s = rb_inspect(RARRAY_AREF(ary, i)); + if (i > 0) rb_str_buf_cat2(str, ", "); + else rb_enc_copy(str, s); + rb_str_buf_append(str, s); } rb_str_buf_cat2(str, "]"); - if (tainted) OBJ_TAINT(str); return str; } /* * call-seq: - * ary.inspect -> string - * ary.to_s -> string + * array.inspect -> new_string * - * Creates a string representation of +self+. + * Returns the new \String formed by calling method <tt>#inspect</tt> + * on each array element: * - * [ "a", "b", "c" ].to_s #=> "[\"a\", \"b\", \"c\"]" + * a = [:foo, 'bar', 2] + * a.inspect # => "[:foo, \"bar\", 2]" + * + * Array#to_s is an alias for Array#inspect. */ static VALUE @@ -2456,40 +3132,59 @@ rb_ary_to_s(VALUE ary) /* * call-seq: - * ary.to_a -> ary + * to_a -> self or new_array * - * Returns +self+. + * When +self+ is an instance of \Array, returns +self+: + * + * a = [:foo, 'bar', 2] + * a.to_a # => [:foo, "bar", 2] + * + * Otherwise, returns a new \Array containing the elements of +self+: + * + * class MyArray < Array; end + * a = MyArray.new(['foo', 'bar', 'two']) + * a.instance_of?(Array) # => false + * a.kind_of?(Array) # => true + * a1 = a.to_a + * a1 # => ["foo", "bar", "two"] + * a1.class # => Array # Not MyArray * - * If called on a subclass of Array, converts the receiver to an Array object. */ static VALUE rb_ary_to_a(VALUE ary) { if (rb_obj_class(ary) != rb_cArray) { - VALUE dup = rb_ary_new2(RARRAY_LEN(ary)); - rb_ary_replace(dup, ary); - return dup; + VALUE dup = rb_ary_new2(RARRAY_LEN(ary)); + rb_ary_replace(dup, ary); + return dup; } return ary; } /* * call-seq: - * ary.to_h -> hash - * ary.to_h {|item| block } -> hash + * array.to_h -> new_hash + * array.to_h {|item| ... } -> new_hash * - * Returns the result of interpreting <i>ary</i> as an array of - * <tt>[key, value]</tt> pairs. + * Returns a new \Hash formed from +self+. * - * [[:foo, :bar], [1, 2]].to_h - * # => {:foo => :bar, 1 => 2} + * When a block is given, calls the block with each array element; + * the block must return a 2-element \Array whose two elements + * form a key-value pair in the returned \Hash: * - * If a block is given, the results of the block on each element of - * the array will be used as pairs. + * a = ['foo', :bar, 1, [2, 3], {baz: 4}] + * h = a.to_h {|item| [item, item] } + * h # => {"foo"=>"foo", :bar=>:bar, 1=>1, [2, 3]=>[2, 3], {:baz=>4}=>{:baz=>4}} + * + * When no block is given, +self+ must be an \Array of 2-element sub-arrays, + * each sub-array is formed into a key-value pair in the new \Hash: + * + * [].to_h # => {} + * a = [['foo', 'zero'], ['bar', 'one'], ['baz', 'two']] + * h = a.to_h + * h # => {"foo"=>"zero", "bar"=>"one", "baz"=>"two"} * - * ["foo", "bar"].to_h {|s| [s.ord, s]} - * # => {102=>"foo", 98=>"bar"} */ static VALUE @@ -2500,25 +3195,25 @@ rb_ary_to_h(VALUE ary) int block_given = rb_block_given_p(); for (i=0; i<RARRAY_LEN(ary); i++) { - const VALUE e = rb_ary_elt(ary, i); - const VALUE elt = block_given ? rb_yield_force_blockarg(e) : e; - const VALUE key_value_pair = rb_check_array_type(elt); - if (NIL_P(key_value_pair)) { - rb_raise(rb_eTypeError, "wrong element type %"PRIsVALUE" at %ld (expected array)", - rb_obj_class(elt), i); - } - if (RARRAY_LEN(key_value_pair) != 2) { - rb_raise(rb_eArgError, "wrong array length at %ld (expected 2, was %ld)", - i, RARRAY_LEN(key_value_pair)); - } - rb_hash_aset(hash, RARRAY_AREF(key_value_pair, 0), RARRAY_AREF(key_value_pair, 1)); + const VALUE e = rb_ary_elt(ary, i); + const VALUE elt = block_given ? rb_yield_force_blockarg(e) : e; + const VALUE key_value_pair = rb_check_array_type(elt); + if (NIL_P(key_value_pair)) { + rb_raise(rb_eTypeError, "wrong element type %"PRIsVALUE" at %ld (expected array)", + rb_obj_class(elt), i); + } + if (RARRAY_LEN(key_value_pair) != 2) { + rb_raise(rb_eArgError, "wrong array length at %ld (expected 2, was %ld)", + i, RARRAY_LEN(key_value_pair)); + } + rb_hash_aset(hash, RARRAY_AREF(key_value_pair, 0), RARRAY_AREF(key_value_pair, 1)); } return hash; } /* * call-seq: - * ary.to_ary -> ary + * array.to_ary -> self * * Returns +self+. */ @@ -2533,9 +3228,9 @@ static void ary_reverse(VALUE *p1, VALUE *p2) { while (p1 < p2) { - VALUE tmp = *p1; - *p1++ = *p2; - *p2-- = tmp; + VALUE tmp = *p1; + *p1++ = *p2; + *p2-- = tmp; } } @@ -2550,20 +3245,20 @@ rb_ary_reverse(VALUE ary) RARRAY_PTR_USE_TRANSIENT(ary, p1, { p2 = p1 + len - 1; /* points last item */ ary_reverse(p1, p2); - }); /* WB: no new reference */ + }); /* WB: no new reference */ } return ary; } /* * call-seq: - * ary.reverse! -> ary + * array.reverse! -> self + * + * Reverses +self+ in place: * - * Reverses +self+ in place. + * a = ['foo', 'bar', 'two'] + * a.reverse! # => ["two", "bar", "foo"] * - * a = [ "a", "b", "c" ] - * a.reverse! #=> ["c", "b", "a"] - * a #=> ["c", "b", "a"] */ static VALUE @@ -2574,12 +3269,14 @@ rb_ary_reverse_bang(VALUE ary) /* * call-seq: - * ary.reverse -> new_ary + * array.reverse -> new_array * - * Returns a new array containing +self+'s elements in reverse order. + * Returns a new \Array with the elements of +self+ in reverse order: + * + * a = ['foo', 'bar', 'two'] + * a1 = a.reverse + * a1 # => ["two", "bar", "foo"] * - * [ "a", "b", "c" ].reverse #=> ["c", "b", "a"] - * [ 1 ].reverse #=> [1] */ static VALUE @@ -2591,7 +3288,7 @@ rb_ary_reverse_m(VALUE ary) if (len > 0) { const VALUE *p1 = RARRAY_CONST_PTR_TRANSIENT(ary); VALUE *p2 = (VALUE *)RARRAY_CONST_PTR_TRANSIENT(dup) + len - 1; - do *p2-- = *p1++; while (--len > 0); + do *p2-- = *p1++; while (--len > 0); } ARY_SET_LEN(dup, RARRAY_LEN(ary)); return dup; @@ -2606,10 +3303,22 @@ rotate_count(long cnt, long len) static void ary_rotate_ptr(VALUE *ptr, long len, long cnt) { - --len; - if (cnt < len) ary_reverse(ptr + cnt, ptr + len); - if (--cnt > 0) ary_reverse(ptr, ptr + cnt); - if (len > 0) ary_reverse(ptr, ptr + len); + if (cnt == 1) { + VALUE tmp = *ptr; + memmove(ptr, ptr + 1, sizeof(VALUE)*(len - 1)); + *(ptr + len - 1) = tmp; + } + else if (cnt == len - 1) { + VALUE tmp = *(ptr + len - 1); + memmove(ptr + 1, ptr, sizeof(VALUE)*(len - 1)); + *ptr = tmp; + } + else { + --len; + if (cnt < len) ary_reverse(ptr + cnt, ptr + len); + if (--cnt > 0) ary_reverse(ptr, ptr + cnt); + if (len > 0) ary_reverse(ptr, ptr + len); + } } VALUE @@ -2619,7 +3328,7 @@ rb_ary_rotate(VALUE ary, long cnt) if (cnt != 0) { long len = RARRAY_LEN(ary); - if (len > 0 && (cnt = rotate_count(cnt, len)) > 0) { + if (len > 1 && (cnt = rotate_count(cnt, len)) > 0) { RARRAY_PTR_USE_TRANSIENT(ary, ptr, ary_rotate_ptr(ptr, len, cnt)); return ary; } @@ -2629,19 +3338,48 @@ rb_ary_rotate(VALUE ary, long cnt) /* * call-seq: - * ary.rotate!(count=1) -> ary + * array.rotate! -> self + * array.rotate!(count) -> self * - * Rotates +self+ in place so that the element at +count+ comes first, and - * returns +self+. + * Rotates +self+ in place by moving elements from one end to the other; returns +self+. + * + * When no argument given, rotates the first element to the last position: + * + * a = [:foo, 'bar', 2, 'bar'] + * a.rotate! # => ["bar", 2, "bar", :foo] + * + * When given a non-negative \Integer +count+, + * rotates +count+ elements from the beginning to the end: + * + * a = [:foo, 'bar', 2] + * a.rotate!(2) + * a # => [2, :foo, "bar"] + * + * If +count+ is large, uses <tt>count % array.size</tt> as the count: * - * If +count+ is negative then it rotates in the opposite direction, starting - * from the end of the array where +-1+ is the last element. + * a = [:foo, 'bar', 2] + * a.rotate!(20) + * a # => [2, :foo, "bar"] + * + * If +count+ is zero, returns +self+ unmodified: + * + * a = [:foo, 'bar', 2] + * a.rotate!(0) + * a # => [:foo, "bar", 2] + * + * When given a negative Integer +count+, rotates in the opposite direction, + * from end to beginning: + * + * a = [:foo, 'bar', 2] + * a.rotate!(-2) + * a # => ["bar", 2, :foo] + * + * If +count+ is small (far from zero), uses <tt>count % array.size</tt> as the count: + * + * a = [:foo, 'bar', 2] + * a.rotate!(-5) + * a # => ["bar", 2, :foo] * - * a = [ "a", "b", "c", "d" ] - * a.rotate! #=> ["b", "c", "d", "a"] - * a #=> ["b", "c", "d", "a"] - * a.rotate!(2) #=> ["d", "a", "b", "c"] - * a.rotate!(-3) #=> ["a", "b", "c", "d"] */ static VALUE @@ -2654,19 +3392,51 @@ rb_ary_rotate_bang(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.rotate(count=1) -> new_ary + * array.rotate -> new_array + * array.rotate(count) -> new_array + * + * Returns a new \Array formed from +self+ with elements + * rotated from one end to the other. + * + * When no argument given, returns a new \Array that is like +self+, + * except that the first element has been rotated to the last position: + * + * a = [:foo, 'bar', 2, 'bar'] + * a1 = a.rotate + * a1 # => ["bar", 2, "bar", :foo] + * + * When given a non-negative \Integer +count+, + * returns a new \Array with +count+ elements rotated from the beginning to the end: * - * Returns a new array by rotating +self+ so that the element at +count+ is - * the first element of the new array. + * a = [:foo, 'bar', 2] + * a1 = a.rotate(2) + * a1 # => [2, :foo, "bar"] * - * If +count+ is negative then it rotates in the opposite direction, starting - * from the end of +self+ where +-1+ is the last element. + * If +count+ is large, uses <tt>count % array.size</tt> as the count: + * + * a = [:foo, 'bar', 2] + * a1 = a.rotate(20) + * a1 # => [2, :foo, "bar"] + * + * If +count+ is zero, returns a copy of +self+, unmodified: + * + * a = [:foo, 'bar', 2] + * a1 = a.rotate(0) + * a1 # => [:foo, "bar", 2] + * + * When given a negative \Integer +count+, rotates in the opposite direction, + * from end to beginning: + * + * a = [:foo, 'bar', 2] + * a1 = a.rotate(-2) + * a1 # => ["bar", 2, :foo] + * + * If +count+ is small (far from zero), uses <tt>count % array.size</tt> as the count: + * + * a = [:foo, 'bar', 2] + * a1 = a.rotate(-5) + * a1 # => ["bar", 2, :foo] * - * a = [ "a", "b", "c", "d" ] - * a.rotate #=> ["b", "c", "d", "a"] - * a #=> ["a", "b", "c", "d"] - * a.rotate(2) #=> ["c", "d", "a", "b"] - * a.rotate(-3) #=> ["b", "c", "d", "a"] */ static VALUE @@ -2680,11 +3450,11 @@ rb_ary_rotate_m(int argc, VALUE *argv, VALUE ary) len = RARRAY_LEN(ary); rotated = rb_ary_new2(len); if (len > 0) { - cnt = rotate_count(cnt, len); + cnt = rotate_count(cnt, len); ptr = RARRAY_CONST_PTR_TRANSIENT(ary); - len -= cnt; - ary_memcpy(rotated, 0, len, ptr + cnt); - ary_memcpy(rotated, len, cnt, ptr); + len -= cnt; + ary_memcpy(rotated, 0, len, ptr + cnt); + ary_memcpy(rotated, len, cnt, ptr); } ARY_SET_LEN(rotated, RARRAY_LEN(ary)); return rotated; @@ -2692,18 +3462,27 @@ rb_ary_rotate_m(int argc, VALUE *argv, VALUE ary) struct ary_sort_data { VALUE ary; - struct cmp_opt_data cmp_opt; + VALUE receiver; }; static VALUE sort_reentered(VALUE ary) { if (RBASIC(ary)->klass) { - rb_raise(rb_eRuntimeError, "sort reentered"); + rb_raise(rb_eRuntimeError, "sort reentered"); } return Qnil; } +static void +sort_returned(struct ary_sort_data *data) +{ + if (rb_obj_frozen_p(data->receiver)) { + rb_raise(rb_eFrozenError, "array frozen during sort"); + } + sort_reentered(data->ary); +} + static int sort_1(const void *ap, const void *bp, void *dummy) { @@ -2717,7 +3496,7 @@ sort_1(const void *ap, const void *bp, void *dummy) args[1] = b; retval = rb_yield_values2(2, args); n = rb_cmpint(retval, a, b); - sort_reentered(data->ary); + sort_returned(data); return n; } @@ -2729,47 +3508,64 @@ sort_2(const void *ap, const void *bp, void *dummy) VALUE a = *(const VALUE *)ap, b = *(const VALUE *)bp; int n; - if (FIXNUM_P(a) && FIXNUM_P(b) && CMP_OPTIMIZABLE(data->cmp_opt, Fixnum)) { - if ((long)a > (long)b) return 1; - if ((long)a < (long)b) return -1; - return 0; + if (FIXNUM_P(a) && FIXNUM_P(b) && CMP_OPTIMIZABLE(INTEGER)) { + if ((long)a > (long)b) return 1; + if ((long)a < (long)b) return -1; + return 0; } - if (STRING_P(a) && STRING_P(b) && CMP_OPTIMIZABLE(data->cmp_opt, String)) { - return rb_str_cmp(a, b); + if (STRING_P(a) && STRING_P(b) && CMP_OPTIMIZABLE(STRING)) { + return rb_str_cmp(a, b); } - if (RB_FLOAT_TYPE_P(a) && CMP_OPTIMIZABLE(data->cmp_opt, Float)) { - return rb_float_cmp(a, b); + if (RB_FLOAT_TYPE_P(a) && CMP_OPTIMIZABLE(FLOAT)) { + return rb_float_cmp(a, b); } retval = rb_funcallv(a, id_cmp, 1, &b); n = rb_cmpint(retval, a, b); - sort_reentered(data->ary); + sort_returned(data); return n; } /* * call-seq: - * ary.sort! -> ary - * ary.sort! {|a, b| block} -> ary + * array.sort! -> self + * array.sort! {|a, b| ... } -> self + * + * Returns +self+ with its elements sorted in place. * - * Sorts +self+ in place. + * With no block, compares elements using operator <tt><=></tt> + * (see Comparable): * - * Comparisons for the sort will be done using the <code><=></code> operator - * or using an optional code block. + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a.sort! + * a # => ["a", "b", "c", "d", "e"] * - * The block must implement a comparison between +a+ and +b+ and return - * an integer less than 0 when +b+ follows +a+, +0+ when +a+ and +b+ - * are equivalent, or an integer greater than 0 when +a+ follows +b+. + * With a block, calls the block with each element pair; + * for each element pair +a+ and +b+, the block should return an integer: * - * The result is not guaranteed to be stable. When the comparison of two - * elements returns +0+, the order of the elements is unpredictable. + * - Negative when +b+ is to follow +a+. + * - Zero when +a+ and +b+ are equivalent. + * - Positive when +a+ is to follow +b+. * - * ary = [ "d", "a", "e", "c", "b" ] - * ary.sort! #=> ["a", "b", "c", "d", "e"] - * ary.sort! {|a, b| b <=> a} #=> ["e", "d", "c", "b", "a"] + * Example: + * + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a.sort! {|a, b| a <=> b } + * a # => ["a", "b", "c", "d", "e"] + * a.sort! {|a, b| b <=> a } + * a # => ["e", "d", "c", "b", "a"] + * + * When the block returns zero, the order for +a+ and +b+ is indeterminate, + * and may be unstable: + * + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a.sort! {|a, b| 0 } + * a # => ["d", "e", "c", "a", "b"] * - * See also Enumerable#sort_by. */ VALUE @@ -2778,24 +3574,23 @@ rb_ary_sort_bang(VALUE ary) rb_ary_modify(ary); assert(!ARY_SHARED_P(ary)); if (RARRAY_LEN(ary) > 1) { - VALUE tmp = ary_make_substitution(ary); /* only ary refers tmp */ - struct ary_sort_data data; - long len = RARRAY_LEN(ary); - RBASIC_CLEAR_CLASS(tmp); - data.ary = tmp; - data.cmp_opt.opt_methods = 0; - data.cmp_opt.opt_inited = 0; - RARRAY_PTR_USE(tmp, ptr, { + VALUE tmp = ary_make_substitution(ary); /* only ary refers tmp */ + struct ary_sort_data data; + long len = RARRAY_LEN(ary); + RBASIC_CLEAR_CLASS(tmp); + data.ary = tmp; + data.receiver = ary; + RARRAY_PTR_USE(tmp, ptr, { ruby_qsort(ptr, len, sizeof(VALUE), rb_block_given_p()?sort_1:sort_2, &data); - }); /* WB: no new reference */ - rb_ary_modify(ary); + }); /* WB: no new reference */ + rb_ary_modify(ary); if (ARY_EMBED_P(tmp)) { if (ARY_SHARED_P(ary)) { /* ary might be destructively operated in the given block */ rb_ary_unshare(ary); - FL_SET_EMBED(ary); + FL_SET_EMBED(ary); } - ary_memcpy(ary, 0, ARY_EMBED_LEN(tmp), ARY_EMBED_PTR(tmp)); + ary_memcpy(ary, 0, ARY_EMBED_LEN(tmp), ARY_EMBED_PTR(tmp)); ARY_SET_LEN(ary, ARY_EMBED_LEN(tmp)); } else { @@ -2834,31 +3629,44 @@ rb_ary_sort_bang(VALUE ary) /* * call-seq: - * ary.sort -> new_ary - * ary.sort {|a, b| block} -> new_ary + * array.sort -> new_array + * array.sort {|a, b| ... } -> new_array * - * Returns a new array created by sorting +self+. + * Returns a new \Array whose elements are those from +self+, sorted. * - * Comparisons for the sort will be done using the <code><=></code> operator - * or using an optional code block. + * With no block, compares elements using operator <tt><=></tt> + * (see Comparable): * - * The block must implement a comparison between +a+ and +b+ and return - * an integer less than 0 when +b+ follows +a+, +0+ when +a+ and +b+ - * are equivalent, or an integer greater than 0 when +a+ follows +b+. + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a1 = a.sort + * a1 # => ["a", "b", "c", "d", "e"] * - * The result is not guaranteed to be stable. When the comparison of two - * elements returns +0+, the order of the elements is unpredictable. + * With a block, calls the block with each element pair; + * for each element pair +a+ and +b+, the block should return an integer: * - * ary = [ "d", "a", "e", "c", "b" ] - * ary.sort #=> ["a", "b", "c", "d", "e"] - * ary.sort {|a, b| b <=> a} #=> ["e", "d", "c", "b", "a"] + * - Negative when +b+ is to follow +a+. + * - Zero when +a+ and +b+ are equivalent. + * - Positive when +a+ is to follow +b+. * - * To produce the reverse order, the following can also be used - * (and may be faster): + * Example: * - * ary.sort.reverse! #=> ["e", "d", "c", "b", "a"] + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a1 = a.sort {|a, b| a <=> b } + * a1 # => ["a", "b", "c", "d", "e"] + * a2 = a.sort {|a, b| b <=> a } + * a2 # => ["e", "d", "c", "b", "a"] * - * See also Enumerable#sort_by. + * When the block returns zero, the order for +a+ and +b+ is indeterminate, + * and may be unstable: + * + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a1 = a.sort {|a, b| 0 } + * a1 # => ["c", "e", "b", "d", "a"] + * + * Related: Enumerable#sort_by. */ VALUE @@ -2873,55 +3681,12 @@ static VALUE rb_ary_bsearch_index(VALUE ary); /* * call-seq: - * ary.bsearch {|x| block } -> elem - * - * By using binary search, finds a value from this array which meets - * the given condition in O(log n) where n is the size of the array. - * - * You can use this method in two modes: a find-minimum mode and - * a find-any mode. In either case, the elements of the array must be - * monotone (or sorted) with respect to the block. - * - * In find-minimum mode (this is a good choice for typical use cases), - * the block must always return true or false, and there must be an index i - * (0 <= i <= ary.size) so that: - * - * - the block returns false for any element whose index is less than - * i, and - * - the block returns true for any element whose index is greater - * than or equal to i. - * - * This method returns the i-th element. If i is equal to ary.size, - * it returns nil. - * - * ary = [0, 4, 7, 10, 12] - * ary.bsearch {|x| x >= 4 } #=> 4 - * ary.bsearch {|x| x >= 6 } #=> 7 - * ary.bsearch {|x| x >= -1 } #=> 0 - * ary.bsearch {|x| x >= 100 } #=> nil - * - * In find-any mode (this behaves like libc's bsearch(3)), the block - * must always return a number, and there must be two indices i and j - * (0 <= i <= j <= ary.size) so that: - * - * - the block returns a positive number for ary[k] if 0 <= k < i, - * - the block returns zero for ary[k] if i <= k < j, and - * - the block returns a negative number for ary[k] if - * j <= k < ary.size. - * - * Under this condition, this method returns any element whose index - * is within i...j. If i is equal to j (i.e., there is no element - * that satisfies the block), this method returns nil. - * - * ary = [0, 4, 7, 10, 12] - * # try to find v such that 4 <= v < 8 - * ary.bsearch {|x| 1 - x / 4 } #=> 4 or 7 - * # try to find v such that 8 <= v < 10 - * ary.bsearch {|x| 4 - x / 2 } #=> nil - * - * You must not mix the two modes at a time; the block must always - * return either true/false, or always return a number. It is - * undefined which value is actually picked up at each iteration. + * array.bsearch {|element| ... } -> object + * array.bsearch -> new_enumerator + * + * Returns an element from +self+ selected by a binary search. + * + * See {Binary Searching}[rdoc-ref:bsearch.rdoc]. */ static VALUE @@ -2930,22 +3695,18 @@ rb_ary_bsearch(VALUE ary) VALUE index_result = rb_ary_bsearch_index(ary); if (FIXNUM_P(index_result)) { - return rb_ary_entry(ary, FIX2LONG(index_result)); + return rb_ary_entry(ary, FIX2LONG(index_result)); } return index_result; } /* * call-seq: - * ary.bsearch_index {|x| block } -> int or nil - * - * By using binary search, finds an index of a value from this array which - * meets the given condition in O(log n) where n is the size of the array. + * array.bsearch_index {|element| ... } -> integer or nil + * array.bsearch_index -> new_enumerator * - * It supports two modes, depending on the nature of the block. They are - * exactly the same as in the case of the #bsearch method, with the only difference - * being that this method returns the index of the element instead of the - * element itself. For more details consult the documentation for #bsearch. + * Searches +self+ as described at method #bsearch, + * but returns the _index_ of the found element instead of the element itself. */ static VALUE @@ -2957,39 +3718,39 @@ rb_ary_bsearch_index(VALUE ary) RETURN_ENUMERATOR(ary, 0, 0); while (low < high) { - mid = low + ((high - low) / 2); - val = rb_ary_entry(ary, mid); - v = rb_yield(val); - if (FIXNUM_P(v)) { - if (v == INT2FIX(0)) return INT2FIX(mid); - smaller = (SIGNED_VALUE)v < 0; /* Fixnum preserves its sign-bit */ - } - else if (v == Qtrue) { - satisfied = 1; - smaller = 1; - } - else if (v == Qfalse || v == Qnil) { - smaller = 0; - } - else if (rb_obj_is_kind_of(v, rb_cNumeric)) { - const VALUE zero = INT2FIX(0); - switch (rb_cmpint(rb_funcallv(v, id_cmp, 1, &zero), v, zero)) { - case 0: return INT2FIX(mid); - case 1: smaller = 1; break; - case -1: smaller = 0; - } - } - else { - rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE - " (must be numeric, true, false or nil)", - rb_obj_class(v)); - } - if (smaller) { - high = mid; - } - else { - low = mid + 1; - } + mid = low + ((high - low) / 2); + val = rb_ary_entry(ary, mid); + v = rb_yield(val); + if (FIXNUM_P(v)) { + if (v == INT2FIX(0)) return INT2FIX(mid); + smaller = (SIGNED_VALUE)v < 0; /* Fixnum preserves its sign-bit */ + } + else if (v == Qtrue) { + satisfied = 1; + smaller = 1; + } + else if (!RTEST(v)) { + smaller = 0; + } + else if (rb_obj_is_kind_of(v, rb_cNumeric)) { + const VALUE zero = INT2FIX(0); + switch (rb_cmpint(rb_funcallv(v, id_cmp, 1, &zero), v, zero)) { + case 0: return INT2FIX(mid); + case 1: smaller = 0; break; + case -1: smaller = 1; + } + } + else { + rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE + " (must be numeric, true, false or nil)", + rb_obj_class(v)); + } + if (smaller) { + high = mid; + } + else { + low = mid + 1; + } } if (!satisfied) return Qnil; return INT2FIX(low); @@ -3004,18 +3765,28 @@ sort_by_i(RB_BLOCK_CALL_FUNC_ARGLIST(i, dummy)) /* * call-seq: - * ary.sort_by! {|obj| block} -> ary - * ary.sort_by! -> Enumerator + * array.sort_by! {|element| ... } -> self + * array.sort_by! -> new_enumerator + * + * Sorts the elements of +self+ in place, + * using an ordering determined by the block; returns self. * - * Sorts +self+ in place using a set of keys generated by mapping the - * values in +self+ through the given block. + * Calls the block with each successive element; + * sorts elements based on the values returned from the block. * - * The result is not guaranteed to be stable. When two keys are equal, - * the order of the corresponding elements is unpredictable. + * For duplicates returned by the block, the ordering is indeterminate, and may be unstable. * - * If no block is given, an Enumerator is returned instead. + * This example sorts strings based on their sizes: + * + * a = ['aaaa', 'bbb', 'cc', 'd'] + * a.sort_by! {|element| element.size } + * a # => ["d", "cc", "bbb", "aaaa"] + * + * Returns a new \Enumerator if no block given: + * + * a = ['aaaa', 'bbb', 'cc', 'd'] + * a.sort_by! # => #<Enumerator: ["aaaa", "bbb", "cc", "d"]:sort_by!> * - * See also Enumerable#sort_by. */ static VALUE @@ -3033,23 +3804,22 @@ rb_ary_sort_by_bang(VALUE ary) /* * call-seq: - * ary.collect {|item| block} -> new_ary - * ary.map {|item| block} -> new_ary - * ary.collect -> Enumerator - * ary.map -> Enumerator + * array.map {|element| ... } -> new_array + * array.map -> new_enumerator * - * Invokes the given block once for each element of +self+. + * Calls the block, if given, with each element of +self+; + * returns a new \Array whose elements are the return values from the block: * - * Creates a new array containing the values returned by the block. + * a = [:foo, 'bar', 2] + * a1 = a.map {|element| element.class } + * a1 # => [Symbol, String, Integer] * - * See also Enumerable#collect. + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2] + * a1 = a.map + * a1 # => #<Enumerator: [:foo, "bar", 2]:map> * - * If no block is given, an Enumerator is returned instead. - * - * a = [ "a", "b", "c", "d" ] - * a.collect {|x| x + "!"} #=> ["a!", "b!", "c!", "d!"] - * a.map.with_index {|x, i| x * i} #=> ["", "b", "cc", "ddd"] - * a #=> ["a", "b", "c", "d"] + * Array#collect is an alias for Array#map. */ static VALUE @@ -3069,23 +3839,22 @@ rb_ary_collect(VALUE ary) /* * call-seq: - * ary.collect! {|item| block } -> ary - * ary.map! {|item| block } -> ary - * ary.collect! -> Enumerator - * ary.map! -> Enumerator + * array.map! {|element| ... } -> self + * array.map! -> new_enumerator + * + * Calls the block, if given, with each element; + * replaces the element with the block's return value: * - * Invokes the given block once for each element of +self+, replacing the - * element with the value returned by the block. + * a = [:foo, 'bar', 2] + * a.map! { |element| element.class } # => [Symbol, String, Integer] * - * See also Enumerable#collect. + * Returns a new \Enumerator if no block given: * - * If no block is given, an Enumerator is returned instead. + * a = [:foo, 'bar', 2] + * a1 = a.map! + * a1 # => #<Enumerator: [:foo, "bar", 2]:map!> * - * a = [ "a", "b", "c", "d" ] - * a.map! {|x| x + "!" } - * a #=> [ "a!", "b!", "c!", "d!" ] - * a.collect!.with_index {|x, i| x[0...i] } - * a #=> ["", "b", "c!", "d!"] + * Array#collect! is an alias for Array#map!. */ static VALUE @@ -3096,7 +3865,7 @@ rb_ary_collect_bang(VALUE ary) RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); rb_ary_modify(ary); for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_ary_store(ary, i, rb_yield(RARRAY_AREF(ary, i))); + rb_ary_store(ary, i, rb_yield(RARRAY_AREF(ary, i))); } return ary; } @@ -3108,21 +3877,21 @@ rb_get_values_at(VALUE obj, long olen, int argc, const VALUE *argv, VALUE (*func long beg, len, i, j; for (i=0; i<argc; i++) { - if (FIXNUM_P(argv[i])) { - rb_ary_push(result, (*func)(obj, FIX2LONG(argv[i]))); - continue; - } - /* check if idx is Range */ - if (rb_range_beg_len(argv[i], &beg, &len, olen, 1)) { - long end = olen < beg+len ? olen : beg+len; - for (j = beg; j < end; j++) { - rb_ary_push(result, (*func)(obj, j)); - } - if (beg + len > j) - rb_ary_resize(result, RARRAY_LEN(result) + (beg + len) - j); - continue; - } - rb_ary_push(result, (*func)(obj, NUM2LONG(argv[i]))); + if (FIXNUM_P(argv[i])) { + rb_ary_push(result, (*func)(obj, FIX2LONG(argv[i]))); + continue; + } + /* check if idx is Range */ + if (rb_range_beg_len(argv[i], &beg, &len, olen, 1)) { + long end = olen < beg+len ? olen : beg+len; + for (j = beg; j < end; j++) { + rb_ary_push(result, (*func)(obj, j)); + } + if (beg + len > j) + rb_ary_resize(result, RARRAY_LEN(result) + (beg + len) - j); + continue; + } + rb_ary_push(result, (*func)(obj, NUM2LONG(argv[i]))); } return result; } @@ -3132,45 +3901,70 @@ append_values_at_single(VALUE result, VALUE ary, long olen, VALUE idx) { long beg, len; if (FIXNUM_P(idx)) { - beg = FIX2LONG(idx); + beg = FIX2LONG(idx); } /* check if idx is Range */ else if (rb_range_beg_len(idx, &beg, &len, olen, 1)) { - if (len > 0) { + if (len > 0) { const VALUE *const src = RARRAY_CONST_PTR_TRANSIENT(ary); - const long end = beg + len; - const long prevlen = RARRAY_LEN(result); - if (beg < olen) { - rb_ary_cat(result, src + beg, end > olen ? olen-beg : len); - } - if (end > olen) { - rb_ary_store(result, prevlen + len - 1, Qnil); - } - } - return result; + const long end = beg + len; + const long prevlen = RARRAY_LEN(result); + if (beg < olen) { + rb_ary_cat(result, src + beg, end > olen ? olen-beg : len); + } + if (end > olen) { + rb_ary_store(result, prevlen + len - 1, Qnil); + } + } + return result; } else { - beg = NUM2LONG(idx); + beg = NUM2LONG(idx); } return rb_ary_push(result, rb_ary_entry(ary, beg)); } /* * call-seq: - * ary.values_at(selector, ...) -> new_ary + * array.values_at(*indexes) -> new_array + * + * Returns a new \Array whose elements are the elements + * of +self+ at the given \Integer or \Range +indexes+. + * + * For each positive +index+, returns the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a.values_at(0, 2) # => [:foo, 2] + * a.values_at(0..1) # => [:foo, "bar"] + * + * The given +indexes+ may be in any order, and may repeat: * - * Returns an array containing the elements in +self+ corresponding to the - * given +selector+(s). + * a = [:foo, 'bar', 2] + * a.values_at(2, 0, 1, 0, 2) # => [2, :foo, "bar", :foo, 2] + * a.values_at(1, 0..2) # => ["bar", :foo, "bar", 2] * - * The selectors may be either integer indices or ranges. + * Assigns +nil+ for an +index+ that is too large: * - * See also Array#select. + * a = [:foo, 'bar', 2] + * a.values_at(0, 3, 1, 3) # => [:foo, nil, "bar", nil] + * + * Returns a new empty \Array if no arguments given. + * + * For each negative +index+, counts backward from the end of the array: + * + * a = [:foo, 'bar', 2] + * a.values_at(-1, -3) # => [2, :foo] + * + * Assigns +nil+ for an +index+ that is too small: + * + * a = [:foo, 'bar', 2] + * a.values_at(0, -5, 1, -6, 2) # => [:foo, nil, "bar", nil, 2] + * + * The given +indexes+ may have a mixture of signs: + * + * a = [:foo, 'bar', 2] + * a.values_at(0, -2, 1, -1) # => [:foo, "bar", "bar", 2] * - * a = %w{ a b c d e f } - * a.values_at(1, 3, 5) # => ["b", "d", "f"] - * a.values_at(1, 3, 5, 7) # => ["b", "d", "f", nil] - * a.values_at(-1, -2, -2, -7) # => ["f", "e", "e", nil] - * a.values_at(4..6, 3...6) # => ["e", "f", nil, "d", "e", "f"] */ static VALUE @@ -3179,7 +3973,7 @@ rb_ary_values_at(int argc, VALUE *argv, VALUE ary) long i, olen = RARRAY_LEN(ary); VALUE result = rb_ary_new_capa(argc); for (i = 0; i < argc; ++i) { - append_values_at_single(result, ary, olen, argv[i]); + append_values_at_single(result, ary, olen, argv[i]); } RB_GC_GUARD(ary); return result; @@ -3188,22 +3982,21 @@ rb_ary_values_at(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.select {|item| block} -> new_ary - * ary.select -> Enumerator - * ary.filter {|item| block} -> new_ary - * ary.filter -> Enumerator - * - * Returns a new array containing all elements of +ary+ - * for which the given +block+ returns a true value. + * array.select {|element| ... } -> new_array + * array.select -> new_enumerator * - * If no block is given, an Enumerator is returned instead. + * Calls the block, if given, with each element of +self+; + * returns a new \Array containing those elements of +self+ + * for which the block returns a truthy value: * - * [1,2,3,4,5].select {|num| num.even? } #=> [2, 4] + * a = [:foo, 'bar', 2, :bam] + * a1 = a.select {|element| element.to_s.start_with?('b') } + * a1 # => ["bar", :bam] * - * a = %w[ a b c d e f ] - * a.select {|v| v =~ /[aeiou]/ } #=> ["a", "e"] + * Returns a new \Enumerator if no block given: * - * See also Enumerable#select. + * a = [:foo, 'bar', 2, :bam] + * a.select # => #<Enumerator: [:foo, "bar", 2, :bam]:select> * * Array#filter is an alias for Array#select. */ @@ -3217,9 +4010,9 @@ rb_ary_select(VALUE ary) RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); result = rb_ary_new2(RARRAY_LEN(ary)); for (i = 0; i < RARRAY_LEN(ary); i++) { - if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { - rb_ary_push(result, rb_ary_elt(ary, i)); - } + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { + rb_ary_push(result, rb_ary_elt(ary, i)); + } } return result; } @@ -3237,12 +4030,12 @@ select_bang_i(VALUE a) long i1, i2; for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); arg->len[0] = ++i1) { - VALUE v = RARRAY_AREF(ary, i1); - if (!RTEST(rb_yield(v))) continue; - if (i1 != i2) { - rb_ary_store(ary, i2, v); - } - arg->len[1] = ++i2; + VALUE v = RARRAY_AREF(ary, i1); + if (!RTEST(rb_yield(v))) continue; + if (i1 != i2) { + rb_ary_store(ary, i2, v); + } + arg->len[1] = ++i2; } return (i1 == i2) ? Qnil : ary; } @@ -3256,35 +4049,38 @@ select_bang_ensure(VALUE a) long i1 = arg->len[0], i2 = arg->len[1]; if (i2 < len && i2 < i1) { - long tail = 0; - if (i1 < len) { - tail = len - i1; + long tail = 0; + rb_ary_modify(ary); + if (i1 < len) { + tail = len - i1; RARRAY_PTR_USE_TRANSIENT(ary, ptr, { - MEMMOVE(ptr + i2, ptr + i1, VALUE, tail); - }); - } - ARY_SET_LEN(ary, i2 + tail); + MEMMOVE(ptr + i2, ptr + i1, VALUE, tail); + }); + } + ARY_SET_LEN(ary, i2 + tail); } return ary; } /* * call-seq: - * ary.select! {|item| block } -> ary or nil - * ary.select! -> Enumerator - * ary.filter! {|item| block } -> ary or nil - * ary.filter! -> Enumerator + * array.select! {|element| ... } -> self or nil + * array.select! -> new_enumerator + * + * Calls the block, if given with each element of +self+; + * removes from +self+ those elements for which the block returns +false+ or +nil+. * - * Invokes the given block passing in successive elements from +self+, - * deleting elements for which the block returns a +false+ value. + * Returns +self+ if any elements were removed: * - * The array may not be changed instantly every time the block is called. + * a = [:foo, 'bar', 2, :bam] + * a.select! {|element| element.to_s.start_with?('b') } # => ["bar", :bam] * - * If changes were made, it will return +self+, otherwise it returns +nil+. + * Returns +nil+ if no elements were removed. * - * If no block is given, an Enumerator is returned instead. + * Returns a new \Enumerator if no block given: * - * See also Array#keep_if. + * a = [:foo, 'bar', 2, :bam] + * a.select! # => #<Enumerator: [:foo, "bar", 2, :bam]:select!> * * Array#filter! is an alias for Array#select!. */ @@ -3304,19 +4100,20 @@ rb_ary_select_bang(VALUE ary) /* * call-seq: - * ary.keep_if {|item| block} -> ary - * ary.keep_if -> Enumerator + * array.keep_if {|element| ... } -> self + * array.keep_if -> new_enumeration * - * Deletes every element of +self+ for which the given block evaluates to - * +false+, and returns +self+. + * Retains those elements for which the block returns a truthy value; + * deletes all other elements; returns +self+: * - * If no block is given, an Enumerator is returned instead. + * a = [:foo, 'bar', 2, :bam] + * a.keep_if {|element| element.to_s.start_with?('b') } # => ["bar", :bam] * - * a = %w[ a b c d e f ] - * a.keep_if {|v| v =~ /[aeiou]/ } #=> ["a", "e"] - * a #=> ["a", "e"] + * Returns a new \Enumerator if no block given: + * + * a = [:foo, 'bar', 2, :bam] + * a.keep_if # => #<Enumerator: [:foo, "bar", 2, :bam]:keep_if> * - * See also Array#select!. */ static VALUE @@ -3332,32 +4129,48 @@ ary_resize_smaller(VALUE ary, long len) { rb_ary_modify(ary); if (RARRAY_LEN(ary) > len) { - ARY_SET_LEN(ary, len); - if (len * 2 < ARY_CAPA(ary) && - ARY_CAPA(ary) > ARY_DEFAULT_SIZE) { - ary_resize_capa(ary, len * 2); - } + ARY_SET_LEN(ary, len); + if (len * 2 < ARY_CAPA(ary) && + ARY_CAPA(ary) > ARY_DEFAULT_SIZE) { + ary_resize_capa(ary, len * 2); + } } } /* * call-seq: - * ary.delete(obj) -> item or nil - * ary.delete(obj) {block} -> item or result of block + * array.delete(obj) -> deleted_object + * array.delete(obj) {|nosuch| ... } -> deleted_object or block_return + * + * Removes zero or more elements from +self+. + * + * When no block is given, + * removes from +self+ each element +ele+ such that <tt>ele == obj</tt>; + * returns the last deleted element: + * + * s1 = 'bar'; s2 = 'bar' + * a = [:foo, s1, 2, s2] + * a.delete('bar') # => "bar" + * a # => [:foo, 2] + * + * Returns +nil+ if no elements removed. + * + * When a block is given, + * removes from +self+ each element +ele+ such that <tt>ele == obj</tt>. + * + * If any such elements are found, ignores the block + * and returns the last deleted element: * - * Deletes all items from +self+ that are equal to +obj+. + * s1 = 'bar'; s2 = 'bar' + * a = [:foo, s1, 2, s2] + * deleted_obj = a.delete('bar') {|obj| fail 'Cannot happen' } + * a # => [:foo, 2] * - * Returns the last deleted item, or +nil+ if no matching item is found. + * If no such elements are found, returns the block's return value: * - * If the optional code block is given, the result of the block is returned if - * the item is not found. (To remove +nil+ elements and get an informative - * return value, use Array#compact!) + * a = [:foo, 'bar', 2] + * a.delete(:nosuch) {|obj| "#{obj} not found" } # => "nosuch not found" * - * a = [ "a", "b", "b", "b", "c" ] - * a.delete("b") #=> "b" - * a #=> ["a", "c"] - * a.delete("z") #=> nil - * a.delete("z") {"not found"} #=> "not found" */ VALUE @@ -3367,22 +4180,22 @@ rb_ary_delete(VALUE ary, VALUE item) long i1, i2; for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); i1++) { - VALUE e = RARRAY_AREF(ary, i1); + VALUE e = RARRAY_AREF(ary, i1); - if (rb_equal(e, item)) { - v = e; - continue; - } - if (i1 != i2) { - rb_ary_store(ary, i2, e); - } - i2++; + if (rb_equal(e, item)) { + v = e; + continue; + } + if (i1 != i2) { + rb_ary_store(ary, i2, e); + } + i2++; } if (RARRAY_LEN(ary) == i2) { - if (rb_block_given_p()) { - return rb_yield(item); - } - return Qnil; + if (rb_block_given_p()) { + return rb_yield(item); + } + return Qnil; } ary_resize_smaller(ary, i2); @@ -3397,18 +4210,18 @@ rb_ary_delete_same(VALUE ary, VALUE item) long i1, i2; for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); i1++) { - VALUE e = RARRAY_AREF(ary, i1); + VALUE e = RARRAY_AREF(ary, i1); - if (e == item) { - continue; - } - if (i1 != i2) { - rb_ary_store(ary, i2, e); - } - i2++; + if (e == item) { + continue; + } + if (i1 != i2) { + rb_ary_store(ary, i2, e); + } + i2++; } if (RARRAY_LEN(ary) == i2) { - return; + return; } ary_resize_smaller(ary, i2); @@ -3422,8 +4235,8 @@ rb_ary_delete_at(VALUE ary, long pos) if (pos >= len) return Qnil; if (pos < 0) { - pos += len; - if (pos < 0) return Qnil; + pos += len; + if (pos < 0) return Qnil; } rb_ary_modify(ary); @@ -3438,17 +4251,25 @@ rb_ary_delete_at(VALUE ary, long pos) /* * call-seq: - * ary.delete_at(index) -> obj or nil + * array.delete_at(index) -> deleted_object or nil + * + * Deletes an element from +self+, per the given \Integer +index+. + * + * When +index+ is non-negative, deletes the element at offset +index+: + * + * a = [:foo, 'bar', 2] + * a.delete_at(1) # => "bar" + * a # => [:foo, 2] * - * Deletes the element at the specified +index+, returning that element, or - * +nil+ if the +index+ is out of range. + * If index is too large, returns +nil+. * - * See also Array#slice! + * When +index+ is negative, counts backward from the end of the array: * - * a = ["ant", "bat", "cat", "dog"] - * a.delete_at(2) #=> "cat" - * a #=> ["ant", "bat", "dog"] - * a.delete_at(99) #=> nil + * a = [:foo, 'bar', 2] + * a.delete_at(-2) # => "bar" + * a # => [:foo, 2] + * + * If +index+ is too small (far from zero), returns nil. */ static VALUE @@ -3457,70 +4278,133 @@ rb_ary_delete_at_m(VALUE ary, VALUE pos) return rb_ary_delete_at(ary, NUM2LONG(pos)); } +static VALUE +ary_slice_bang_by_rb_ary_splice(VALUE ary, long pos, long len) +{ + const long orig_len = RARRAY_LEN(ary); + + if (len < 0) { + return Qnil; + } + else if (pos < -orig_len) { + return Qnil; + } + else if (pos < 0) { + pos += orig_len; + } + else if (orig_len < pos) { + return Qnil; + } + if (orig_len < pos + len) { + len = orig_len - pos; + } + if (len == 0) { + return rb_ary_new2(0); + } + else { + VALUE arg2 = rb_ary_new4(len, RARRAY_CONST_PTR_TRANSIENT(ary)+pos); + rb_ary_splice(ary, pos, len, 0, 0); + return arg2; + } +} + /* * call-seq: - * ary.slice!(index) -> obj or nil - * ary.slice!(start, length) -> new_ary or nil - * ary.slice!(range) -> new_ary or nil - * - * Deletes the element(s) given by an +index+ (optionally up to +length+ - * elements) or by a +range+. - * - * Returns the deleted object (or objects), or +nil+ if the +index+ is out of - * range. - * - * a = [ "a", "b", "c" ] - * a.slice!(1) #=> "b" - * a #=> ["a", "c"] - * a.slice!(-1) #=> "c" - * a #=> ["a"] - * a.slice!(100) #=> nil - * a #=> ["a"] + * array.slice!(n) -> object or nil + * array.slice!(start, length) -> new_array or nil + * array.slice!(range) -> new_array or nil + * + * Removes and returns elements from +self+. + * + * When the only argument is an \Integer +n+, + * removes and returns the _nth_ element in +self+: + * + * a = [:foo, 'bar', 2] + * a.slice!(1) # => "bar" + * a # => [:foo, 2] + * + * If +n+ is negative, counts backwards from the end of +self+: + * + * a = [:foo, 'bar', 2] + * a.slice!(-1) # => 2 + * a # => [:foo, "bar"] + * + * If +n+ is out of range, returns +nil+. + * + * When the only arguments are Integers +start+ and +length+, + * removes +length+ elements from +self+ beginning at offset +start+; + * returns the deleted objects in a new \Array: + * + * a = [:foo, 'bar', 2] + * a.slice!(0, 2) # => [:foo, "bar"] + * a # => [2] + * + * If <tt>start + length</tt> exceeds the array size, + * removes and returns all elements from offset +start+ to the end: + * + * a = [:foo, 'bar', 2] + * a.slice!(1, 50) # => ["bar", 2] + * a # => [:foo] + * + * If <tt>start == a.size</tt> and +length+ is non-negative, + * returns a new empty \Array. + * + * If +length+ is negative, returns +nil+. + * + * When the only argument is a \Range object +range+, + * treats <tt>range.min</tt> as +start+ above and <tt>range.size</tt> as +length+ above: + * + * a = [:foo, 'bar', 2] + * a.slice!(1..2) # => ["bar", 2] + * a # => [:foo] + * + * If <tt>range.start == a.size</tt>, returns a new empty \Array. + * + * If <tt>range.start</tt> is larger than the array size, returns +nil+. + * + * If <tt>range.end</tt> is negative, counts backwards from the end of the array: + * + * a = [:foo, 'bar', 2] + * a.slice!(0..-2) # => [:foo, "bar"] + * a # => [2] + * + * If <tt>range.start</tt> is negative, + * calculates the start index backwards from the end of the array: + * + * a = [:foo, 'bar', 2] + * a.slice!(-2..2) # => ["bar", 2] + * a # => [:foo] + * */ static VALUE rb_ary_slice_bang(int argc, VALUE *argv, VALUE ary) { - VALUE arg1, arg2; - long pos, len, orig_len; + VALUE arg1; + long pos, len; rb_ary_modify_check(ary); - if (argc == 2) { - pos = NUM2LONG(argv[0]); - len = NUM2LONG(argv[1]); - delete_pos_len: - if (len < 0) return Qnil; - orig_len = RARRAY_LEN(ary); - if (pos < 0) { - pos += orig_len; - if (pos < 0) return Qnil; - } - else if (orig_len < pos) return Qnil; - if (orig_len < pos + len) { - len = orig_len - pos; - } - if (len == 0) return rb_ary_new2(0); - arg2 = rb_ary_new4(len, RARRAY_CONST_PTR_TRANSIENT(ary)+pos); - RBASIC_SET_CLASS(arg2, rb_obj_class(ary)); - rb_ary_splice(ary, pos, len, 0, 0); - return arg2; - } - rb_check_arity(argc, 1, 2); arg1 = argv[0]; + if (argc == 2) { + pos = NUM2LONG(argv[0]); + len = NUM2LONG(argv[1]); + return ary_slice_bang_by_rb_ary_splice(ary, pos, len); + } + if (!FIXNUM_P(arg1)) { - switch (rb_range_beg_len(arg1, &pos, &len, RARRAY_LEN(ary), 0)) { - case Qtrue: - /* valid range */ - goto delete_pos_len; - case Qnil: - /* invalid range */ - return Qnil; - default: - /* not a range */ - break; - } + switch (rb_range_beg_len(arg1, &pos, &len, RARRAY_LEN(ary), 0)) { + case Qtrue: + /* valid range */ + return ary_slice_bang_by_rb_ary_splice(ary, pos, len); + case Qnil: + /* invalid range */ + return Qnil; + default: + /* not a range */ + break; + } } return rb_ary_delete_at(ary, NUM2LONG(arg1)); @@ -3532,11 +4416,11 @@ ary_reject(VALUE orig, VALUE result) long i; for (i = 0; i < RARRAY_LEN(orig); i++) { - VALUE v = RARRAY_AREF(orig, i); + VALUE v = RARRAY_AREF(orig, i); if (!RTEST(rb_yield(v))) { - rb_ary_push(result, v); - } + rb_ary_push(result, v); + } } return result; } @@ -3549,12 +4433,12 @@ reject_bang_i(VALUE a) long i1, i2; for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); arg->len[0] = ++i1) { - VALUE v = RARRAY_AREF(ary, i1); - if (RTEST(rb_yield(v))) continue; - if (i1 != i2) { - rb_ary_store(ary, i2, v); - } - arg->len[1] = ++i2; + VALUE v = RARRAY_AREF(ary, i1); + if (RTEST(rb_yield(v))) continue; + if (i1 != i2) { + rb_ary_store(ary, i2, v); + } + arg->len[1] = ++i2; } return (i1 == i2) ? Qnil : ary; } @@ -3571,17 +4455,23 @@ ary_reject_bang(VALUE ary) /* * call-seq: - * ary.reject! {|item| block} -> ary or nil - * ary.reject! -> Enumerator + * array.reject! {|element| ... } -> self or nil + * array.reject! -> new_enumerator + * + * Removes each element for which the block returns a truthy value. + * + * Returns +self+ if any elements removed: + * + * a = [:foo, 'bar', 2, 'bat'] + * a.reject! {|element| element.to_s.start_with?('b') } # => [:foo, 2] * - * Deletes every element of +self+ for which the block evaluates to +true+, - * if no changes were made returns +nil+. + * Returns +nil+ if no elements removed. * - * The array may not be changed instantly every time the block is called. + * Returns a new \Enumerator if no block given: * - * See also Enumerable#reject and Array#delete_if. + * a = [:foo, 'bar', 2] + * a.reject! # => #<Enumerator: [:foo, "bar", 2]:reject!> * - * If no block is given, an Enumerator is returned instead. */ static VALUE @@ -3594,15 +4484,21 @@ rb_ary_reject_bang(VALUE ary) /* * call-seq: - * ary.reject {|item| block } -> new_ary - * ary.reject -> Enumerator + * array.reject {|element| ... } -> new_array + * array.reject -> new_enumerator * - * Returns a new array containing the items in +self+ for which the given - * block is not +true+. The ordering of non-rejected elements is maintained. + * Returns a new \Array whose elements are all those from +self+ + * for which the block returns +false+ or +nil+: * - * See also Array#delete_if + * a = [:foo, 'bar', 2, 'bat'] + * a1 = a.reject {|element| element.to_s.start_with?('b') } + * a1 # => [:foo, 2] + * + * Returns a new \Enumerator if no block given: + * + * a = [:foo, 'bar', 2] + * a.reject # => #<Enumerator: [:foo, "bar", 2]:reject> * - * If no block is given, an Enumerator is returned instead. */ static VALUE @@ -3618,21 +4514,21 @@ rb_ary_reject(VALUE ary) /* * call-seq: - * ary.delete_if {|item| block} -> ary - * ary.delete_if -> Enumerator + * array.delete_if {|element| ... } -> self + * array.delete_if -> Enumerator * - * Deletes every element of +self+ for which block evaluates to +true+. + * Removes each element in +self+ for which the block returns a truthy value; + * returns +self+: * - * The array is changed instantly every time the block is called, not after - * the iteration is over. + * a = [:foo, 'bar', 2, 'bat'] + * a.delete_if {|element| element.to_s.start_with?('b') } # => [:foo, 2] * - * See also Array#reject! + * Returns a new \Enumerator if no block given: * - * If no block is given, an Enumerator is returned instead. + * a = [:foo, 'bar', 2] + * a.delete_if # => #<Enumerator: [:foo, "bar", 2]:delete_if> * - * scores = [ 97, 42, 75 ] - * scores.delete_if {|score| score < 80 } #=> [97] - */ +3 */ static VALUE rb_ary_delete_if(VALUE ary) @@ -3647,10 +4543,9 @@ static VALUE take_i(RB_BLOCK_CALL_FUNC_ARGLIST(val, cbarg)) { VALUE *args = (VALUE *)cbarg; - if (args[1] == 0) rb_iter_break(); - else args[1]--; if (argc > 1) val = rb_ary_new4(argc, argv); rb_ary_push(args[0], val); + if (--args[1] == 0) rb_iter_break(); return Qnil; } @@ -3660,38 +4555,71 @@ take_items(VALUE obj, long n) VALUE result = rb_check_array_type(obj); VALUE args[2]; + if (n == 0) return result; if (!NIL_P(result)) return rb_ary_subseq(result, 0, n); result = rb_ary_new2(n); args[0] = result; args[1] = (VALUE)n; - if (rb_check_block_call(obj, idEach, 0, 0, take_i, (VALUE)args) == Qundef) - rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE" (must respond to :each)", - rb_obj_class(obj)); + if (UNDEF_P(rb_check_block_call(obj, idEach, 0, 0, take_i, (VALUE)args))) + rb_raise(rb_eTypeError, "wrong argument type %"PRIsVALUE" (must respond to :each)", + rb_obj_class(obj)); return result; } /* * call-seq: - * ary.zip(arg, ...) -> new_ary - * ary.zip(arg, ...) {|arr| block} -> nil + * array.zip(*other_arrays) -> new_array + * array.zip(*other_arrays) {|other_array| ... } -> nil + * + * When no block given, returns a new \Array +new_array+ of size <tt>self.size</tt> + * whose elements are Arrays. + * + * Each nested array <tt>new_array[n]</tt> is of size <tt>other_arrays.size+1</tt>, + * and contains: + * + * - The _nth_ element of +self+. + * - The _nth_ element of each of the +other_arrays+. + * + * If all +other_arrays+ and +self+ are the same size: + * + * a = [:a0, :a1, :a2, :a3] + * b = [:b0, :b1, :b2, :b3] + * c = [:c0, :c1, :c2, :c3] + * d = a.zip(b, c) + * d # => [[:a0, :b0, :c0], [:a1, :b1, :c1], [:a2, :b2, :c2], [:a3, :b3, :c3]] + * + * If any array in +other_arrays+ is smaller than +self+, + * fills to <tt>self.size</tt> with +nil+: + * + * a = [:a0, :a1, :a2, :a3] + * b = [:b0, :b1, :b2] + * c = [:c0, :c1] + * d = a.zip(b, c) + * d # => [[:a0, :b0, :c0], [:a1, :b1, :c1], [:a2, :b2, nil], [:a3, nil, nil]] * - * Converts any arguments to arrays, then merges elements of +self+ with - * corresponding elements from each argument. + * If any array in +other_arrays+ is larger than +self+, + * its trailing elements are ignored: * - * This generates a sequence of <code>ary.size</code> _n_-element arrays, - * where _n_ is one more than the count of arguments. + * a = [:a0, :a1, :a2, :a3] + * b = [:b0, :b1, :b2, :b3, :b4] + * c = [:c0, :c1, :c2, :c3, :c4, :c5] + * d = a.zip(b, c) + * d # => [[:a0, :b0, :c0], [:a1, :b1, :c1], [:a2, :b2, :c2], [:a3, :b3, :c3]] * - * If the size of any argument is less than the size of the initial array, - * +nil+ values are supplied. + * When a block is given, calls the block with each of the sub-arrays (formed as above); returns +nil+: * - * If a block is given, it is invoked for each output +array+, otherwise an - * array of arrays is returned. + * a = [:a0, :a1, :a2, :a3] + * b = [:b0, :b1, :b2, :b3] + * c = [:c0, :c1, :c2, :c3] + * a.zip(b, c) {|sub_array| p sub_array} # => nil + * + * Output: + * + * [:a0, :b0, :c0] + * [:a1, :b1, :c1] + * [:a2, :b2, :c2] + * [:a3, :b3, :c3] * - * a = [ 4, 5, 6 ] - * b = [ 7, 8, 9 ] - * [1, 2, 3].zip(a, b) #=> [[1, 4, 7], [2, 5, 8], [3, 6, 9]] - * [1, 2].zip(a, b) #=> [[1, 4, 7], [2, 5, 8]] - * a.zip([1, 2], [8]) #=> [[4, 1, 8], [5, 2, nil], [6, nil, nil]] */ static VALUE @@ -3702,51 +4630,51 @@ rb_ary_zip(int argc, VALUE *argv, VALUE ary) VALUE result = Qnil; for (i=0; i<argc; i++) { - argv[i] = take_items(argv[i], len); + argv[i] = take_items(argv[i], len); } if (rb_block_given_p()) { - int arity = rb_block_arity(); - - if (arity > 1) { - VALUE work, *tmp; - - tmp = ALLOCV_N(VALUE, work, argc+1); - - for (i=0; i<RARRAY_LEN(ary); i++) { - tmp[0] = RARRAY_AREF(ary, i); - for (j=0; j<argc; j++) { - tmp[j+1] = rb_ary_elt(argv[j], i); - } - rb_yield_values2(argc+1, tmp); - } - - if (work) ALLOCV_END(work); - } - else { - for (i=0; i<RARRAY_LEN(ary); i++) { - VALUE tmp = rb_ary_new2(argc+1); - - rb_ary_push(tmp, RARRAY_AREF(ary, i)); - for (j=0; j<argc; j++) { - rb_ary_push(tmp, rb_ary_elt(argv[j], i)); - } - rb_yield(tmp); - } - } + int arity = rb_block_arity(); + + if (arity > 1) { + VALUE work, *tmp; + + tmp = ALLOCV_N(VALUE, work, argc+1); + + for (i=0; i<RARRAY_LEN(ary); i++) { + tmp[0] = RARRAY_AREF(ary, i); + for (j=0; j<argc; j++) { + tmp[j+1] = rb_ary_elt(argv[j], i); + } + rb_yield_values2(argc+1, tmp); + } + + if (work) ALLOCV_END(work); + } + else { + for (i=0; i<RARRAY_LEN(ary); i++) { + VALUE tmp = rb_ary_new2(argc+1); + + rb_ary_push(tmp, RARRAY_AREF(ary, i)); + for (j=0; j<argc; j++) { + rb_ary_push(tmp, rb_ary_elt(argv[j], i)); + } + rb_yield(tmp); + } + } } else { - result = rb_ary_new_capa(len); + result = rb_ary_new_capa(len); - for (i=0; i<len; i++) { - VALUE tmp = rb_ary_new_capa(argc+1); + for (i=0; i<len; i++) { + VALUE tmp = rb_ary_new_capa(argc+1); - rb_ary_push(tmp, RARRAY_AREF(ary, i)); - for (j=0; j<argc; j++) { - rb_ary_push(tmp, rb_ary_elt(argv[j], i)); - } - rb_ary_push(result, tmp); - } + rb_ary_push(tmp, RARRAY_AREF(ary, i)); + for (j=0; j<argc; j++) { + rb_ary_push(tmp, rb_ary_elt(argv[j], i)); + } + rb_ary_push(result, tmp); + } } return result; @@ -3754,15 +4682,14 @@ rb_ary_zip(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.transpose -> new_ary + * array.transpose -> new_array * - * Assumes that +self+ is an array of arrays and transposes the rows and - * columns. + * Transposes the rows and columns in an \Array of Arrays; + * the nested Arrays must all be the same size: * - * a = [[1,2], [3,4], [5,6]] - * a.transpose #=> [[1, 3, 5], [2, 4, 6]] + * a = [[:a0, :a1], [:b0, :b1], [:c0, :c1]] + * a.transpose # => [[:a0, :b0, :c0], [:a1, :b1, :c1]] * - * If the length of the subarrays don't match, an IndexError is raised. */ static VALUE @@ -3774,36 +4701,34 @@ rb_ary_transpose(VALUE ary) alen = RARRAY_LEN(ary); if (alen == 0) return rb_ary_dup(ary); for (i=0; i<alen; i++) { - tmp = to_ary(rb_ary_elt(ary, i)); - if (elen < 0) { /* first element */ - elen = RARRAY_LEN(tmp); - result = rb_ary_new2(elen); - for (j=0; j<elen; j++) { - rb_ary_store(result, j, rb_ary_new2(alen)); - } - } - else if (elen != RARRAY_LEN(tmp)) { - rb_raise(rb_eIndexError, "element size differs (%ld should be %ld)", - RARRAY_LEN(tmp), elen); - } - for (j=0; j<elen; j++) { - rb_ary_store(rb_ary_elt(result, j), i, rb_ary_elt(tmp, j)); - } + tmp = to_ary(rb_ary_elt(ary, i)); + if (elen < 0) { /* first element */ + elen = RARRAY_LEN(tmp); + result = rb_ary_new2(elen); + for (j=0; j<elen; j++) { + rb_ary_store(result, j, rb_ary_new2(alen)); + } + } + else if (elen != RARRAY_LEN(tmp)) { + rb_raise(rb_eIndexError, "element size differs (%ld should be %ld)", + RARRAY_LEN(tmp), elen); + } + for (j=0; j<elen; j++) { + rb_ary_store(rb_ary_elt(result, j), i, rb_ary_elt(tmp, j)); + } } return result; } /* * call-seq: - * ary.replace(other_ary) -> ary - * ary.initialize_copy(other_ary) -> ary + * array.replace(other_array) -> self * - * Replaces the contents of +self+ with the contents of +other_ary+, - * truncating or expanding if necessary. + * Replaces the content of +self+ with the content of +other_array+; returns +self+: + * + * a = [:foo, 'bar', 2] + * a.replace(['foo', :bar, 3]) # => ["foo", :bar, 3] * - * a = [ "a", "b", "c", "d", "e" ] - * a.replace([ "x", "y", "z" ]) #=> ["x", "y", "z"] - * a #=> ["x", "y", "z"] */ VALUE @@ -3813,31 +4738,35 @@ rb_ary_replace(VALUE copy, VALUE orig) orig = to_ary(orig); if (copy == orig) return copy; - if (RARRAY_LEN(orig) <= RARRAY_EMBED_LEN_MAX) { - VALUE shared_root = 0; + rb_ary_reset(copy); - if (ARY_OWNS_HEAP_P(copy)) { - ary_heap_free(copy); - } - else if (ARY_SHARED_P(copy)) { - shared_root = ARY_SHARED_ROOT(copy); - FL_UNSET_SHARED(copy); - } - FL_SET_EMBED(copy); + /* orig has enough space to embed the contents of orig. */ + if (RARRAY_LEN(orig) <= ary_embed_capa(copy)) { + assert(ARY_EMBED_P(copy)); ary_memcpy(copy, 0, RARRAY_LEN(orig), RARRAY_CONST_PTR_TRANSIENT(orig)); - if (shared_root) { - rb_ary_decrement_share(shared_root); - } - ARY_SET_LEN(copy, RARRAY_LEN(orig)); + ARY_SET_EMBED_LEN(copy, RARRAY_LEN(orig)); } +#if USE_RVARGC + /* orig is embedded but copy does not have enough space to embed the + * contents of orig. */ + else if (ARY_EMBED_P(orig)) { + long len = ARY_EMBED_LEN(orig); + VALUE *ptr = ary_heap_alloc(copy, len); + + FL_UNSET_EMBED(copy); + ARY_SET_PTR(copy, ptr); + ARY_SET_LEN(copy, len); + ARY_SET_CAPA(copy, len); + + // No allocation and exception expected that could leave `copy` in a + // bad state from the edits above. + ary_memcpy(copy, 0, len, RARRAY_CONST_PTR_TRANSIENT(orig)); + } +#endif + /* Otherwise, orig is on heap and copy does not have enough space to embed + * the contents of orig. */ else { VALUE shared_root = ary_make_shared(orig); - if (ARY_OWNS_HEAP_P(copy)) { - ary_heap_free(copy); - } - else { - rb_ary_unshare_safe(copy); - } FL_UNSET_EMBED(copy); ARY_SET_PTR(copy, ARY_HEAP_PTR(orig)); ARY_SET_LEN(copy, ARY_HEAP_LEN(orig)); @@ -3849,12 +4778,13 @@ rb_ary_replace(VALUE copy, VALUE orig) /* * call-seq: - * ary.clear -> ary + * array.clear -> self + * + * Removes all elements from +self+: * - * Removes all elements from +self+. + * a = [:foo, 'bar', 2] + * a.clear # => [] * - * a = [ "a", "b", "c", "d", "e" ] - * a.clear #=> [ ] */ VALUE @@ -3862,11 +4792,11 @@ rb_ary_clear(VALUE ary) { rb_ary_modify_check(ary); if (ARY_SHARED_P(ary)) { - if (!ARY_EMBED_P(ary)) { - rb_ary_unshare(ary); - FL_SET_EMBED(ary); + if (!ARY_EMBED_P(ary)) { + rb_ary_unshare(ary); + FL_SET_EMBED(ary); ARY_SET_EMBED_LEN(ary, 0); - } + } } else { ARY_SET_LEN(ary, 0); @@ -3880,32 +4810,198 @@ rb_ary_clear(VALUE ary) /* * call-seq: - * ary.fill(obj) -> ary - * ary.fill(obj, start [, length]) -> ary - * ary.fill(obj, range) -> ary - * ary.fill {|index| block} -> ary - * ary.fill(start [, length]) {|index| block} -> ary - * ary.fill(range) {|index| block} -> ary + * array.fill(obj) -> self + * array.fill(obj, start) -> self + * array.fill(obj, start, length) -> self + * array.fill(obj, range) -> self + * array.fill {|index| ... } -> self + * array.fill(start) {|index| ... } -> self + * array.fill(start, length) {|index| ... } -> self + * array.fill(range) {|index| ... } -> self + * + * Replaces specified elements in +self+ with specified objects; returns +self+. + * + * With argument +obj+ and no block given, replaces all elements with that one object: + * + * a = ['a', 'b', 'c', 'd'] + * a # => ["a", "b", "c", "d"] + * a.fill(:X) # => [:X, :X, :X, :X] + * + * With arguments +obj+ and \Integer +start+, and no block given, + * replaces elements based on the given start. + * + * If +start+ is in range (<tt>0 <= start < array.size</tt>), + * replaces all elements from offset +start+ through the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 2) # => ["a", "b", :X, :X] + * + * If +start+ is too large (<tt>start >= array.size</tt>), does nothing: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 4) # => ["a", "b", "c", "d"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 5) # => ["a", "b", "c", "d"] + * + * If +start+ is negative, counts from the end (starting index is <tt>start + array.size</tt>): + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, -2) # => ["a", "b", :X, :X] + * + * If +start+ is too small (less than and far from zero), replaces all elements: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, -6) # => [:X, :X, :X, :X] + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, -50) # => [:X, :X, :X, :X] + * + * With arguments +obj+, \Integer +start+, and \Integer +length+, and no block given, + * replaces elements based on the given +start+ and +length+. + * + * If +start+ is in range, replaces +length+ elements beginning at offset +start+: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 1, 1) # => ["a", :X, "c", "d"] + * + * If +start+ is negative, counts from the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, -2, 1) # => ["a", "b", :X, "d"] + * + * If +start+ is large (<tt>start >= array.size</tt>), extends +self+ with +nil+: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 5, 0) # => ["a", "b", "c", "d", nil] + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 5, 2) # => ["a", "b", "c", "d", nil, :X, :X] + * + * If +length+ is zero or negative, replaces no elements: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, 1, 0) # => ["a", "b", "c", "d"] + * a.fill(:X, 1, -1) # => ["a", "b", "c", "d"] + * + * With arguments +obj+ and \Range +range+, and no block given, + * replaces elements based on the given range. + * + * If the range is positive and ascending (<tt>0 < range.begin <= range.end</tt>), + * replaces elements from <tt>range.begin</tt> to <tt>range.end</tt>: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (1..1)) # => ["a", :X, "c", "d"] + * + * If <tt>range.first</tt> is negative, replaces no elements: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (-1..1)) # => ["a", "b", "c", "d"] * - * The first three forms set the selected elements of +self+ (which - * may be the entire array) to +obj+. + * If <tt>range.last</tt> is negative, counts from the end: * - * A +start+ of +nil+ is equivalent to zero. + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (0..-2)) # => [:X, :X, :X, "d"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (1..-2)) # => ["a", :X, :X, "d"] * - * A +length+ of +nil+ is equivalent to the length of the array. + * If <tt>range.last</tt> and <tt>range.last</tt> are both negative, + * both count from the end of the array: * - * The last three forms fill the array with the value of the given block, - * which is passed the absolute index of each element to be filled. + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (-1..-1)) # => ["a", "b", "c", :X] + * a = ['a', 'b', 'c', 'd'] + * a.fill(:X, (-2..-2)) # => ["a", "b", :X, "d"] * - * Negative values of +start+ count from the end of the array, where +-1+ is - * the last element. + * With no arguments and a block given, calls the block with each index; + * replaces the corresponding element with the block's return value: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill { |index| "new_#{index}" } # => ["new_0", "new_1", "new_2", "new_3"] + * + * With argument +start+ and a block given, calls the block with each index + * from offset +start+ to the end; replaces the corresponding element + * with the block's return value. + * + * If start is in range (<tt>0 <= start < array.size</tt>), + * replaces from offset +start+ to the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(1) { |index| "new_#{index}" } # => ["a", "new_1", "new_2", "new_3"] + * + * If +start+ is too large(<tt>start >= array.size</tt>), does nothing: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(4) { |index| fail 'Cannot happen' } # => ["a", "b", "c", "d"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(4) { |index| fail 'Cannot happen' } # => ["a", "b", "c", "d"] + * + * If +start+ is negative, counts from the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(-2) { |index| "new_#{index}" } # => ["a", "b", "new_2", "new_3"] + * + * If start is too small (<tt>start <= -array.size</tt>, replaces all elements: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(-6) { |index| "new_#{index}" } # => ["new_0", "new_1", "new_2", "new_3"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(-50) { |index| "new_#{index}" } # => ["new_0", "new_1", "new_2", "new_3"] + * + * With arguments +start+ and +length+, and a block given, + * calls the block for each index specified by start length; + * replaces the corresponding element with the block's return value. + * + * If +start+ is in range, replaces +length+ elements beginning at offset +start+: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(1, 1) { |index| "new_#{index}" } # => ["a", "new_1", "c", "d"] + * + * If start is negative, counts from the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(-2, 1) { |index| "new_#{index}" } # => ["a", "b", "new_2", "d"] + * + * If +start+ is large (<tt>start >= array.size</tt>), extends +self+ with +nil+: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(5, 0) { |index| "new_#{index}" } # => ["a", "b", "c", "d", nil] + * a = ['a', 'b', 'c', 'd'] + * a.fill(5, 2) { |index| "new_#{index}" } # => ["a", "b", "c", "d", nil, "new_5", "new_6"] + * + * If +length+ is zero or less, replaces no elements: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(1, 0) { |index| "new_#{index}" } # => ["a", "b", "c", "d"] + * a.fill(1, -1) { |index| "new_#{index}" } # => ["a", "b", "c", "d"] + * + * With arguments +obj+ and +range+, and a block given, + * calls the block with each index in the given range; + * replaces the corresponding element with the block's return value. + * + * If the range is positive and ascending (<tt>range 0 < range.begin <= range.end</tt>, + * replaces elements from <tt>range.begin</tt> to <tt>range.end</tt>: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(1..1) { |index| "new_#{index}" } # => ["a", "new_1", "c", "d"] + * + * If +range.first+ is negative, does nothing: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(-1..1) { |index| fail 'Cannot happen' } # => ["a", "b", "c", "d"] + * + * If <tt>range.last</tt> is negative, counts from the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(0..-2) { |index| "new_#{index}" } # => ["new_0", "new_1", "new_2", "d"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(1..-2) { |index| "new_#{index}" } # => ["a", "new_1", "new_2", "d"] + * + * If <tt>range.first</tt> and <tt>range.last</tt> are both negative, + * both count from the end: + * + * a = ['a', 'b', 'c', 'd'] + * a.fill(-1..-1) { |index| "new_#{index}" } # => ["a", "b", "c", "new_3"] + * a = ['a', 'b', 'c', 'd'] + * a.fill(-2..-2) { |index| "new_#{index}" } # => ["a", "b", "new_2", "d"] * - * a = [ "a", "b", "c", "d" ] - * a.fill("x") #=> ["x", "x", "x", "x"] - * a.fill("z", 2, 2) #=> ["x", "x", "z", "z"] - * a.fill("y", 0..1) #=> ["y", "y", "z", "z"] - * a.fill {|i| i*i} #=> [0, 1, 4, 9] - * a.fill(-2) {|i| i*i*i} #=> [0, 1, 8, 27] */ static VALUE @@ -3915,84 +5011,74 @@ rb_ary_fill(int argc, VALUE *argv, VALUE ary) long beg = 0, end = 0, len = 0; if (rb_block_given_p()) { - rb_scan_args(argc, argv, "02", &arg1, &arg2); - argc += 1; /* hackish */ + rb_scan_args(argc, argv, "02", &arg1, &arg2); + argc += 1; /* hackish */ } else { - rb_scan_args(argc, argv, "12", &item, &arg1, &arg2); + rb_scan_args(argc, argv, "12", &item, &arg1, &arg2); } switch (argc) { case 1: - beg = 0; - len = RARRAY_LEN(ary); - break; + beg = 0; + len = RARRAY_LEN(ary); + break; case 2: - if (rb_range_beg_len(arg1, &beg, &len, RARRAY_LEN(ary), 1)) { - break; - } - /* fall through */ + if (rb_range_beg_len(arg1, &beg, &len, RARRAY_LEN(ary), 1)) { + break; + } + /* fall through */ case 3: - beg = NIL_P(arg1) ? 0 : NUM2LONG(arg1); - if (beg < 0) { - beg = RARRAY_LEN(ary) + beg; - if (beg < 0) beg = 0; - } - len = NIL_P(arg2) ? RARRAY_LEN(ary) - beg : NUM2LONG(arg2); - break; + beg = NIL_P(arg1) ? 0 : NUM2LONG(arg1); + if (beg < 0) { + beg = RARRAY_LEN(ary) + beg; + if (beg < 0) beg = 0; + } + len = NIL_P(arg2) ? RARRAY_LEN(ary) - beg : NUM2LONG(arg2); + break; } rb_ary_modify(ary); if (len < 0) { return ary; } if (beg >= ARY_MAX_SIZE || len > ARY_MAX_SIZE - beg) { - rb_raise(rb_eArgError, "argument too big"); + rb_raise(rb_eArgError, "argument too big"); } end = beg + len; if (RARRAY_LEN(ary) < end) { - if (end >= ARY_CAPA(ary)) { - ary_resize_capa(ary, end); - } - ary_mem_clear(ary, RARRAY_LEN(ary), end - RARRAY_LEN(ary)); - ARY_SET_LEN(ary, end); + if (end >= ARY_CAPA(ary)) { + ary_resize_capa(ary, end); + } + ary_mem_clear(ary, RARRAY_LEN(ary), end - RARRAY_LEN(ary)); + ARY_SET_LEN(ary, end); } - if (item == Qundef) { - VALUE v; - long i; + if (UNDEF_P(item)) { + VALUE v; + long i; - for (i=beg; i<end; i++) { - v = rb_yield(LONG2NUM(i)); - if (i>=RARRAY_LEN(ary)) break; - ARY_SET(ary, i, v); - } + for (i=beg; i<end; i++) { + v = rb_yield(LONG2NUM(i)); + if (i>=RARRAY_LEN(ary)) break; + ARY_SET(ary, i, v); + } } else { - ary_memfill(ary, beg, len, item); + ary_memfill(ary, beg, len, item); } return ary; } /* * call-seq: - * ary + other_ary -> new_ary - * - * Concatenation --- Returns a new array built by concatenating the - * two arrays together to produce a third array. + * array + other_array -> new_array * - * [ 1, 2, 3 ] + [ 4, 5 ] #=> [ 1, 2, 3, 4, 5 ] - * a = [ "a", "b", "c" ] - * c = a + [ "d", "e", "f" ] - * c #=> [ "a", "b", "c", "d", "e", "f" ] - * a #=> [ "a", "b", "c" ] + * Returns a new \Array containing all elements of +array+ + * followed by all elements of +other_array+: * - * Note that - * x += y - * is the same as - * x = x + y - * This means that it produces a new array. As a consequence, - * repeated use of <code>+=</code> on arrays can be quite inefficient. + * a = [0, 1] + [2, 3] + * a # => [0, 1, 2, 3] * - * See also Array#concat. + * Related: #concat. */ VALUE @@ -4020,27 +5106,18 @@ ary_append(VALUE x, VALUE y) if (n > 0) { rb_ary_splice(x, RARRAY_LEN(x), 0, RARRAY_CONST_PTR_TRANSIENT(y), n); } + RB_GC_GUARD(y); return x; } /* * call-seq: - * ary.concat(other_ary1, other_ary2, ...) -> ary - * - * Appends the elements of <code>other_ary</code>s to +self+. - * - * [ "a", "b" ].concat( ["c", "d"]) #=> [ "a", "b", "c", "d" ] - * [ "a" ].concat( ["b"], ["c", "d"]) #=> [ "a", "b", "c", "d" ] - * [ "a" ].concat #=> [ "a" ] - * - * a = [ 1, 2, 3 ] - * a.concat( [ 4, 5 ]) - * a #=> [ 1, 2, 3, 4, 5 ] + * array.concat(*other_arrays) -> self * - * a = [ 1, 2 ] - * a.concat(a, a) #=> [1, 2, 1, 2, 1, 2] + * Adds to +array+ all elements from each \Array in +other_arrays+; returns +self+: * - * See also Array#+. + * a = [0, 1] + * a.concat([2, 3], [4, 5]) # => [0, 1, 2, 3, 4, 5] */ static VALUE @@ -4049,15 +5126,15 @@ rb_ary_concat_multi(int argc, VALUE *argv, VALUE ary) rb_ary_modify_check(ary); if (argc == 1) { - rb_ary_concat(ary, argv[0]); + rb_ary_concat(ary, argv[0]); } else if (argc > 1) { - int i; - VALUE args = rb_ary_tmp_new(argc); - for (i = 0; i < argc; i++) { - rb_ary_concat(args, argv[i]); - } - ary_append(ary, args); + int i; + VALUE args = rb_ary_hidden_new(argc); + for (i = 0; i < argc; i++) { + rb_ary_concat(args, argv[i]); + } + ary_append(ary, args); } ary_verify(ary); @@ -4072,18 +5149,19 @@ rb_ary_concat(VALUE x, VALUE y) /* * call-seq: - * ary * int -> new_ary - * ary * str -> new_string + * array * n -> new_array + * array * string_separator -> new_string * - * Repetition --- With a String argument, equivalent to - * <code>ary.join(str)</code>. + * When non-negative argument \Integer +n+ is given, + * returns a new \Array built by concatenating the +n+ copies of +self+: * - * Otherwise, returns a new array built by concatenating the +int+ copies of - * +self+. + * a = ['x', 'y'] + * a * 3 # => ["x", "y", "x", "y", "x", "y"] * + * When \String argument +string_separator+ is given, + * equivalent to <tt>array.join(string_separator)</tt>: * - * [ 1, 2, 3 ] * 3 #=> [ 1, 2, 3, 1, 2, 3, 1, 2, 3 ] - * [ 1, 2, 3 ] * "," #=> "1,2,3" + * [0, [0, 1], {foo: 0}] * ', ' # => "0, 0, 1, {:foo=>0}" * */ @@ -4096,30 +5174,30 @@ rb_ary_times(VALUE ary, VALUE times) tmp = rb_check_string_type(times); if (!NIL_P(tmp)) { - return rb_ary_join(ary, tmp); + return rb_ary_join(ary, tmp); } len = NUM2LONG(times); if (len == 0) { - ary2 = ary_new(rb_obj_class(ary), 0); - goto out; + ary2 = ary_new(rb_cArray, 0); + goto out; } if (len < 0) { - rb_raise(rb_eArgError, "negative argument"); + rb_raise(rb_eArgError, "negative argument"); } if (ARY_MAX_SIZE/len < RARRAY_LEN(ary)) { - rb_raise(rb_eArgError, "argument too big"); + rb_raise(rb_eArgError, "argument too big"); } len *= RARRAY_LEN(ary); - ary2 = ary_new(rb_obj_class(ary), len); + ary2 = ary_new(rb_cArray, len); ARY_SET_LEN(ary2, len); ptr = RARRAY_CONST_PTR_TRANSIENT(ary); t = RARRAY_LEN(ary); if (0 < t) { - ary_memcpy(ary2, 0, t, ptr); - while (t <= len/2) { + ary_memcpy(ary2, 0, t, ptr); + while (t <= len/2) { ary_memcpy(ary2, t, t, RARRAY_CONST_PTR_TRANSIENT(ary2)); t *= 2; } @@ -4128,29 +5206,22 @@ rb_ary_times(VALUE ary, VALUE times) } } out: - OBJ_INFECT(ary2, ary); - return ary2; } /* * call-seq: - * ary.assoc(obj) -> element_ary or nil + * array.assoc(obj) -> found_array or nil * - * Searches through an array whose elements are also arrays comparing +obj+ - * with the first element of each contained array using <code>obj.==</code>. + * Returns the first element in +self+ that is an \Array + * whose first element <tt>==</tt> +obj+: * - * Returns the first contained array that matches (that is, the first - * associated array), or +nil+ if no match is found. + * a = [{foo: 0}, [2, 4], [4, 5, 6], [4, 5]] + * a.assoc(4) # => [4, 5, 6] * - * See also Array#rassoc + * Returns +nil+ if no such element is found. * - * s1 = [ "colors", "red", "blue", "green" ] - * s2 = [ "letters", "a", "b", "c" ] - * s3 = "foo" - * a = [ s1, s2, s3 ] - * a.assoc("letters") #=> [ "letters", "a", "b", "c" ] - * a.assoc("foo") #=> nil + * Related: #rassoc. */ VALUE @@ -4160,30 +5231,27 @@ rb_ary_assoc(VALUE ary, VALUE key) VALUE v; for (i = 0; i < RARRAY_LEN(ary); ++i) { - v = rb_check_array_type(RARRAY_AREF(ary, i)); - if (!NIL_P(v) && RARRAY_LEN(v) > 0 && - rb_equal(RARRAY_AREF(v, 0), key)) - return v; + v = rb_check_array_type(RARRAY_AREF(ary, i)); + if (!NIL_P(v) && RARRAY_LEN(v) > 0 && + rb_equal(RARRAY_AREF(v, 0), key)) + return v; } return Qnil; } /* * call-seq: - * ary.rassoc(obj) -> element_ary or nil - * - * Searches through the array whose elements are also arrays. + * array.rassoc(obj) -> found_array or nil * - * Compares +obj+ with the second element of each contained array using - * <code>obj.==</code>. + * Returns the first element in +self+ that is an \Array + * whose second element <tt>==</tt> +obj+: * - * Returns the first contained array that matches +obj+. + * a = [{foo: 0}, [2, 4], [4, 5, 6], [4, 5]] + * a.rassoc(4) # => [2, 4] * - * See also Array#assoc. + * Returns +nil+ if no such element is found. * - * a = [ [ 1, "one"], [2, "two"], [3, "three"], ["ii", "two"] ] - * a.rassoc("two") #=> [2, "two"] - * a.rassoc("four") #=> nil + * Related: #assoc. */ VALUE @@ -4193,11 +5261,11 @@ rb_ary_rassoc(VALUE ary, VALUE value) VALUE v; for (i = 0; i < RARRAY_LEN(ary); ++i) { - v = RARRAY_AREF(ary, i); - if (RB_TYPE_P(v, T_ARRAY) && - RARRAY_LEN(v) > 1 && - rb_equal(RARRAY_AREF(v, 1), value)) - return v; + v = RARRAY_AREF(ary, i); + if (RB_TYPE_P(v, T_ARRAY) && + RARRAY_LEN(v) > 1 && + rb_equal(RARRAY_AREF(v, 1), value)) + return v; } return Qnil; } @@ -4216,38 +5284,42 @@ recursive_equal(VALUE ary1, VALUE ary2, int recur) len1 = RARRAY_LEN(ary1); for (i = 0; i < len1; i++) { - if (*p1 != *p2) { - if (rb_equal(*p1, *p2)) { - len1 = RARRAY_LEN(ary1); - if (len1 != RARRAY_LEN(ary2)) - return Qfalse; - if (len1 < i) - return Qtrue; + if (*p1 != *p2) { + if (rb_equal(*p1, *p2)) { + len1 = RARRAY_LEN(ary1); + if (len1 != RARRAY_LEN(ary2)) + return Qfalse; + if (len1 < i) + return Qtrue; p1 = RARRAY_CONST_PTR(ary1) + i; p2 = RARRAY_CONST_PTR(ary2) + i; - } - else { - return Qfalse; - } - } - p1++; - p2++; + } + else { + return Qfalse; + } + } + p1++; + p2++; } return Qtrue; } /* * call-seq: - * ary == other_ary -> bool + * array == other_array -> true or false + * + * Returns +true+ if both <tt>array.size == other_array.size</tt> + * and for each index +i+ in +array+, <tt>array[i] == other_array[i]</tt>: * - * Equality --- Two arrays are equal if they contain the same number of - * elements and if each element is equal to (according to Object#==) the - * corresponding element in +other_ary+. + * a0 = [:foo, 'bar', 2] + * a1 = [:foo, 'bar', 2.0] + * a1 == a0 # => true + * [] == [] # => true * - * [ "a", "c" ] == [ "a", "c", 7 ] #=> false - * [ "a", "c", 7 ] == [ "a", "c", 7 ] #=> true - * [ "a", "c", 7 ] == [ "a", "d", "f" ] #=> false + * Otherwise, returns +false+. * + * This method is different from method Array#eql?, + * which compares elements using <tt>Object#eql?</tt>. */ static VALUE @@ -4255,10 +5327,10 @@ rb_ary_equal(VALUE ary1, VALUE ary2) { if (ary1 == ary2) return Qtrue; if (!RB_TYPE_P(ary2, T_ARRAY)) { - if (!rb_respond_to(ary2, idTo_ary)) { - return Qfalse; - } - return rb_equal(ary2, ary1); + if (!rb_respond_to(ary2, idTo_ary)) { + return Qfalse; + } + return rb_equal(ary2, ary1); } if (RARRAY_LEN(ary1) != RARRAY_LEN(ary2)) return Qfalse; if (RARRAY_CONST_PTR_TRANSIENT(ary1) == RARRAY_CONST_PTR_TRANSIENT(ary2)) return Qtrue; @@ -4272,18 +5344,27 @@ recursive_eql(VALUE ary1, VALUE ary2, int recur) if (recur) return Qtrue; /* Subtle! */ for (i=0; i<RARRAY_LEN(ary1); i++) { - if (!rb_eql(rb_ary_elt(ary1, i), rb_ary_elt(ary2, i))) - return Qfalse; + if (!rb_eql(rb_ary_elt(ary1, i), rb_ary_elt(ary2, i))) + return Qfalse; } return Qtrue; } /* * call-seq: - * ary.eql?(other) -> true or false + * array.eql? other_array -> true or false * - * Returns +true+ if +self+ and +other+ are the same object, - * or are both arrays with the same content (according to Object#eql?). + * Returns +true+ if +self+ and +other_array+ are the same size, + * and if, for each index +i+ in +self+, <tt>self[i].eql? other_array[i]</tt>: + * + * a0 = [:foo, 'bar', 2] + * a1 = [:foo, 'bar', 2] + * a1.eql?(a0) # => true + * + * Otherwise, returns +false+. + * + * This method is different from method Array#==, + * which compares using method <tt>Object#==</tt>. */ static VALUE @@ -4298,14 +5379,15 @@ rb_ary_eql(VALUE ary1, VALUE ary2) /* * call-seq: - * ary.hash -> integer + * array.hash -> integer * - * Compute a hash-code for this array. + * Returns the integer hash value for +self+. * - * Two arrays with the same content will have the same hash code (and will - * compare using #eql?). + * Two arrays with the same content will have the same hash code (and will compare using eql?): + * + * [0, 1, 2].hash == [0, 1, 2].hash # => true + * [0, 1, 2].hash == [0, 1, 3].hash # => false * - * See also Object#hash. */ static VALUE @@ -4318,8 +5400,8 @@ rb_ary_hash(VALUE ary) h = rb_hash_start(RARRAY_LEN(ary)); h = rb_hash_uint(h, (st_index_t)rb_ary_hash); for (i=0; i<RARRAY_LEN(ary); i++) { - n = rb_hash(RARRAY_AREF(ary, i)); - h = rb_hash_uint(h, NUM2LONG(n)); + n = rb_hash(RARRAY_AREF(ary, i)); + h = rb_hash_uint(h, NUM2LONG(n)); } h = rb_hash_end(h); return ST2FIX(h); @@ -4327,14 +5409,13 @@ rb_ary_hash(VALUE ary) /* * call-seq: - * ary.include?(object) -> true or false + * array.include?(obj) -> true or false * - * Returns +true+ if the given +object+ is present in +self+ (that is, if any - * element <code>==</code> +object+), otherwise returns +false+. + * Returns +true+ if for some index +i+ in +self+, <tt>obj == self[i]</tt>; + * otherwise +false+: * - * a = [ "a", "b", "c" ] - * a.include?("b") #=> true - * a.include?("z") #=> false + * [0, 1, 2].include?(2) # => true + * [0, 1, 2].include?(3) # => false */ VALUE @@ -4344,10 +5425,10 @@ rb_ary_includes(VALUE ary, VALUE item) VALUE e; for (i=0; i<RARRAY_LEN(ary); i++) { - e = RARRAY_AREF(ary, i); - if (rb_equal(e, item)) { - return Qtrue; - } + e = RARRAY_AREF(ary, i); + if (rb_equal(e, item)) { + return Qtrue; + } } return Qfalse; } @@ -4359,10 +5440,10 @@ rb_ary_includes_by_eql(VALUE ary, VALUE item) VALUE e; for (i=0; i<RARRAY_LEN(ary); i++) { - e = RARRAY_AREF(ary, i); - if (rb_eql(item, e)) { - return Qtrue; - } + e = RARRAY_AREF(ary, i); + if (rb_eql(item, e)) { + return Qtrue; + } } return Qfalse; } @@ -4375,45 +5456,46 @@ recursive_cmp(VALUE ary1, VALUE ary2, int recur) if (recur) return Qundef; /* Subtle! */ len = RARRAY_LEN(ary1); if (len > RARRAY_LEN(ary2)) { - len = RARRAY_LEN(ary2); + len = RARRAY_LEN(ary2); } for (i=0; i<len; i++) { - VALUE e1 = rb_ary_elt(ary1, i), e2 = rb_ary_elt(ary2, i); - VALUE v = rb_funcallv(e1, id_cmp, 1, &e2); - if (v != INT2FIX(0)) { - return v; - } + VALUE e1 = rb_ary_elt(ary1, i), e2 = rb_ary_elt(ary2, i); + VALUE v = rb_funcallv(e1, id_cmp, 1, &e2); + if (v != INT2FIX(0)) { + return v; + } } return Qundef; } /* * call-seq: - * ary <=> other_ary -> -1, 0, +1 or nil + * array <=> other_array -> -1, 0, or 1 + * + * Returns -1, 0, or 1 as +self+ is less than, equal to, or greater than +other_array+. + * For each index +i+ in +self+, evaluates <tt>result = self[i] <=> other_array[i]</tt>. + * + * Returns -1 if any result is -1: * - * Comparison --- Returns an integer (+-1+, +0+, or <code>+1</code>) if this - * array is less than, equal to, or greater than +other_ary+. + * [0, 1, 2] <=> [0, 1, 3] # => -1 * - * Each object in each array is compared (using the <=> operator). + * Returns 1 if any result is 1: * - * Arrays are compared in an "element-wise" manner; the first element of +ary+ - * is compared with the first one of +other_ary+ using the <=> operator, then - * each of the second elements, etc... - * As soon as the result of any such comparison is non zero (i.e. the two - * corresponding elements are not equal), that result is returned for the - * whole array comparison. + * [0, 1, 2] <=> [0, 1, 1] # => 1 * - * If all the elements are equal, then the result is based on a comparison of - * the array lengths. Thus, two arrays are "equal" according to Array#<=> if, - * and only if, they have the same length and the value of each element is - * equal to the value of the corresponding element in the other array. + * When all results are zero: * - * +nil+ is returned if the +other_ary+ is not an array or if the comparison - * of two elements returned +nil+. + * - Returns -1 if +array+ is smaller than +other_array+: * - * [ "a", "a", "c" ] <=> [ "a", "b", "c" ] #=> -1 - * [ 1, 2, 3, 4, 5, 6 ] <=> [ 1, 2 ] #=> +1 - * [ 1, 2 ] <=> [ 1, :two ] #=> nil + * [0, 1, 2] <=> [0, 1, 2, 3] # => -1 + * + * - Returns 1 if +array+ is larger than +other_array+: + * + * [0, 1, 2] <=> [0, 1] # => 1 + * + * - Returns 0 if +array+ and +other_array+ are the same size: + * + * [0, 1, 2] <=> [0, 1, 2] # => 0 * */ @@ -4427,7 +5509,7 @@ rb_ary_cmp(VALUE ary1, VALUE ary2) if (NIL_P(ary2)) return Qnil; if (ary1 == ary2) return INT2FIX(0); v = rb_exec_recursive_paired(recursive_cmp, ary1, ary2, ary2); - if (v != Qundef) return v; + if (!UNDEF_P(v)) return v; len = RARRAY_LEN(ary1) - RARRAY_LEN(ary2); if (len == 0) return INT2FIX(0); if (len > 0) return INT2FIX(1); @@ -4440,8 +5522,8 @@ ary_add_hash(VALUE hash, VALUE ary) long i; for (i=0; i<RARRAY_LEN(ary); i++) { - VALUE elt = RARRAY_AREF(ary, i); - rb_hash_add_new_element(hash, elt, elt); + VALUE elt = RARRAY_AREF(ary, i); + rb_hash_add_new_element(hash, elt, elt); } return hash; } @@ -4469,8 +5551,8 @@ ary_add_hash_by(VALUE hash, VALUE ary) long i; for (i = 0; i < RARRAY_LEN(ary); ++i) { - VALUE v = rb_ary_elt(ary, i), k = rb_yield(v); - rb_hash_add_new_element(hash, k, v); + VALUE v = rb_ary_elt(ary, i), k = rb_yield(v); + rb_hash_add_new_element(hash, k, v); } return hash; } @@ -4488,35 +5570,28 @@ ary_recycle_hash(VALUE hash) assert(RBASIC_CLASS(hash) == 0); if (RHASH_ST_TABLE_P(hash)) { st_table *tbl = RHASH_ST_TABLE(hash); - st_free_table(tbl); + st_free_table(tbl); RHASH_ST_CLEAR(hash); } } /* * call-seq: - * ary - other_ary -> new_ary - * - * Array Difference + * array - other_array -> new_array * - * Returns a new array that is a copy of the original array, removing all - * occurences of any item that also appear in +other_ary+. The order is - * preserved from the original array. + * Returns a new \Array containing only those elements from +array+ + * that are not found in \Array +other_array+; + * items are compared using <tt>eql?</tt>; + * the order from +array+ is preserved: * - * It compares elements using their #hash and #eql? methods for efficiency. + * [0, 1, 1, 2, 1, 1, 3, 1, 1] - [1] # => [0, 2, 3] + * [0, 1, 2, 3] - [3, 0] # => [1, 2] + * [0, 1, 2] - [4] # => [0, 1, 2] * - * [ 1, 1, 2, 2, 3, 3, 4, 5 ] - [ 1, 2, 4 ] #=> [ 3, 3, 5 ] - * - * Note that while 1 and 2 were only present once in the array argument, and - * were present twice in the receiver array, all occurences of each Integer are - * removed in the returned array. - * - * If you need set-like behavior, see the library class Set. - * - * See also Array#difference. + * Related: Array#difference. */ -static VALUE +VALUE rb_ary_diff(VALUE ary1, VALUE ary2) { VALUE ary3; @@ -4524,21 +5599,22 @@ rb_ary_diff(VALUE ary1, VALUE ary2) long i; ary2 = to_ary(ary2); + if (RARRAY_LEN(ary2) == 0) { return ary_make_shared_copy(ary1); } ary3 = rb_ary_new(); if (RARRAY_LEN(ary1) <= SMALL_ARRAY_LEN || RARRAY_LEN(ary2) <= SMALL_ARRAY_LEN) { - for (i=0; i<RARRAY_LEN(ary1); i++) { - VALUE elt = rb_ary_elt(ary1, i); - if (rb_ary_includes_by_eql(ary2, elt)) continue; - rb_ary_push(ary3, elt); - } - return ary3; + for (i=0; i<RARRAY_LEN(ary1); i++) { + VALUE elt = rb_ary_elt(ary1, i); + if (rb_ary_includes_by_eql(ary2, elt)) continue; + rb_ary_push(ary3, elt); + } + return ary3; } hash = ary_make_hash(ary2); for (i=0; i<RARRAY_LEN(ary1); i++) { if (rb_hash_stlike_lookup(hash, RARRAY_AREF(ary1, i), NULL)) continue; - rb_ary_push(ary3, rb_ary_elt(ary1, i)); + rb_ary_push(ary3, rb_ary_elt(ary1, i)); } ary_recycle_hash(hash); return ary3; @@ -4546,31 +5622,19 @@ rb_ary_diff(VALUE ary1, VALUE ary2) /* * call-seq: - * ary.difference(other_ary1, other_ary2, ...) -> new_ary - * - * Array Difference - * - * Returns a new array that is a copy of the original array, removing all - * occurences of any item that also appear in +other_ary+. The order is - * preserved from the original array. - * - * It compares elements using their #hash and #eql? methods for efficiency. - * - * [ 1, 1, 2, 2, 3, 3, 4, 5 ].difference([ 1, 2, 4 ]) #=> [ 3, 3, 5 ] - * - * Note that while 1 and 2 were only present once in the array argument, and - * were present twice in the receiver array, all occurences of each Integer are - * removed in the returned array. + * array.difference(*other_arrays) -> new_array * - * Multiple array arguments can be supplied and all occurences of any element - * in those supplied arrays that match the receiver will be removed from the - * returned array. + * Returns a new \Array containing only those elements from +self+ + * that are not found in any of the Arrays +other_arrays+; + * items are compared using <tt>eql?</tt>; order from +self+ is preserved: * - * [ 1, 'c', :s, 'yep' ].difference([ 1 ], [ 'a', 'c' ]) #=> [ :s, "yep" ] + * [0, 1, 1, 2, 1, 1, 3, 1, 1].difference([1]) # => [0, 2, 3] + * [0, 1, 2, 3].difference([3, 0], [1, 3]) # => [2] + * [0, 1, 2].difference([4]) # => [0, 1, 2] * - * If you need set-like behavior, see the library class Set. + * Returns a copy of +self+ if no arguments given. * - * See also Array#-. + * Related: Array#-. */ static VALUE @@ -4612,17 +5676,19 @@ rb_ary_difference_multi(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary & other_ary -> new_ary + * array & other_array -> new_array * - * Set Intersection --- Returns a new array containing unique elements common to the - * two arrays. The order is preserved from the original array. + * Returns a new \Array containing each element found in both +array+ and \Array +other_array+; + * duplicates are omitted; items are compared using <tt>eql?</tt>: * - * It compares elements using their #hash and #eql? methods for efficiency. + * [0, 1, 2, 3] & [1, 2] # => [1, 2] + * [0, 1, 0, 1] & [0, 1] # => [0, 1] * - * [ 1, 1, 3, 5 ] & [ 3, 2, 1 ] #=> [ 1, 3 ] - * [ 'a', 'b', 'b', 'z' ] & [ 'a', 'b', 'c' ] #=> [ 'a', 'b' ] + * Preserves order from +array+: * - * See also Array#uniq. + * [0, 1, 2] & [3, 2, 1, 0] # => [0, 1, 2] + * + * Related: Array#intersection. */ @@ -4638,29 +5704,62 @@ rb_ary_and(VALUE ary1, VALUE ary2) if (RARRAY_LEN(ary1) == 0 || RARRAY_LEN(ary2) == 0) return ary3; if (RARRAY_LEN(ary1) <= SMALL_ARRAY_LEN && RARRAY_LEN(ary2) <= SMALL_ARRAY_LEN) { - for (i=0; i<RARRAY_LEN(ary1); i++) { - v = RARRAY_AREF(ary1, i); - if (!rb_ary_includes_by_eql(ary2, v)) continue; - if (rb_ary_includes_by_eql(ary3, v)) continue; - rb_ary_push(ary3, v); - } - return ary3; + for (i=0; i<RARRAY_LEN(ary1); i++) { + v = RARRAY_AREF(ary1, i); + if (!rb_ary_includes_by_eql(ary2, v)) continue; + if (rb_ary_includes_by_eql(ary3, v)) continue; + rb_ary_push(ary3, v); + } + return ary3; } hash = ary_make_hash(ary2); for (i=0; i<RARRAY_LEN(ary1); i++) { - v = RARRAY_AREF(ary1, i); - vv = (st_data_t)v; + v = RARRAY_AREF(ary1, i); + vv = (st_data_t)v; if (rb_hash_stlike_delete(hash, &vv, 0)) { - rb_ary_push(ary3, v); - } + rb_ary_push(ary3, v); + } } ary_recycle_hash(hash); return ary3; } +/* + * call-seq: + * array.intersection(*other_arrays) -> new_array + * + * Returns a new \Array containing each element found both in +self+ + * and in all of the given Arrays +other_arrays+; + * duplicates are omitted; items are compared using <tt>eql?</tt>: + * + * [0, 1, 2, 3].intersection([0, 1, 2], [0, 1, 3]) # => [0, 1] + * [0, 0, 1, 1, 2, 3].intersection([0, 1, 2], [0, 1, 3]) # => [0, 1] + * + * Preserves order from +self+: + * + * [0, 1, 2].intersection([2, 1, 0]) # => [0, 1, 2] + * + * Returns a copy of +self+ if no arguments given. + * + * Related: Array#&. + */ + +static VALUE +rb_ary_intersection_multi(int argc, VALUE *argv, VALUE ary) +{ + VALUE result = rb_ary_dup(ary); + int i; + + for (i = 0; i < argc; i++) { + result = rb_ary_and(result, argv[i]); + } + + return result; +} + static int ary_hash_orset(st_data_t *key, st_data_t *value, st_data_t arg, int existing) { @@ -4694,17 +5793,17 @@ rb_ary_union_hash(VALUE hash, VALUE ary2) /* * call-seq: - * ary | other_ary -> new_ary - * - * Set Union --- Returns a new array by joining +ary+ with +other_ary+, - * excluding any duplicates and preserving the order from the given arrays. + * array | other_array -> new_array * - * It compares elements using their #hash and #eql? methods for efficiency. + * Returns the union of +array+ and \Array +other_array+; + * duplicates are removed; order is preserved; + * items are compared using <tt>eql?</tt>: * - * [ "a", "b", "c" ] | [ "c", "d", "a" ] #=> [ "a", "b", "c", "d" ] - * [ "c", "d", "a" ] | [ "a", "b", "c" ] #=> [ "c", "d", "a", "b" ] + * [0, 1] | [2, 3] # => [0, 1, 2, 3] + * [0, 1, 1] | [2, 2, 3] # => [0, 1, 2, 3] + * [0, 1, 2] | [3, 2, 1, 0] # => [0, 1, 2, 3] * - * See also Array#union. + * Related: Array#union. */ static VALUE @@ -4714,10 +5813,10 @@ rb_ary_or(VALUE ary1, VALUE ary2) ary2 = to_ary(ary2); if (RARRAY_LEN(ary1) + RARRAY_LEN(ary2) <= SMALL_ARRAY_LEN) { - ary3 = rb_ary_new(); + ary3 = rb_ary_new(); rb_ary_union(ary3, ary1); rb_ary_union(ary3, ary2); - return ary3; + return ary3; } hash = ary_make_hash(ary1); @@ -4730,18 +5829,18 @@ rb_ary_or(VALUE ary1, VALUE ary2) /* * call-seq: - * ary.union(other_ary1, other_ary2, ...) -> new_ary + * array.union(*other_arrays) -> new_array * - * Set Union --- Returns a new array by joining <code>other_ary</code>s with +self+, - * excluding any duplicates and preserving the order from the given arrays. + * Returns a new \Array that is the union of +self+ and all given Arrays +other_arrays+; + * duplicates are removed; order is preserved; items are compared using <tt>eql?</tt>: * - * It compares elements using their #hash and #eql? methods for efficiency. + * [0, 1, 2, 3].union([4, 5], [6, 7]) # => [0, 1, 2, 3, 4, 5, 6, 7] + * [0, 1, 1].union([2, 1], [3, 1]) # => [0, 1, 2, 3] + * [0, 1, 2, 3].union([3, 2], [1, 0]) # => [0, 1, 2, 3] * - * [ "a", "b", "c" ].union( [ "c", "d", "a" ] ) #=> [ "a", "b", "c", "d" ] - * [ "a" ].union( ["e", "b"], ["a", "c", "b"] ) #=> [ "a", "e", "b", "c" ] - * [ "a" ].union #=> [ "a" ] + * Returns a copy of +self+ if no arguments given. * - * See also Array#|. + * Related: Array#|. */ static VALUE @@ -4776,30 +5875,191 @@ rb_ary_union_multi(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.max -> obj - * ary.max {|a, b| block} -> obj - * ary.max(n) -> array - * ary.max(n) {|a, b| block} -> array + * ary.intersect?(other_ary) -> true or false * - * Returns the object in _ary_ with the maximum value. The - * first form assumes all objects implement Comparable; - * the second uses the block to return <em>a <=> b</em>. + * Returns +true+ if the array and +other_ary+ have at least one element in + * common, otherwise returns +false+: * - * ary = %w(albatross dog horse) - * ary.max #=> "horse" - * ary.max {|a, b| a.length <=> b.length} #=> "albatross" + * a = [ 1, 2, 3 ] + * b = [ 3, 4, 5 ] + * c = [ 5, 6, 7 ] + * a.intersect?(b) #=> true + * a.intersect?(c) #=> false * - * If the +n+ argument is given, maximum +n+ elements are returned - * as an array. + */ + +static VALUE +rb_ary_intersect_p(VALUE ary1, VALUE ary2) +{ + VALUE hash, v, result, shorter, longer; + st_data_t vv; + long i; + + ary2 = to_ary(ary2); + if (RARRAY_LEN(ary1) == 0 || RARRAY_LEN(ary2) == 0) return Qfalse; + + if (RARRAY_LEN(ary1) <= SMALL_ARRAY_LEN && RARRAY_LEN(ary2) <= SMALL_ARRAY_LEN) { + for (i=0; i<RARRAY_LEN(ary1); i++) { + v = RARRAY_AREF(ary1, i); + if (rb_ary_includes_by_eql(ary2, v)) return Qtrue; + } + return Qfalse; + } + + shorter = ary1; + longer = ary2; + if (RARRAY_LEN(ary1) > RARRAY_LEN(ary2)) { + longer = ary1; + shorter = ary2; + } + + hash = ary_make_hash(shorter); + result = Qfalse; + + for (i=0; i<RARRAY_LEN(longer); i++) { + v = RARRAY_AREF(longer, i); + vv = (st_data_t)v; + if (rb_hash_stlike_lookup(hash, vv, 0)) { + result = Qtrue; + break; + } + } + ary_recycle_hash(hash); + + return result; +} + +static VALUE +ary_max_generic(VALUE ary, long i, VALUE vmax) +{ + RUBY_ASSERT(i > 0 && i < RARRAY_LEN(ary)); + + VALUE v; + for (; i < RARRAY_LEN(ary); ++i) { + v = RARRAY_AREF(ary, i); + + if (rb_cmpint(rb_funcallv(vmax, id_cmp, 1, &v), vmax, v) < 0) { + vmax = v; + } + } + + return vmax; +} + +static VALUE +ary_max_opt_fixnum(VALUE ary, long i, VALUE vmax) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(FIXNUM_P(vmax)); + + VALUE v; + for (; i < n; ++i) { + v = RARRAY_AREF(ary, i); + + if (FIXNUM_P(v)) { + if ((long)vmax < (long)v) { + vmax = v; + } + } + else { + return ary_max_generic(ary, i, vmax); + } + } + + return vmax; +} + +static VALUE +ary_max_opt_float(VALUE ary, long i, VALUE vmax) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(RB_FLOAT_TYPE_P(vmax)); + + VALUE v; + for (; i < n; ++i) { + v = RARRAY_AREF(ary, i); + + if (RB_FLOAT_TYPE_P(v)) { + if (rb_float_cmp(vmax, v) < 0) { + vmax = v; + } + } + else { + return ary_max_generic(ary, i, vmax); + } + } + + return vmax; +} + +static VALUE +ary_max_opt_string(VALUE ary, long i, VALUE vmax) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(STRING_P(vmax)); + + VALUE v; + for (; i < n; ++i) { + v = RARRAY_AREF(ary, i); + + if (STRING_P(v)) { + if (rb_str_cmp(vmax, v) < 0) { + vmax = v; + } + } + else { + return ary_max_generic(ary, i, vmax); + } + } + + return vmax; +} + +/* + * call-seq: + * array.max -> element + * array.max {|a, b| ... } -> element + * array.max(n) -> new_array + * array.max(n) {|a, b| ... } -> new_array + * + * Returns one of the following: + * + * - The maximum-valued element from +self+. + * - A new \Array of maximum-valued elements selected from +self+. + * + * When no block is given, each element in +self+ must respond to method <tt><=></tt> + * with an \Integer. + * + * With no argument and no block, returns the element in +self+ + * having the maximum value per method <tt><=></tt>: + * + * [0, 1, 2].max # => 2 + * + * With an argument \Integer +n+ and no block, returns a new \Array with at most +n+ elements, + * in descending order per method <tt><=></tt>: + * + * [0, 1, 2, 3].max(3) # => [3, 2, 1] + * [0, 1, 2, 3].max(6) # => [3, 2, 1, 0] + * + * When a block is given, the block must return an \Integer. + * + * With a block and no argument, calls the block <tt>self.size-1</tt> times to compare elements; + * returns the element having the maximum value per the block: + * + * ['0', '00', '000'].max {|a, b| a.size <=> b.size } # => "000" + * + * With an argument +n+ and a block, returns a new \Array with at most +n+ elements, + * in descending order per the block: + * + * ['0', '00', '000'].max(2) {|a, b| a.size <=> b.size } # => ["000", "00"] * - * ary = %w[albatross dog horse] - * ary.max(2) #=> ["horse", "dog"] - * ary.max(2) {|a, b| a.length <=> b.length } #=> ["albatross", "horse"] */ static VALUE rb_ary_max(int argc, VALUE *argv, VALUE ary) { - struct cmp_opt_data cmp_opt = { 0, 0 }; VALUE result = Qundef, v; VALUE num; long i; @@ -4807,52 +6067,167 @@ rb_ary_max(int argc, VALUE *argv, VALUE ary) if (rb_check_arity(argc, 0, 1) && !NIL_P(num = argv[0])) return rb_nmin_run(ary, num, 0, 1, 1); + const long n = RARRAY_LEN(ary); if (rb_block_given_p()) { - for (i = 0; i < RARRAY_LEN(ary); i++) { - v = RARRAY_AREF(ary, i); - if (result == Qundef || rb_cmpint(rb_yield_values(2, v, result), v, result) > 0) { - result = v; - } - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + v = RARRAY_AREF(ary, i); + if (UNDEF_P(result) || rb_cmpint(rb_yield_values(2, v, result), v, result) > 0) { + result = v; + } + } } - else { - for (i = 0; i < RARRAY_LEN(ary); i++) { - v = RARRAY_AREF(ary, i); - if (result == Qundef || OPTIMIZED_CMP(v, result, cmp_opt) > 0) { - result = v; - } - } - } - if (result == Qundef) return Qnil; + else if (n > 0) { + result = RARRAY_AREF(ary, 0); + if (n > 1) { + if (FIXNUM_P(result) && CMP_OPTIMIZABLE(INTEGER)) { + return ary_max_opt_fixnum(ary, 1, result); + } + else if (STRING_P(result) && CMP_OPTIMIZABLE(STRING)) { + return ary_max_opt_string(ary, 1, result); + } + else if (RB_FLOAT_TYPE_P(result) && CMP_OPTIMIZABLE(FLOAT)) { + return ary_max_opt_float(ary, 1, result); + } + else { + return ary_max_generic(ary, 1, result); + } + } + } + if (UNDEF_P(result)) return Qnil; return result; } +static VALUE +ary_min_generic(VALUE ary, long i, VALUE vmin) +{ + RUBY_ASSERT(i > 0 && i < RARRAY_LEN(ary)); + + VALUE v; + for (; i < RARRAY_LEN(ary); ++i) { + v = RARRAY_AREF(ary, i); + + if (rb_cmpint(rb_funcallv(vmin, id_cmp, 1, &v), vmin, v) > 0) { + vmin = v; + } + } + + return vmin; +} + +static VALUE +ary_min_opt_fixnum(VALUE ary, long i, VALUE vmin) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(FIXNUM_P(vmin)); + + VALUE a; + for (; i < n; ++i) { + a = RARRAY_AREF(ary, i); + + if (FIXNUM_P(a)) { + if ((long)vmin > (long)a) { + vmin = a; + } + } + else { + return ary_min_generic(ary, i, vmin); + } + } + + return vmin; +} + +static VALUE +ary_min_opt_float(VALUE ary, long i, VALUE vmin) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(RB_FLOAT_TYPE_P(vmin)); + + VALUE a; + for (; i < n; ++i) { + a = RARRAY_AREF(ary, i); + + if (RB_FLOAT_TYPE_P(a)) { + if (rb_float_cmp(vmin, a) > 0) { + vmin = a; + } + } + else { + return ary_min_generic(ary, i, vmin); + } + } + + return vmin; +} + +static VALUE +ary_min_opt_string(VALUE ary, long i, VALUE vmin) +{ + const long n = RARRAY_LEN(ary); + RUBY_ASSERT(i > 0 && i < n); + RUBY_ASSERT(STRING_P(vmin)); + + VALUE a; + for (; i < n; ++i) { + a = RARRAY_AREF(ary, i); + + if (STRING_P(a)) { + if (rb_str_cmp(vmin, a) > 0) { + vmin = a; + } + } + else { + return ary_min_generic(ary, i, vmin); + } + } + + return vmin; +} + /* * call-seq: - * ary.min -> obj - * ary.min {| a,b | block } -> obj - * ary.min(n) -> array - * ary.min(n) {| a,b | block } -> array + * array.min -> element + * array.min { |a, b| ... } -> element + * array.min(n) -> new_array + * array.min(n) { |a, b| ... } -> new_array + * + * Returns one of the following: + * + * - The minimum-valued element from +self+. + * - A new \Array of minimum-valued elements selected from +self+. * - * Returns the object in _ary_ with the minimum value. The - * first form assumes all objects implement Comparable; - * the second uses the block to return <em>a <=> b</em>. + * When no block is given, each element in +self+ must respond to method <tt><=></tt> + * with an \Integer. * - * ary = %w(albatross dog horse) - * ary.min #=> "albatross" - * ary.min {|a, b| a.length <=> b.length} #=> "dog" + * With no argument and no block, returns the element in +self+ + * having the minimum value per method <tt><=></tt>: * - * If the +n+ argument is given, minimum +n+ elements are returned - * as an array. + * [0, 1, 2].min # => 0 + * + * With \Integer argument +n+ and no block, returns a new \Array with at most +n+ elements, + * in ascending order per method <tt><=></tt>: + * + * [0, 1, 2, 3].min(3) # => [0, 1, 2] + * [0, 1, 2, 3].min(6) # => [0, 1, 2, 3] + * + * When a block is given, the block must return an Integer. + * + * With a block and no argument, calls the block <tt>self.size-1</tt> times to compare elements; + * returns the element having the minimum value per the block: + * + * ['0', '00', '000'].min { |a, b| a.size <=> b.size } # => "0" + * + * With an argument +n+ and a block, returns a new \Array with at most +n+ elements, + * in ascending order per the block: + * + * ['0', '00', '000'].min(2) {|a, b| a.size <=> b.size } # => ["0", "00"] * - * ary = %w[albatross dog horse] - * ary.min(2) #=> ["albatross", "dog"] - * ary.min(2) {|a, b| a.length <=> b.length } #=> ["dog", "horse"] */ static VALUE rb_ary_min(int argc, VALUE *argv, VALUE ary) { - struct cmp_opt_data cmp_opt = { 0, 0 }; VALUE result = Qundef, v; VALUE num; long i; @@ -4860,36 +6235,58 @@ rb_ary_min(int argc, VALUE *argv, VALUE ary) if (rb_check_arity(argc, 0, 1) && !NIL_P(num = argv[0])) return rb_nmin_run(ary, num, 0, 0, 1); + const long n = RARRAY_LEN(ary); if (rb_block_given_p()) { - for (i = 0; i < RARRAY_LEN(ary); i++) { - v = RARRAY_AREF(ary, i); - if (result == Qundef || rb_cmpint(rb_yield_values(2, v, result), v, result) < 0) { - result = v; - } - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + v = RARRAY_AREF(ary, i); + if (UNDEF_P(result) || rb_cmpint(rb_yield_values(2, v, result), v, result) < 0) { + result = v; + } + } } - else { - for (i = 0; i < RARRAY_LEN(ary); i++) { - v = RARRAY_AREF(ary, i); - if (result == Qundef || OPTIMIZED_CMP(v, result, cmp_opt) < 0) { - result = v; - } - } - } - if (result == Qundef) return Qnil; + else if (n > 0) { + result = RARRAY_AREF(ary, 0); + if (n > 1) { + if (FIXNUM_P(result) && CMP_OPTIMIZABLE(INTEGER)) { + return ary_min_opt_fixnum(ary, 1, result); + } + else if (STRING_P(result) && CMP_OPTIMIZABLE(STRING)) { + return ary_min_opt_string(ary, 1, result); + } + else if (RB_FLOAT_TYPE_P(result) && CMP_OPTIMIZABLE(FLOAT)) { + return ary_min_opt_float(ary, 1, result); + } + else { + return ary_min_generic(ary, 1, result); + } + } + } + if (UNDEF_P(result)) return Qnil; return result; } /* * call-seq: - * ary.minmax -> [obj, obj] - * ary.minmax {| a,b | block } -> [obj, obj] + * array.minmax -> [min_val, max_val] + * array.minmax {|a, b| ... } -> [min_val, max_val] * - * Returns a two element array which contains the minimum and the - * maximum value in the array. + * Returns a new 2-element \Array containing the minimum and maximum values + * from +self+, either per method <tt><=></tt> or per a given block:. + * + * When no block is given, each element in +self+ must respond to method <tt><=></tt> + * with an \Integer; + * returns a new 2-element \Array containing the minimum and maximum values + * from +self+, per method <tt><=></tt>: + * + * [0, 1, 2].minmax # => [0, 2] + * + * When a block is given, the block must return an \Integer; + * the block is called <tt>self.size-1</tt> times to compare elements; + * returns a new 2-element \Array containing the minimum and maximum values + * from +self+, per the block: + * + * ['0', '00', '000'].minmax {|a, b| a.size <=> b.size } # => ["0", "000"] * - * Can be given an optional block to override the default comparison - * method <code>a <=> b</code>. */ static VALUE rb_ary_minmax(VALUE ary) @@ -4909,31 +6306,33 @@ push_value(st_data_t key, st_data_t val, st_data_t ary) /* * call-seq: - * ary.uniq! -> ary or nil - * ary.uniq! {|item| ...} -> ary or nil + * array.uniq! -> self or nil + * array.uniq! {|element| ... } -> self or nil * - * Removes duplicate elements from +self+. + * Removes duplicate elements from +self+, the first occurrence always being retained; + * returns +self+ if any elements removed, +nil+ otherwise. * - * If a block is given, it will use the return value of the block for - * comparison. + * With no block given, identifies and removes elements using method <tt>eql?</tt> + * to compare. * - * It compares values using their #hash and #eql? methods for efficiency. + * Returns +self+ if any elements removed: * - * +self+ is traversed in order, and the first occurrence is kept. + * a = [0, 0, 1, 1, 2, 2] + * a.uniq! # => [0, 1, 2] * - * Returns +nil+ if no changes are made (that is, no duplicates are found). + * Returns +nil+ if no elements removed. * - * a = [ "a", "a", "b", "b", "c" ] - * a.uniq! # => ["a", "b", "c"] + * With a block given, calls the block for each element; + * identifies (using method <tt>eql?</tt>) and removes + * elements for which the block returns duplicate values. * - * b = [ "a", "b", "c" ] - * b.uniq! # => nil + * Returns +self+ if any elements removed: * - * c = [["student","sam"], ["student","george"], ["teacher","matz"]] - * c.uniq! {|s| s.first} # => [["student", "sam"], ["teacher", "matz"]] + * a = ['a', 'aa', 'aaa', 'b', 'bb', 'bbb'] + * a.uniq! {|element| element.size } # => ['a', 'aa', 'aaa'] * + * Returns +nil+ if no elements removed. */ - static VALUE rb_ary_uniq_bang(VALUE ary) { @@ -4944,19 +6343,19 @@ rb_ary_uniq_bang(VALUE ary) if (RARRAY_LEN(ary) <= 1) return Qnil; if (rb_block_given_p()) - hash = ary_make_hash_by(ary); + hash = ary_make_hash_by(ary); else - hash = ary_make_hash(ary); + hash = ary_make_hash(ary); hash_size = RHASH_SIZE(hash); if (RARRAY_LEN(ary) == hash_size) { - return Qnil; + return Qnil; } rb_ary_modify_check(ary); ARY_SET_LEN(ary, 0); if (ARY_SHARED_P(ary) && !ARY_EMBED_P(ary)) { - rb_ary_unshare(ary); - FL_SET_EMBED(ary); + rb_ary_unshare(ary); + FL_SET_EMBED(ary); } ary_resize_capa(ary, hash_size); rb_hash_foreach(hash, push_value, ary); @@ -4967,22 +6366,24 @@ rb_ary_uniq_bang(VALUE ary) /* * call-seq: - * ary.uniq -> new_ary - * ary.uniq {|item| ...} -> new_ary - * - * Returns a new array by removing duplicate values in +self+. + * array.uniq -> new_array + * array.uniq {|element| ... } -> new_array * - * If a block is given, it will use the return value of the block for comparison. + * Returns a new \Array containing those elements from +self+ that are not duplicates, + * the first occurrence always being retained. * - * It compares values using their #hash and #eql? methods for efficiency. + * With no block given, identifies and omits duplicates using method <tt>eql?</tt> + * to compare: * - * +self+ is traversed in order, and the first occurrence is kept. + * a = [0, 0, 1, 1, 2, 2] + * a.uniq # => [0, 1, 2] * - * a = [ "a", "a", "b", "b", "c" ] - * a.uniq # => ["a", "b", "c"] + * With a block given, calls the block for each element; + * identifies (using method <tt>eql?</tt>) and omits duplicate values, + * that is, those elements for which the block returns the same value: * - * b = [["student","sam"], ["student","george"], ["teacher","matz"]] - * b.uniq {|s| s.first} # => [["student", "sam"], ["teacher", "matz"]] + * a = ['a', 'aa', 'aaa', 'b', 'bb', 'bbb'] + * a.uniq {|element| element.size } # => ["a", "aa", "aaa"] * */ @@ -4991,32 +6392,32 @@ rb_ary_uniq(VALUE ary) { VALUE hash, uniq; - if (RARRAY_LEN(ary) <= 1) - return rb_ary_dup(ary); - if (rb_block_given_p()) { - hash = ary_make_hash_by(ary); - uniq = rb_hash_values(hash); + if (RARRAY_LEN(ary) <= 1) { + hash = 0; + uniq = rb_ary_dup(ary); + } + else if (rb_block_given_p()) { + hash = ary_make_hash_by(ary); + uniq = rb_hash_values(hash); } else { - hash = ary_make_hash(ary); - uniq = rb_hash_values(hash); + hash = ary_make_hash(ary); + uniq = rb_hash_values(hash); + } + if (hash) { + ary_recycle_hash(hash); } - RBASIC_SET_CLASS(uniq, rb_obj_class(ary)); - ary_recycle_hash(hash); return uniq; } /* * call-seq: - * ary.compact! -> ary or nil + * array.compact! -> self or nil * - * Removes +nil+ elements from the array. + * Removes all +nil+ elements from +self+. * - * Returns +nil+ if no changes were made, otherwise returns the array. - * - * [ "a", nil, "b", nil, "c" ].compact! #=> [ "a", "b", "c" ] - * [ "a", "b", "c" ].compact! #=> nil + * Returns +self+ if any elements removed, otherwise +nil+. */ static VALUE @@ -5030,12 +6431,12 @@ rb_ary_compact_bang(VALUE ary) end = p + RARRAY_LEN(ary); while (t < end) { - if (NIL_P(*t)) t++; - else *p++ = *t++; + if (NIL_P(*t)) t++; + else *p++ = *t++; } n = p - RARRAY_CONST_PTR_TRANSIENT(ary); if (RARRAY_LEN(ary) == n) { - return Qnil; + return Qnil; } ary_resize_smaller(ary, n); @@ -5044,12 +6445,12 @@ rb_ary_compact_bang(VALUE ary) /* * call-seq: - * ary.compact -> new_ary + * array.compact -> new_array * - * Returns a copy of +self+ with all +nil+ elements removed. + * Returns a new \Array containing all non-+nil+ elements from +self+: * - * [ "a", nil, "b", nil, "c", nil ].compact - * #=> [ "a", "b", "c" ] + * a = [nil, 0, nil, 1, nil, 2, nil] + * a.compact # => [0, 1, 2] */ static VALUE @@ -5062,23 +6463,29 @@ rb_ary_compact(VALUE ary) /* * call-seq: - * ary.count -> int - * ary.count(obj) -> int - * ary.count {|item| block} -> int + * array.count -> an_integer + * array.count(obj) -> an_integer + * array.count {|element| ... } -> an_integer * - * Returns the number of elements. + * Returns a count of specified elements. * - * If an argument is given, counts the number of elements which equal +obj+ - * using <code>==</code>. + * With no argument and no block, returns the count of all elements: * - * If a block is given, counts the number of elements for which the block - * returns a true value. + * [0, 1, 2].count # => 3 + * [].count # => 0 * - * ary = [1, 2, 4, 2] - * ary.count #=> 4 - * ary.count(2) #=> 2 - * ary.count {|x| x%2 == 0} #=> 3 + * With argument +obj+, returns the count of elements <tt>==</tt> to +obj+: * + * [0, 1, 2, 0.0].count(0) # => 2 + * [0, 1, 2].count(3) # => 0 + * + * With no argument and a block given, calls the block with each element; + * returns the count of elements for which the block returns a truthy value: + * + * [0, 1, 2, 3].count {|element| element > 1} # => 2 + * + * With argument +obj+ and a block given, issues a warning, ignores the block, + * and returns the count of elements <tt>==</tt> to +obj+. */ static VALUE @@ -5087,106 +6494,151 @@ rb_ary_count(int argc, VALUE *argv, VALUE ary) long i, n = 0; if (rb_check_arity(argc, 0, 1) == 0) { - VALUE v; + VALUE v; - if (!rb_block_given_p()) - return LONG2NUM(RARRAY_LEN(ary)); + if (!rb_block_given_p()) + return LONG2NUM(RARRAY_LEN(ary)); - for (i = 0; i < RARRAY_LEN(ary); i++) { - v = RARRAY_AREF(ary, i); - if (RTEST(rb_yield(v))) n++; - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + v = RARRAY_AREF(ary, i); + if (RTEST(rb_yield(v))) n++; + } } else { VALUE obj = argv[0]; - if (rb_block_given_p()) { - rb_warn("given block not used"); - } - for (i = 0; i < RARRAY_LEN(ary); i++) { - if (rb_equal(RARRAY_AREF(ary, i), obj)) n++; - } + if (rb_block_given_p()) { + rb_warn("given block not used"); + } + for (i = 0; i < RARRAY_LEN(ary); i++) { + if (rb_equal(RARRAY_AREF(ary, i), obj)) n++; + } } return LONG2NUM(n); } static VALUE -flatten(VALUE ary, int level, int *modified) +flatten(VALUE ary, int level) { - long i = 0; - VALUE stack, result, tmp, elt; - st_table *memo; + long i; + VALUE stack, result, tmp = 0, elt, vmemo; + st_table *memo = 0; st_data_t id; - stack = ary_new(0, ARY_DEFAULT_SIZE); + for (i = 0; i < RARRAY_LEN(ary); i++) { + elt = RARRAY_AREF(ary, i); + tmp = rb_check_array_type(elt); + if (!NIL_P(tmp)) { + break; + } + } + if (i == RARRAY_LEN(ary)) { + return ary; + } + result = ary_new(0, RARRAY_LEN(ary)); - memo = st_init_numtable(); - st_insert(memo, (st_data_t)ary, (st_data_t)Qtrue); - *modified = 0; + ary_memcpy(result, 0, i, RARRAY_CONST_PTR_TRANSIENT(ary)); + ARY_SET_LEN(result, i); + + stack = ary_new(0, ARY_DEFAULT_SIZE); + rb_ary_push(stack, ary); + rb_ary_push(stack, LONG2NUM(i + 1)); + + if (level < 0) { + vmemo = rb_hash_new(); + RBASIC_CLEAR_CLASS(vmemo); + memo = st_init_numtable(); + rb_hash_st_table_set(vmemo, memo); + st_insert(memo, (st_data_t)ary, (st_data_t)Qtrue); + st_insert(memo, (st_data_t)tmp, (st_data_t)Qtrue); + } + + ary = tmp; + i = 0; while (1) { - while (i < RARRAY_LEN(ary)) { - elt = RARRAY_AREF(ary, i++); - if (level >= 0 && RARRAY_LEN(stack) / 2 >= level) { - rb_ary_push(result, elt); - continue; - } - tmp = rb_check_array_type(elt); - if (RBASIC(result)->klass) { - rb_raise(rb_eRuntimeError, "flatten reentered"); - } - if (NIL_P(tmp)) { - rb_ary_push(result, elt); - } - else { - *modified = 1; - id = (st_data_t)tmp; - if (st_lookup(memo, id, 0)) { - st_free_table(memo); - rb_raise(rb_eArgError, "tried to flatten recursive array"); - } - st_insert(memo, id, (st_data_t)Qtrue); - rb_ary_push(stack, ary); - rb_ary_push(stack, LONG2NUM(i)); - ary = tmp; - i = 0; - } - } - if (RARRAY_LEN(stack) == 0) { - break; - } - id = (st_data_t)ary; - st_delete(memo, &id, 0); - tmp = rb_ary_pop(stack); - i = NUM2LONG(tmp); - ary = rb_ary_pop(stack); - } - - st_free_table(memo); - - RBASIC_SET_CLASS(result, rb_obj_class(ary)); + while (i < RARRAY_LEN(ary)) { + elt = RARRAY_AREF(ary, i++); + if (level >= 0 && RARRAY_LEN(stack) / 2 >= level) { + rb_ary_push(result, elt); + continue; + } + tmp = rb_check_array_type(elt); + if (RBASIC(result)->klass) { + if (memo) { + RB_GC_GUARD(vmemo); + st_clear(memo); + } + rb_raise(rb_eRuntimeError, "flatten reentered"); + } + if (NIL_P(tmp)) { + rb_ary_push(result, elt); + } + else { + if (memo) { + id = (st_data_t)tmp; + if (st_is_member(memo, id)) { + st_clear(memo); + rb_raise(rb_eArgError, "tried to flatten recursive array"); + } + st_insert(memo, id, (st_data_t)Qtrue); + } + rb_ary_push(stack, ary); + rb_ary_push(stack, LONG2NUM(i)); + ary = tmp; + i = 0; + } + } + if (RARRAY_LEN(stack) == 0) { + break; + } + if (memo) { + id = (st_data_t)ary; + st_delete(memo, &id, 0); + } + tmp = rb_ary_pop(stack); + i = NUM2LONG(tmp); + ary = rb_ary_pop(stack); + } + + if (memo) { + st_clear(memo); + } + + RBASIC_SET_CLASS(result, rb_cArray); return result; } /* * call-seq: - * ary.flatten! -> ary or nil - * ary.flatten!(level) -> ary or nil + * array.flatten! -> self or nil + * array.flatten!(level) -> self or nil + * + * Replaces each nested \Array in +self+ with the elements from that \Array; + * returns +self+ if any changes, +nil+ otherwise. + * + * With non-negative \Integer argument +level+, flattens recursively through +level+ levels: * - * Flattens +self+ in place. + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten!(1) # => [0, 1, [2, 3], 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten!(2) # => [0, 1, 2, 3, 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten!(3) # => [0, 1, 2, 3, 4, 5] + * [0, 1, 2].flatten!(1) # => nil * - * Returns +nil+ if no modifications were made (i.e., the array contains no - * subarrays.) + * With no argument, a +nil+ argument, or with negative argument +level+, flattens all levels: * - * The optional +level+ argument determines the level of recursion to flatten. + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten! # => [0, 1, 2, 3, 4, 5] + * [0, 1, 2].flatten! # => nil + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten!(-1) # => [0, 1, 2, 3, 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten!(-2) # => [0, 1, 2, 3, 4, 5] + * [0, 1, 2].flatten!(-1) # => nil * - * a = [ 1, 2, [3, [4, 5] ] ] - * a.flatten! #=> [1, 2, 3, 4, 5] - * a.flatten! #=> nil - * a #=> [1, 2, 3, 4, 5] - * a = [ 1, 2, [3, [4, 5] ] ] - * a.flatten!(1) #=> [1, 2, 3, [4, 5]] */ static VALUE @@ -5200,10 +6652,9 @@ rb_ary_flatten_bang(int argc, VALUE *argv, VALUE ary) if (!NIL_P(lv)) level = NUM2INT(lv); if (level == 0) return Qnil; - result = flatten(ary, level, &mod); - if (mod == 0) { - ary_discard(result); - return Qnil; + result = flatten(ary, level); + if (result == ary) { + return Qnil; } if (!(mod = ARY_EMBED_P(result))) rb_obj_freeze(result); rb_ary_replace(ary, result); @@ -5214,30 +6665,41 @@ rb_ary_flatten_bang(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.flatten -> new_ary - * ary.flatten(level) -> new_ary - * - * Returns a new array that is a one-dimensional flattening of +self+ - * (recursively). + * array.flatten -> new_array + * array.flatten(level) -> new_array + * + * Returns a new \Array that is a recursive flattening of +self+: + * - Each non-Array element is unchanged. + * - Each \Array is replaced by its individual elements. + * + * With non-negative \Integer argument +level+, flattens recursively through +level+ levels: + * + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(0) # => [0, [1, [2, 3], 4], 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(1) # => [0, 1, [2, 3], 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(2) # => [0, 1, 2, 3, 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(3) # => [0, 1, 2, 3, 4, 5] + * + * With no argument, a +nil+ argument, or with negative argument +level+, flattens all levels: + * + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten # => [0, 1, 2, 3, 4, 5] + * [0, 1, 2].flatten # => [0, 1, 2] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(-1) # => [0, 1, 2, 3, 4, 5] + * a = [ 0, [ 1, [2, 3], 4 ], 5 ] + * a.flatten(-2) # => [0, 1, 2, 3, 4, 5] + * [0, 1, 2].flatten(-1) # => [0, 1, 2] * - * That is, for every element that is an array, extract its elements into - * the new array. - * - * The optional +level+ argument determines the level of recursion to - * flatten. - * - * s = [ 1, 2, 3 ] #=> [1, 2, 3] - * t = [ 4, 5, 6, [7, 8] ] #=> [4, 5, 6, [7, 8]] - * a = [ s, t, 9, 10 ] #=> [[1, 2, 3], [4, 5, 6, [7, 8]], 9, 10] - * a.flatten #=> [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] - * a = [ 1, 2, [3, [4, 5] ] ] - * a.flatten(1) #=> [1, 2, 3, [4, 5]] */ static VALUE rb_ary_flatten(int argc, VALUE *argv, VALUE ary) { - int mod = 0, level = -1; + int level = -1; VALUE result; if (rb_check_arity(argc, 0, 1) && !NIL_P(argv[0])) { @@ -5245,257 +6707,170 @@ rb_ary_flatten(int argc, VALUE *argv, VALUE ary) if (level == 0) return ary_make_shared_copy(ary); } - result = flatten(ary, level, &mod); - OBJ_INFECT(result, ary); + result = flatten(ary, level); + if (result == ary) { + result = ary_make_shared_copy(ary); + } return result; } -#define OPTHASH_GIVEN_P(opts) \ - (argc > 0 && !NIL_P((opts) = rb_check_hash_type(argv[argc-1])) && (--argc, 1)) -static ID id_random; - #define RAND_UPTO(max) (long)rb_random_ulong_limited((randgen), (max)-1) -/* - * call-seq: - * ary.shuffle! -> ary - * ary.shuffle!(random: rng) -> ary - * - * Shuffles elements in +self+ in place. - * - * a = [ 1, 2, 3 ] #=> [1, 2, 3] - * a.shuffle! #=> [2, 3, 1] - * a #=> [2, 3, 1] - * - * The optional +rng+ argument will be used as the random number generator. - * - * a.shuffle!(random: Random.new(1)) #=> [1, 3, 2] - */ - static VALUE -rb_ary_shuffle_bang(int argc, VALUE *argv, VALUE ary) +rb_ary_shuffle_bang(rb_execution_context_t *ec, VALUE ary, VALUE randgen) { - VALUE opts, randgen = rb_cRandom; long i, len; - if (OPTHASH_GIVEN_P(opts)) { - VALUE rnd; - ID keyword_ids[1]; - - keyword_ids[0] = id_random; - rb_get_kwargs(opts, keyword_ids, 0, 1, &rnd); - if (rnd != Qundef) { - randgen = rnd; - } - } - rb_check_arity(argc, 0, 0); rb_ary_modify(ary); i = len = RARRAY_LEN(ary); RARRAY_PTR_USE(ary, ptr, { - while (i) { - long j = RAND_UPTO(i); - VALUE tmp; + while (i) { + long j = RAND_UPTO(i); + VALUE tmp; if (len != RARRAY_LEN(ary) || ptr != RARRAY_CONST_PTR_TRANSIENT(ary)) { rb_raise(rb_eRuntimeError, "modified during shuffle"); - } - tmp = ptr[--i]; - ptr[i] = ptr[j]; - ptr[j] = tmp; - } + } + tmp = ptr[--i]; + ptr[i] = ptr[j]; + ptr[j] = tmp; + } }); /* WB: no new reference */ return ary; } - -/* - * call-seq: - * ary.shuffle -> new_ary - * ary.shuffle(random: rng) -> new_ary - * - * Returns a new array with elements of +self+ shuffled. - * - * a = [ 1, 2, 3 ] #=> [1, 2, 3] - * a.shuffle #=> [2, 3, 1] - * a #=> [1, 2, 3] - * - * The optional +rng+ argument will be used as the random number generator. - * - * a.shuffle(random: Random.new(1)) #=> [1, 3, 2] - */ - static VALUE -rb_ary_shuffle(int argc, VALUE *argv, VALUE ary) +rb_ary_shuffle(rb_execution_context_t *ec, VALUE ary, VALUE randgen) { ary = rb_ary_dup(ary); - rb_ary_shuffle_bang(argc, argv, ary); + rb_ary_shuffle_bang(ec, ary, randgen); return ary; } - -/* - * call-seq: - * ary.sample -> obj - * ary.sample(random: rng) -> obj - * ary.sample(n) -> new_ary - * ary.sample(n, random: rng) -> new_ary - * - * Choose a random element or +n+ random elements from the array. - * - * The elements are chosen by using random and unique indices into the array - * in order to ensure that an element doesn't repeat itself unless the array - * already contained duplicate elements. - * - * If the array is empty the first form returns +nil+ and the second form - * returns an empty array. - * - * a = [ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ] - * a.sample #=> 7 - * a.sample(4) #=> [6, 4, 2, 5] - * - * The optional +rng+ argument will be used as the random number generator. - * - * a.sample(random: Random.new(1)) #=> 6 - * a.sample(4, random: Random.new(1)) #=> [6, 10, 9, 2] - */ - - static VALUE -rb_ary_sample(int argc, VALUE *argv, VALUE ary) +ary_sample(rb_execution_context_t *ec, VALUE ary, VALUE randgen, VALUE nv, VALUE to_array) { - VALUE nv, result; - VALUE opts, randgen = rb_cRandom; + VALUE result; long n, len, i, j, k, idx[10]; long rnds[numberof(idx)]; long memo_threshold; - if (OPTHASH_GIVEN_P(opts)) { - VALUE rnd; - ID keyword_ids[1]; - - keyword_ids[0] = id_random; - rb_get_kwargs(opts, keyword_ids, 0, 1, &rnd); - if (rnd != Qundef) { - randgen = rnd; - } - } len = RARRAY_LEN(ary); - if (rb_check_arity(argc, 0, 1) == 0) { - if (len < 2) - i = 0; - else - i = RAND_UPTO(len); + if (!to_array) { + if (len < 2) + i = 0; + else + i = RAND_UPTO(len); - return rb_ary_elt(ary, i); + return rb_ary_elt(ary, i); } - nv = argv[0]; n = NUM2LONG(nv); if (n < 0) rb_raise(rb_eArgError, "negative sample number"); if (n > len) n = len; if (n <= numberof(idx)) { - for (i = 0; i < n; ++i) { - rnds[i] = RAND_UPTO(len - i); - } + for (i = 0; i < n; ++i) { + rnds[i] = RAND_UPTO(len - i); + } } k = len; len = RARRAY_LEN(ary); if (len < k && n <= numberof(idx)) { - for (i = 0; i < n; ++i) { - if (rnds[i] >= len) return rb_ary_new_capa(0); - } + for (i = 0; i < n; ++i) { + if (rnds[i] >= len) return rb_ary_new_capa(0); + } } if (n > len) n = len; switch (n) { case 0: - return rb_ary_new_capa(0); + return rb_ary_new_capa(0); case 1: - i = rnds[0]; - return rb_ary_new_from_values(1, &RARRAY_AREF(ary, i)); + i = rnds[0]; + return rb_ary_new_from_args(1, RARRAY_AREF(ary, i)); case 2: - i = rnds[0]; - j = rnds[1]; - if (j >= i) j++; - return rb_ary_new_from_args(2, RARRAY_AREF(ary, i), RARRAY_AREF(ary, j)); + i = rnds[0]; + j = rnds[1]; + if (j >= i) j++; + return rb_ary_new_from_args(2, RARRAY_AREF(ary, i), RARRAY_AREF(ary, j)); case 3: - i = rnds[0]; - j = rnds[1]; - k = rnds[2]; - { - long l = j, g = i; - if (j >= i) l = i, g = ++j; - if (k >= l && (++k >= g)) ++k; - } - return rb_ary_new_from_args(3, RARRAY_AREF(ary, i), RARRAY_AREF(ary, j), RARRAY_AREF(ary, k)); + i = rnds[0]; + j = rnds[1]; + k = rnds[2]; + { + long l = j, g = i; + if (j >= i) l = i, g = ++j; + if (k >= l && (++k >= g)) ++k; + } + return rb_ary_new_from_args(3, RARRAY_AREF(ary, i), RARRAY_AREF(ary, j), RARRAY_AREF(ary, k)); } memo_threshold = - len < 2560 ? len / 128 : - len < 5120 ? len / 64 : - len < 10240 ? len / 32 : - len / 16; + len < 2560 ? len / 128 : + len < 5120 ? len / 64 : + len < 10240 ? len / 32 : + len / 16; if (n <= numberof(idx)) { - long sorted[numberof(idx)]; - sorted[0] = idx[0] = rnds[0]; - for (i=1; i<n; i++) { - k = rnds[i]; - for (j = 0; j < i; ++j) { - if (k < sorted[j]) break; - ++k; - } - memmove(&sorted[j+1], &sorted[j], sizeof(sorted[0])*(i-j)); - sorted[j] = idx[i] = k; - } - result = rb_ary_new_capa(n); + long sorted[numberof(idx)]; + sorted[0] = idx[0] = rnds[0]; + for (i=1; i<n; i++) { + k = rnds[i]; + for (j = 0; j < i; ++j) { + if (k < sorted[j]) break; + ++k; + } + memmove(&sorted[j+1], &sorted[j], sizeof(sorted[0])*(i-j)); + sorted[j] = idx[i] = k; + } + result = rb_ary_new_capa(n); RARRAY_PTR_USE_TRANSIENT(result, ptr_result, { - for (i=0; i<n; i++) { - ptr_result[i] = RARRAY_AREF(ary, idx[i]); - } - }); + for (i=0; i<n; i++) { + ptr_result[i] = RARRAY_AREF(ary, idx[i]); + } + }); } else if (n <= memo_threshold / 2) { - long max_idx = 0; + long max_idx = 0; #undef RUBY_UNTYPED_DATA_WARNING #define RUBY_UNTYPED_DATA_WARNING 0 - VALUE vmemo = Data_Wrap_Struct(0, 0, st_free_table, 0); - st_table *memo = st_init_numtable_with_size(n); - DATA_PTR(vmemo) = memo; - result = rb_ary_new_capa(n); - RARRAY_PTR_USE(result, ptr_result, { - for (i=0; i<n; i++) { - long r = RAND_UPTO(len-i) + i; - ptr_result[i] = r; - if (r > max_idx) max_idx = r; - } - len = RARRAY_LEN(ary); - if (len <= max_idx) n = 0; - else if (n > len) n = len; + VALUE vmemo = Data_Wrap_Struct(0, 0, st_free_table, 0); + st_table *memo = st_init_numtable_with_size(n); + DATA_PTR(vmemo) = memo; + result = rb_ary_new_capa(n); + RARRAY_PTR_USE(result, ptr_result, { + for (i=0; i<n; i++) { + long r = RAND_UPTO(len-i) + i; + ptr_result[i] = r; + if (r > max_idx) max_idx = r; + } + len = RARRAY_LEN(ary); + if (len <= max_idx) n = 0; + else if (n > len) n = len; RARRAY_PTR_USE_TRANSIENT(ary, ptr_ary, { - for (i=0; i<n; i++) { - long j2 = j = ptr_result[i]; - long i2 = i; - st_data_t value; - if (st_lookup(memo, (st_data_t)i, &value)) i2 = (long)value; - if (st_lookup(memo, (st_data_t)j, &value)) j2 = (long)value; - st_insert(memo, (st_data_t)j, (st_data_t)i2); - ptr_result[i] = ptr_ary[j2]; - } - }); - }); - DATA_PTR(vmemo) = 0; - st_free_table(memo); + for (i=0; i<n; i++) { + long j2 = j = ptr_result[i]; + long i2 = i; + st_data_t value; + if (st_lookup(memo, (st_data_t)i, &value)) i2 = (long)value; + if (st_lookup(memo, (st_data_t)j, &value)) j2 = (long)value; + st_insert(memo, (st_data_t)j, (st_data_t)i2); + ptr_result[i] = ptr_ary[j2]; + } + }); + }); + DATA_PTR(vmemo) = 0; + st_free_table(memo); } else { - result = rb_ary_dup(ary); - RBASIC_CLEAR_CLASS(result); - RB_GC_GUARD(ary); - RARRAY_PTR_USE(result, ptr_result, { - for (i=0; i<n; i++) { - j = RAND_UPTO(len-i) + i; - nv = ptr_result[j]; - ptr_result[j] = ptr_result[i]; - ptr_result[i] = nv; - } - }); - RBASIC_SET_CLASS_RAW(result, rb_cArray); + result = rb_ary_dup(ary); + RBASIC_CLEAR_CLASS(result); + RB_GC_GUARD(ary); + RARRAY_PTR_USE(result, ptr_result, { + for (i=0; i<n; i++) { + j = RAND_UPTO(len-i) + i; + nv = ptr_result[j]; + ptr_result[j] = ptr_result[i]; + ptr_result[i] = nv; + } + }); + RBASIC_SET_CLASS_RAW(result, rb_cArray); } ARY_SET_LEN(result, n); @@ -5503,15 +6878,21 @@ rb_ary_sample(int argc, VALUE *argv, VALUE ary) } static VALUE +ary_sample0(rb_execution_context_t *ec, VALUE ary) +{ + return ary_sample(ec, ary, rb_cRandom, Qfalse, Qfalse); +} + +static VALUE rb_ary_cycle_size(VALUE self, VALUE args, VALUE eobj) { long mul; VALUE n = Qnil; if (args && (RARRAY_LEN(args) > 0)) { - n = RARRAY_AREF(args, 0); + n = RARRAY_AREF(args, 0); } if (RARRAY_LEN(self) == 0) return INT2FIX(0); - if (n == Qnil) return DBL2NUM(HUGE_VAL); + if (NIL_P(n)) return DBL2NUM(HUGE_VAL); mul = NUM2LONG(n); if (mul <= 0) return INT2FIX(0); n = LONG2FIX(mul); @@ -5520,24 +6901,37 @@ rb_ary_cycle_size(VALUE self, VALUE args, VALUE eobj) /* * call-seq: - * ary.cycle(n=nil) {|obj| block} -> nil - * ary.cycle(n=nil) -> Enumerator + * array.cycle {|element| ... } -> nil + * array.cycle(count) {|element| ... } -> nil + * array.cycle -> new_enumerator + * array.cycle(count) -> new_enumerator + * + * When called with positive \Integer argument +count+ and a block, + * calls the block with each element, then does so again, + * until it has done so +count+ times; returns +nil+: + * + * output = [] + * [0, 1].cycle(2) {|element| output.push(element) } # => nil + * output # => [0, 1, 0, 1] * - * Calls the given block for each element +n+ times or forever if +nil+ is - * given. + * If +count+ is zero or negative, does not call the block: * - * Does nothing if a non-positive number is given or the array is empty. + * [0, 1].cycle(0) {|element| fail 'Cannot happen' } # => nil + * [0, 1].cycle(-1) {|element| fail 'Cannot happen' } # => nil * - * Returns +nil+ if the loop has finished without getting interrupted. + * When a block is given, and argument is omitted or +nil+, cycles forever: * - * If no block is given, an Enumerator is returned instead. + * # Prints 0 and 1 forever. + * [0, 1].cycle {|element| puts element } + * [0, 1].cycle(nil) {|element| puts element } * - * a = ["a", "b", "c"] - * a.cycle {|x| puts x} # print, a, b, c, a, b, c,.. forever. - * a.cycle(2) {|x| puts x} # print, a, b, c, a, b, c. + * When no block is given, returns a new \Enumerator: + * + * [0, 1].cycle(2) # => #<Enumerator: [0, 1]:cycle(2)> + * [0, 1].cycle # => # => #<Enumerator: [0, 1]:cycle> + * [0, 1].cycle.first(5) # => [0, 1, 0, 1, 0] * */ - static VALUE rb_ary_cycle(int argc, VALUE *argv, VALUE ary) { @@ -5562,9 +6956,6 @@ rb_ary_cycle(int argc, VALUE *argv, VALUE ary) return Qnil; } -#define tmpary(n) rb_ary_tmp_new(n) -#define tmpary_discard(a) (ary_discard(a), RBASIC_SET_CLASS_RAW(a, rb_cArray)) - /* * Build a ruby array of the corresponding values and yield it to the * associated block. @@ -5600,52 +6991,52 @@ permute0(const long n, const long r, long *const p, char *const used, const VALU long i = 0, index = 0; for (;;) { - const char *const unused = memchr(&used[i], 0, n-i); - if (!unused) { - if (!index) break; - i = p[--index]; /* pop index */ - used[i++] = 0; /* index unused */ - } - else { - i = unused - used; - p[index] = i; - used[i] = 1; /* mark index used */ - ++index; - if (index < r-1) { /* if not done yet */ - p[index] = i = 0; - continue; - } - for (i = 0; i < n; ++i) { - if (used[i]) continue; - p[index] = i; - if (!yield_indexed_values(values, r, p)) { - rb_raise(rb_eRuntimeError, "permute reentered"); - } - } - i = p[--index]; /* pop index */ - used[i] = 0; /* index unused */ - p[index] = ++i; - } + const char *const unused = memchr(&used[i], 0, n-i); + if (!unused) { + if (!index) break; + i = p[--index]; /* pop index */ + used[i++] = 0; /* index unused */ + } + else { + i = unused - used; + p[index] = i; + used[i] = 1; /* mark index used */ + ++index; + if (index < r-1) { /* if not done yet */ + p[index] = i = 0; + continue; + } + for (i = 0; i < n; ++i) { + if (used[i]) continue; + p[index] = i; + if (!yield_indexed_values(values, r, p)) { + rb_raise(rb_eRuntimeError, "permute reentered"); + } + } + i = p[--index]; /* pop index */ + used[i] = 0; /* index unused */ + p[index] = ++i; + } } } /* * Returns the product of from, from-1, ..., from - how_many + 1. - * http://en.wikipedia.org/wiki/Pochhammer_symbol + * https://en.wikipedia.org/wiki/Pochhammer_symbol */ static VALUE descending_factorial(long from, long how_many) { VALUE cnt; if (how_many > 0) { - cnt = LONG2FIX(from); - while (--how_many > 0) { - long v = --from; - cnt = rb_int_mul(cnt, LONG2FIX(v)); - } + cnt = LONG2FIX(from); + while (--how_many > 0) { + long v = --from; + cnt = rb_int_mul(cnt, LONG2FIX(v)); + } } else { - cnt = LONG2FIX(how_many == 0); + cnt = LONG2FIX(how_many == 0); } return cnt; } @@ -5656,18 +7047,18 @@ binomial_coefficient(long comb, long size) VALUE r; long i; if (comb > size-comb) { - comb = size-comb; + comb = size-comb; } if (comb < 0) { - return LONG2FIX(0); + return LONG2FIX(0); } else if (comb == 0) { - return LONG2FIX(1); + return LONG2FIX(1); } r = LONG2FIX(size); for (i = 1; i < comb; ++i) { - r = rb_int_mul(r, LONG2FIX(size - i)); - r = rb_int_idiv(r, LONG2FIX(i + 1)); + r = rb_int_mul(r, LONG2FIX(size - i)); + r = rb_int_idiv(r, LONG2FIX(i + 1)); } return r; } @@ -5683,30 +7074,82 @@ rb_ary_permutation_size(VALUE ary, VALUE args, VALUE eobj) /* * call-seq: - * ary.permutation {|p| block} -> ary - * ary.permutation -> Enumerator - * ary.permutation(n) {|p| block} -> ary - * ary.permutation(n) -> Enumerator + * array.permutation {|element| ... } -> self + * array.permutation(n) {|element| ... } -> self + * array.permutation -> new_enumerator + * array.permutation(n) -> new_enumerator + * + * When invoked with a block, yield all permutations of elements of +self+; returns +self+. + * The order of permutations is indeterminate. + * + * When a block and an in-range positive \Integer argument +n+ (<tt>0 < n <= self.size</tt>) + * are given, calls the block with all +n+-tuple permutations of +self+. + * + * Example: + * + * a = [0, 1, 2] + * a.permutation(2) {|permutation| p permutation } + * + * Output: + * + * [0, 1] + * [0, 2] + * [1, 0] + * [1, 2] + * [2, 0] + * [2, 1] + * + * Another example: + * + * a = [0, 1, 2] + * a.permutation(3) {|permutation| p permutation } + * + * Output: + * + * [0, 1, 2] + * [0, 2, 1] + * [1, 0, 2] + * [1, 2, 0] + * [2, 0, 1] + * [2, 1, 0] + * + * When +n+ is zero, calls the block once with a new empty \Array: + * + * a = [0, 1, 2] + * a.permutation(0) {|permutation| p permutation } + * + * Output: * - * When invoked with a block, yield all permutations of length +n+ of the - * elements of the array, then return the array itself. + * [] * - * If +n+ is not specified, yield all permutations of all elements. + * When +n+ is out of range (negative or larger than <tt>self.size</tt>), + * does not call the block: * - * The implementation makes no guarantees about the order in which the - * permutations are yielded. + * a = [0, 1, 2] + * a.permutation(-1) {|permutation| fail 'Cannot happen' } + * a.permutation(4) {|permutation| fail 'Cannot happen' } * - * If no block is given, an Enumerator is returned instead. + * When a block given but no argument, + * behaves the same as <tt>a.permutation(a.size)</tt>: * - * Examples: + * a = [0, 1, 2] + * a.permutation {|permutation| p permutation } + * + * Output: + * + * [0, 1, 2] + * [0, 2, 1] + * [1, 0, 2] + * [1, 2, 0] + * [2, 0, 1] + * [2, 1, 0] + * + * Returns a new \Enumerator if no block given: + * + * a = [0, 1, 2] + * a.permutation # => #<Enumerator: [0, 1, 2]:permutation> + * a.permutation(2) # => #<Enumerator: [0, 1, 2]:permutation(2)> * - * a = [1, 2, 3] - * a.permutation.to_a #=> [[1,2,3],[1,3,2],[2,1,3],[2,3,1],[3,1,2],[3,2,1]] - * a.permutation(1).to_a #=> [[1],[2],[3]] - * a.permutation(2).to_a #=> [[1,2],[1,3],[2,1],[2,3],[3,1],[3,2]] - * a.permutation(3).to_a #=> [[1,2,3],[1,3,2],[2,1,3],[2,3,1],[3,1,2],[3,2,1]] - * a.permutation(0).to_a #=> [[]] # one permutation of length 0 - * a.permutation(4).to_a #=> [] # no permutations of length 4 */ static VALUE @@ -5721,28 +7164,28 @@ rb_ary_permutation(int argc, VALUE *argv, VALUE ary) r = NUM2LONG(argv[0]); /* Permutation size from argument */ if (r < 0 || n < r) { - /* no permutations: yield nothing */ + /* no permutations: yield nothing */ } else if (r == 0) { /* exactly one permutation: the zero-length array */ - rb_yield(rb_ary_new2(0)); + rb_yield(rb_ary_new2(0)); } else if (r == 1) { /* this is a special, easy case */ - for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); + } } else { /* this is the general case */ - volatile VALUE t0; - long *p = ALLOCV_N(long, t0, r+roomof(n, sizeof(long))); - char *used = (char*)(p + r); - VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - RBASIC_CLEAR_CLASS(ary0); + volatile VALUE t0; + long *p = ALLOCV_N(long, t0, r+roomof(n, sizeof(long))); + char *used = (char*)(p + r); + VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ + RBASIC_CLEAR_CLASS(ary0); - MEMZERO(used, char, n); /* initialize array */ + MEMZERO(used, char, n); /* initialize array */ - permute0(n, r, p, used, ary0); /* compute and yield permutations */ - ALLOCV_END(t0); - RBASIC_SET_CLASS_RAW(ary0, rb_cArray); + permute0(n, r, p, used, ary0); /* compute and yield permutations */ + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } @@ -5755,16 +7198,16 @@ combinate0(const long len, const long n, long *const stack, const VALUE values) MEMZERO(stack+1, long, n); stack[0] = -1; for (;;) { - for (lev++; lev < n; lev++) { - stack[lev+1] = stack[lev]+1; - } - if (!yield_indexed_values(values, n, stack+1)) { - rb_raise(rb_eRuntimeError, "combination reentered"); - } - do { - if (lev == 0) return; - stack[lev--]++; - } while (stack[lev+1]+n == len+lev+1); + for (lev++; lev < n; lev++) { + stack[lev+1] = stack[lev]+1; + } + if (!yield_indexed_values(values, n, stack+1)) { + rb_raise(rb_eRuntimeError, "combination reentered"); + } + do { + if (lev == 0) return; + stack[lev--]++; + } while (stack[lev+1]+n == len+lev+1); } } @@ -5779,26 +7222,55 @@ rb_ary_combination_size(VALUE ary, VALUE args, VALUE eobj) /* * call-seq: - * ary.combination(n) {|c| block} -> ary - * ary.combination(n) -> Enumerator + * array.combination(n) {|element| ... } -> self + * array.combination(n) -> new_enumerator + * + * Calls the block, if given, with combinations of elements of +self+; + * returns +self+. The order of combinations is indeterminate. + * + * When a block and an in-range positive \Integer argument +n+ (<tt>0 < n <= self.size</tt>) + * are given, calls the block with all +n+-tuple combinations of +self+. + * + * Example: + * + * a = [0, 1, 2] + * a.combination(2) {|combination| p combination } * - * When invoked with a block, yields all combinations of length +n+ of elements - * from the array and then returns the array itself. + * Output: * - * The implementation makes no guarantees about the order in which the - * combinations are yielded. + * [0, 1] + * [0, 2] + * [1, 2] * - * If no block is given, an Enumerator is returned instead. + * Another example: * - * Examples: + * a = [0, 1, 2] + * a.combination(3) {|combination| p combination } * - * a = [1, 2, 3, 4] - * a.combination(1).to_a #=> [[1],[2],[3],[4]] - * a.combination(2).to_a #=> [[1,2],[1,3],[1,4],[2,3],[2,4],[3,4]] - * a.combination(3).to_a #=> [[1,2,3],[1,2,4],[1,3,4],[2,3,4]] - * a.combination(4).to_a #=> [[1,2,3,4]] - * a.combination(0).to_a #=> [[]] # one combination of length 0 - * a.combination(5).to_a #=> [] # no combinations of length 5 + * Output: + * + * [0, 1, 2] + * + * When +n+ is zero, calls the block once with a new empty \Array: + * + * a = [0, 1, 2] + * a1 = a.combination(0) {|combination| p combination } + * + * Output: + * + * [] + * + * When +n+ is out of range (negative or larger than <tt>self.size</tt>), + * does not call the block: + * + * a = [0, 1, 2] + * a.combination(-1) {|combination| fail 'Cannot happen' } + * a.combination(4) {|combination| fail 'Cannot happen' } + * + * Returns a new \Enumerator if no block given: + * + * a = [0, 1, 2] + * a.combination(2) # => #<Enumerator: [0, 1, 2]:combination(2)> * */ @@ -5811,25 +7283,25 @@ rb_ary_combination(VALUE ary, VALUE num) RETURN_SIZED_ENUMERATOR(ary, 1, &num, rb_ary_combination_size); len = RARRAY_LEN(ary); if (n < 0 || len < n) { - /* yield nothing */ + /* yield nothing */ } else if (n == 0) { - rb_yield(rb_ary_new2(0)); + rb_yield(rb_ary_new2(0)); } else if (n == 1) { - for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); + } } else { - VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - volatile VALUE t0; - long *stack = ALLOCV_N(long, t0, n+1); + VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ + volatile VALUE t0; + long *stack = ALLOCV_N(long, t0, n+1); - RBASIC_CLEAR_CLASS(ary0); - combinate0(len, n, stack, ary0); - ALLOCV_END(t0); - RBASIC_SET_CLASS_RAW(ary0, rb_cArray); + RBASIC_CLEAR_CLASS(ary0); + combinate0(len, n, stack, ary0); + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } @@ -5853,19 +7325,19 @@ rpermute0(const long n, const long r, long *const p, const VALUE values) p[index] = i; for (;;) { - if (++index < r-1) { - p[index] = i = 0; - continue; - } - for (i = 0; i < n; ++i) { - p[index] = i; - if (!yield_indexed_values(values, r, p)) { - rb_raise(rb_eRuntimeError, "repeated permute reentered"); - } - } - do { - if (index <= 0) return; - } while ((i = ++p[--index]) >= n); + if (++index < r-1) { + p[index] = i = 0; + continue; + } + for (i = 0; i < n; ++i) { + p[index] = i; + if (!yield_indexed_values(values, r, p)) { + rb_raise(rb_eRuntimeError, "repeated permute reentered"); + } + } + do { + if (index <= 0) return; + } while ((i = ++p[--index]) >= n); } } @@ -5876,37 +7348,79 @@ rb_ary_repeated_permutation_size(VALUE ary, VALUE args, VALUE eobj) long k = NUM2LONG(RARRAY_AREF(args, 0)); if (k < 0) { - return LONG2FIX(0); + return LONG2FIX(0); } if (n <= 0) { - return LONG2FIX(!k); + return LONG2FIX(!k); } return rb_int_positive_pow(n, (unsigned long)k); } /* * call-seq: - * ary.repeated_permutation(n) {|p| block} -> ary - * ary.repeated_permutation(n) -> Enumerator + * array.repeated_permutation(n) {|permutation| ... } -> self + * array.repeated_permutation(n) -> new_enumerator + * + * Calls the block with each repeated permutation of length +n+ of the elements of +self+; + * each permutation is an \Array; + * returns +self+. The order of the permutations is indeterminate. + * + * When a block and a positive \Integer argument +n+ are given, calls the block with each + * +n+-tuple repeated permutation of the elements of +self+. + * The number of permutations is <tt>self.size**n</tt>. + * + * +n+ = 1: + * + * a = [0, 1, 2] + * a.repeated_permutation(1) {|permutation| p permutation } + * + * Output: + * + * [0] + * [1] + * [2] + * + * +n+ = 2: + * + * a.repeated_permutation(2) {|permutation| p permutation } * - * When invoked with a block, yield all repeated permutations of length +n+ of - * the elements of the array, then return the array itself. + * Output: * - * The implementation makes no guarantees about the order in which the repeated - * permutations are yielded. + * [0, 0] + * [0, 1] + * [0, 2] + * [1, 0] + * [1, 1] + * [1, 2] + * [2, 0] + * [2, 1] + * [2, 2] * - * If no block is given, an Enumerator is returned instead. + * If +n+ is zero, calls the block once with an empty \Array. * - * Examples: + * If +n+ is negative, does not call the block: + * + * a.repeated_permutation(-1) {|permutation| fail 'Cannot happen' } + * + * Returns a new \Enumerator if no block given: + * + * a = [0, 1, 2] + * a.repeated_permutation(2) # => #<Enumerator: [0, 1, 2]:permutation(2)> + * + * Using Enumerators, it's convenient to show the permutations and counts + * for some values of +n+: + * + * e = a.repeated_permutation(0) + * e.size # => 1 + * e.to_a # => [[]] + * e = a.repeated_permutation(1) + * e.size # => 3 + * e.to_a # => [[0], [1], [2]] + * e = a.repeated_permutation(2) + * e.size # => 9 + * e.to_a # => [[0, 0], [0, 1], [0, 2], [1, 0], [1, 1], [1, 2], [2, 0], [2, 1], [2, 2]] * - * a = [1, 2] - * a.repeated_permutation(1).to_a #=> [[1], [2]] - * a.repeated_permutation(2).to_a #=> [[1,1],[1,2],[2,1],[2,2]] - * a.repeated_permutation(3).to_a #=> [[1,1,1],[1,1,2],[1,2,1],[1,2,2], - * # [2,1,1],[2,1,2],[2,2,1],[2,2,2]] - * a.repeated_permutation(0).to_a #=> [[]] # one permutation of length 0 */ - static VALUE rb_ary_repeated_permutation(VALUE ary, VALUE num) { @@ -5917,25 +7431,25 @@ rb_ary_repeated_permutation(VALUE ary, VALUE num) r = NUM2LONG(num); /* Permutation size from argument */ if (r < 0) { - /* no permutations: yield nothing */ + /* no permutations: yield nothing */ } else if (r == 0) { /* exactly one permutation: the zero-length array */ - rb_yield(rb_ary_new2(0)); + rb_yield(rb_ary_new2(0)); } else if (r == 1) { /* this is a special, easy case */ - for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); + } } else { /* this is the general case */ - volatile VALUE t0; - long *p = ALLOCV_N(long, t0, r); - VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - RBASIC_CLEAR_CLASS(ary0); + volatile VALUE t0; + long *p = ALLOCV_N(long, t0, r); + VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ + RBASIC_CLEAR_CLASS(ary0); - rpermute0(n, r, p, ary0); /* compute and yield repeated permutations */ - ALLOCV_END(t0); - RBASIC_SET_CLASS_RAW(ary0, rb_cArray); + rpermute0(n, r, p, ary0); /* compute and yield repeated permutations */ + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } @@ -5947,19 +7461,19 @@ rcombinate0(const long n, const long r, long *const p, const long rest, const VA p[index] = i; for (;;) { - if (++index < r-1) { - p[index] = i; - continue; - } - for (; i < n; ++i) { - p[index] = i; - if (!yield_indexed_values(values, r, p)) { - rb_raise(rb_eRuntimeError, "repeated combination reentered"); - } - } - do { - if (index <= 0) return; - } while ((i = ++p[--index]) >= n); + if (++index < r-1) { + p[index] = i; + continue; + } + for (; i < n; ++i) { + p[index] = i; + if (!yield_indexed_values(values, r, p)) { + rb_raise(rb_eRuntimeError, "repeated combination reentered"); + } + } + do { + if (index <= 0) return; + } while ((i = ++p[--index]) >= n); } } @@ -5969,35 +7483,71 @@ rb_ary_repeated_combination_size(VALUE ary, VALUE args, VALUE eobj) long n = RARRAY_LEN(ary); long k = NUM2LONG(RARRAY_AREF(args, 0)); if (k == 0) { - return LONG2FIX(1); + return LONG2FIX(1); } return binomial_coefficient(k, n + k - 1); } /* * call-seq: - * ary.repeated_combination(n) {|c| block} -> ary - * ary.repeated_combination(n) -> Enumerator + * array.repeated_combination(n) {|combination| ... } -> self + * array.repeated_combination(n) -> new_enumerator * - * When invoked with a block, yields all repeated combinations of length +n+ of - * elements from the array and then returns the array itself. + * Calls the block with each repeated combination of length +n+ of the elements of +self+; + * each combination is an \Array; + * returns +self+. The order of the combinations is indeterminate. * - * The implementation makes no guarantees about the order in which the repeated - * combinations are yielded. + * When a block and a positive \Integer argument +n+ are given, calls the block with each + * +n+-tuple repeated combination of the elements of +self+. + * The number of combinations is <tt>(n+1)(n+2)/2</tt>. * - * If no block is given, an Enumerator is returned instead. + * +n+ = 1: * - * Examples: + * a = [0, 1, 2] + * a.repeated_combination(1) {|combination| p combination } * - * a = [1, 2, 3] - * a.repeated_combination(1).to_a #=> [[1], [2], [3]] - * a.repeated_combination(2).to_a #=> [[1,1],[1,2],[1,3],[2,2],[2,3],[3,3]] - * a.repeated_combination(3).to_a #=> [[1,1,1],[1,1,2],[1,1,3],[1,2,2],[1,2,3], - * # [1,3,3],[2,2,2],[2,2,3],[2,3,3],[3,3,3]] - * a.repeated_combination(4).to_a #=> [[1,1,1,1],[1,1,1,2],[1,1,1,3],[1,1,2,2],[1,1,2,3], - * # [1,1,3,3],[1,2,2,2],[1,2,2,3],[1,2,3,3],[1,3,3,3], - * # [2,2,2,2],[2,2,2,3],[2,2,3,3],[2,3,3,3],[3,3,3,3]] - * a.repeated_combination(0).to_a #=> [[]] # one combination of length 0 + * Output: + * + * [0] + * [1] + * [2] + * + * +n+ = 2: + * + * a.repeated_combination(2) {|combination| p combination } + * + * Output: + * + * [0, 0] + * [0, 1] + * [0, 2] + * [1, 1] + * [1, 2] + * [2, 2] + * + * If +n+ is zero, calls the block once with an empty \Array. + * + * If +n+ is negative, does not call the block: + * + * a.repeated_combination(-1) {|combination| fail 'Cannot happen' } + * + * Returns a new \Enumerator if no block given: + * + * a = [0, 1, 2] + * a.repeated_combination(2) # => #<Enumerator: [0, 1, 2]:combination(2)> + * + * Using Enumerators, it's convenient to show the combinations and counts + * for some values of +n+: + * + * e = a.repeated_combination(0) + * e.size # => 1 + * e.to_a # => [[]] + * e = a.repeated_combination(1) + * e.size # => 3 + * e.to_a # => [[0], [1], [2]] + * e = a.repeated_combination(2) + * e.size # => 6 + * e.to_a # => [[0, 0], [0, 1], [0, 2], [1, 1], [1, 2], [2, 2]] * */ @@ -6010,58 +7560,97 @@ rb_ary_repeated_combination(VALUE ary, VALUE num) RETURN_SIZED_ENUMERATOR(ary, 1, &num, rb_ary_repeated_combination_size); /* Return enumerator if no block */ len = RARRAY_LEN(ary); if (n < 0) { - /* yield nothing */ + /* yield nothing */ } else if (n == 0) { - rb_yield(rb_ary_new2(0)); + rb_yield(rb_ary_new2(0)); } else if (n == 1) { - for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); - } + for (i = 0; i < RARRAY_LEN(ary); i++) { + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); + } } else if (len == 0) { - /* yield nothing */ + /* yield nothing */ } else { - volatile VALUE t0; - long *p = ALLOCV_N(long, t0, n); - VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - RBASIC_CLEAR_CLASS(ary0); + volatile VALUE t0; + long *p = ALLOCV_N(long, t0, n); + VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ + RBASIC_CLEAR_CLASS(ary0); - rcombinate0(len, n, p, n, ary0); /* compute and yield repeated combinations */ - ALLOCV_END(t0); - RBASIC_SET_CLASS_RAW(ary0, rb_cArray); + rcombinate0(len, n, p, n, ary0); /* compute and yield repeated combinations */ + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } /* * call-seq: - * ary.product(other_ary, ...) -> new_ary - * ary.product(other_ary, ...) {|p| block} -> ary + * array.product(*other_arrays) -> new_array + * array.product(*other_arrays) {|combination| ... } -> self + * + * Computes and returns or yields all combinations of elements from all the Arrays, + * including both +self+ and +other_arrays+: + * + * - The number of combinations is the product of the sizes of all the arrays, + * including both +self+ and +other_arrays+. + * - The order of the returned combinations is indeterminate. + * + * When no block is given, returns the combinations as an \Array of Arrays: + * + * a = [0, 1, 2] + * a1 = [3, 4] + * a2 = [5, 6] + * p = a.product(a1) + * p.size # => 6 # a.size * a1.size + * p # => [[0, 3], [0, 4], [1, 3], [1, 4], [2, 3], [2, 4]] + * p = a.product(a1, a2) + * p.size # => 12 # a.size * a1.size * a2.size + * p # => [[0, 3, 5], [0, 3, 6], [0, 4, 5], [0, 4, 6], [1, 3, 5], [1, 3, 6], [1, 4, 5], [1, 4, 6], [2, 3, 5], [2, 3, 6], [2, 4, 5], [2, 4, 6]] + * + * If any argument is an empty \Array, returns an empty \Array. + * + * If no argument is given, returns an \Array of 1-element Arrays, + * each containing an element of +self+: + * + * a.product # => [[0], [1], [2]] + * + * When a block is given, yields each combination as an \Array; returns +self+: * - * Returns an array of all combinations of elements from all arrays. + * a.product(a1) {|combination| p combination } * - * The length of the returned array is the product of the length of +self+ and - * the argument arrays. + * Output: * - * If given a block, #product will yield all combinations and return +self+ - * instead. + * [0, 3] + * [0, 4] + * [1, 3] + * [1, 4] + * [2, 3] + * [2, 4] + * + * If any argument is an empty \Array, does not call the block: + * + * a.product(a1, a2, []) {|combination| fail 'Cannot happen' } + * + * If no argument is given, yields each element of +self+ as a 1-element \Array: + * + * a.product {|combination| p combination } + * + * Output: + * + * [0] + * [1] + * [2] * - * [1,2,3].product([4,5]) #=> [[1,4],[1,5],[2,4],[2,5],[3,4],[3,5]] - * [1,2].product([1,2]) #=> [[1,1],[1,2],[2,1],[2,2]] - * [1,2].product([3,4],[5,6]) #=> [[1,3,5],[1,3,6],[1,4,5],[1,4,6], - * # [2,3,5],[2,3,6],[2,4,5],[2,4,6]] - * [1,2].product() #=> [[1],[2]] - * [1,2].product([]) #=> [] */ static VALUE rb_ary_product(int argc, VALUE *argv, VALUE ary) { int n = argc+1; /* How many arrays we're operating on */ - volatile VALUE t0 = tmpary(n); + volatile VALUE t0 = rb_ary_hidden_new(n); volatile VALUE t1 = Qundef; VALUE *arrays = RARRAY_PTR(t0); /* The arrays we're computing the product of */ int *counters = ALLOCV_N(int, t1, n); /* The current position in each one */ @@ -6082,64 +7671,64 @@ rb_ary_product(int argc, VALUE *argv, VALUE ary) /* Otherwise, allocate and fill in an array of results */ if (rb_block_given_p()) { - /* Make defensive copies of arrays; exit if any is empty */ - for (i = 0; i < n; i++) { - if (RARRAY_LEN(arrays[i]) == 0) goto done; - arrays[i] = ary_make_shared_copy(arrays[i]); - } + /* Make defensive copies of arrays; exit if any is empty */ + for (i = 0; i < n; i++) { + if (RARRAY_LEN(arrays[i]) == 0) goto done; + arrays[i] = ary_make_shared_copy(arrays[i]); + } } else { - /* Compute the length of the result array; return [] if any is empty */ - for (i = 0; i < n; i++) { - long k = RARRAY_LEN(arrays[i]); - if (k == 0) { - result = rb_ary_new2(0); - goto done; - } + /* Compute the length of the result array; return [] if any is empty */ + for (i = 0; i < n; i++) { + long k = RARRAY_LEN(arrays[i]); + if (k == 0) { + result = rb_ary_new2(0); + goto done; + } if (MUL_OVERFLOW_LONG_P(resultlen, k)) - rb_raise(rb_eRangeError, "too big to product"); - resultlen *= k; - } - result = rb_ary_new2(resultlen); + rb_raise(rb_eRangeError, "too big to product"); + resultlen *= k; + } + result = rb_ary_new2(resultlen); } for (;;) { - int m; - /* fill in one subarray */ - VALUE subarray = rb_ary_new2(n); - for (j = 0; j < n; j++) { - rb_ary_push(subarray, rb_ary_entry(arrays[j], counters[j])); - } - - /* put it on the result array */ - if (NIL_P(result)) { - FL_SET(t0, FL_USER5); - rb_yield(subarray); - if (! FL_TEST(t0, FL_USER5)) { - rb_raise(rb_eRuntimeError, "product reentered"); - } - else { - FL_UNSET(t0, FL_USER5); - } - } - else { - rb_ary_push(result, subarray); - } - - /* - * Increment the last counter. If it overflows, reset to 0 - * and increment the one before it. - */ - m = n-1; - counters[m]++; - while (counters[m] == RARRAY_LEN(arrays[m])) { - counters[m] = 0; - /* If the first counter overflows, we are done */ - if (--m < 0) goto done; - counters[m]++; - } + int m; + /* fill in one subarray */ + VALUE subarray = rb_ary_new2(n); + for (j = 0; j < n; j++) { + rb_ary_push(subarray, rb_ary_entry(arrays[j], counters[j])); + } + + /* put it on the result array */ + if (NIL_P(result)) { + FL_SET(t0, RARRAY_SHARED_ROOT_FLAG); + rb_yield(subarray); + if (!FL_TEST(t0, RARRAY_SHARED_ROOT_FLAG)) { + rb_raise(rb_eRuntimeError, "product reentered"); + } + else { + FL_UNSET(t0, RARRAY_SHARED_ROOT_FLAG); + } + } + else { + rb_ary_push(result, subarray); + } + + /* + * Increment the last counter. If it overflows, reset to 0 + * and increment the one before it. + */ + m = n-1; + counters[m]++; + while (counters[m] == RARRAY_LEN(arrays[m])) { + counters[m] = 0; + /* If the first counter overflows, we are done */ + if (--m < 0) goto done; + counters[m]++; + } } + done: - tmpary_discard(t0); ALLOCV_END(t1); return NIL_P(result) ? ary : result; @@ -6147,16 +7736,19 @@ done: /* * call-seq: - * ary.take(n) -> new_ary - * - * Returns first +n+ elements from the array. + * array.take(n) -> new_array * - * If a negative number is given, raises an ArgumentError. + * Returns a new \Array containing the first +n+ element of +self+, + * where +n+ is a non-negative \Integer; + * does not modify +self+. * - * See also Array#drop + * Examples: * - * a = [1, 2, 3, 4, 5, 0] - * a.take(3) #=> [1, 2, 3] + * a = [0, 1, 2, 3, 4, 5] + * a.take(1) # => [0] + * a.take(2) # => [0, 1] + * a.take(50) # => [0, 1, 2, 3, 4, 5] + * a # => [0, 1, 2, 3, 4, 5] * */ @@ -6165,25 +7757,31 @@ rb_ary_take(VALUE obj, VALUE n) { long len = NUM2LONG(n); if (len < 0) { - rb_raise(rb_eArgError, "attempt to take negative size"); + rb_raise(rb_eArgError, "attempt to take negative size"); } return rb_ary_subseq(obj, 0, len); } /* * call-seq: - * ary.take_while {|obj| block} -> new_ary - * ary.take_while -> Enumerator + * array.take_while {|element| ... } -> new_array + * array.take_while -> new_enumerator + * + * Returns a new \Array containing zero or more leading elements of +self+; + * does not modify +self+. * - * Passes elements to the block until the block returns +nil+ or +false+, then - * stops iterating and returns an array of all prior elements. + * With a block given, calls the block with each successive element of +self+; + * stops if the block returns +false+ or +nil+; + * returns a new \Array containing those elements for which the block returned a truthy value: * - * If no block is given, an Enumerator is returned instead. + * a = [0, 1, 2, 3, 4, 5] + * a.take_while {|element| element < 3 } # => [0, 1, 2] + * a.take_while {|element| true } # => [0, 1, 2, 3, 4, 5] + * a # => [0, 1, 2, 3, 4, 5] * - * See also Array#drop_while + * With no block given, returns a new \Enumerator: * - * a = [1, 2, 3, 4, 5, 0] - * a.take_while {|i| i < 3} #=> [1, 2] + * [0, 1].take_while # => #<Enumerator: [0, 1]:take_while> * */ @@ -6194,24 +7792,25 @@ rb_ary_take_while(VALUE ary) RETURN_ENUMERATOR(ary, 0, 0); for (i = 0; i < RARRAY_LEN(ary); i++) { - if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) break; + if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) break; } return rb_ary_take(ary, LONG2FIX(i)); } /* * call-seq: - * ary.drop(n) -> new_ary + * array.drop(n) -> new_array * - * Drops first +n+ elements from +ary+ and returns the rest of the elements in - * an array. + * Returns a new \Array containing all but the first +n+ element of +self+, + * where +n+ is a non-negative \Integer; + * does not modify +self+. * - * If a negative number is given, raises an ArgumentError. + * Examples: * - * See also Array#take - * - * a = [1, 2, 3, 4, 5, 0] - * a.drop(3) #=> [4, 5, 0] + * a = [0, 1, 2, 3, 4, 5] + * a.drop(0) # => [0, 1, 2, 3, 4, 5] + * a.drop(1) # => [1, 2, 3, 4, 5] + * a.drop(2) # => [2, 3, 4, 5] * */ @@ -6221,29 +7820,32 @@ rb_ary_drop(VALUE ary, VALUE n) VALUE result; long pos = NUM2LONG(n); if (pos < 0) { - rb_raise(rb_eArgError, "attempt to drop negative size"); + rb_raise(rb_eArgError, "attempt to drop negative size"); } result = rb_ary_subseq(ary, pos, RARRAY_LEN(ary)); - if (result == Qnil) result = rb_ary_new(); + if (NIL_P(result)) result = rb_ary_new(); return result; } /* * call-seq: - * ary.drop_while {|obj| block} -> new_ary - * ary.drop_while -> Enumerator + * array.drop_while {|element| ... } -> new_array + * array.drop_while -> new_enumerator + + * Returns a new \Array containing zero or more trailing elements of +self+; + * does not modify +self+. * - * Drops elements up to, but not including, the first element for which the - * block returns +nil+ or +false+ and returns an array containing the - * remaining elements. + * With a block given, calls the block with each successive element of +self+; + * stops if the block returns +false+ or +nil+; + * returns a new \Array _omitting_ those elements for which the block returned a truthy value: * - * If no block is given, an Enumerator is returned instead. + * a = [0, 1, 2, 3, 4, 5] + * a.drop_while {|element| element < 3 } # => [3, 4, 5] * - * See also Array#take_while + * With no block given, returns a new \Enumerator: * - * a = [1, 2, 3, 4, 5, 0] - * a.drop_while {|i| i < 3 } #=> [3, 4, 5, 0] + * [0, 1].drop_while # => # => #<Enumerator: [0, 1]:drop_while> * */ @@ -6254,17 +7856,42 @@ rb_ary_drop_while(VALUE ary) RETURN_ENUMERATOR(ary, 0, 0); for (i = 0; i < RARRAY_LEN(ary); i++) { - if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) break; + if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) break; } return rb_ary_drop(ary, LONG2FIX(i)); } /* * call-seq: - * ary.any? [{|obj| block} ] -> true or false - * ary.any?(pattern) -> true or false + * array.any? -> true or false + * array.any? {|element| ... } -> true or false + * array.any?(obj) -> true or false + * + * Returns +true+ if any element of +self+ meets a given criterion. + * + * With no block given and no argument, returns +true+ if +self+ has any truthy element, + * +false+ otherwise: + * + * [nil, 0, false].any? # => true + * [nil, false].any? # => false + * [].any? # => false + * + * With a block given and no argument, calls the block with each element in +self+; + * returns +true+ if the block returns any truthy value, +false+ otherwise: + * + * [0, 1, 2].any? {|element| element > 1 } # => true + * [0, 1, 2].any? {|element| element > 2 } # => false + * + * If argument +obj+ is given, returns +true+ if +obj+.<tt>===</tt> any element, + * +false+ otherwise: * - * See also Enumerable#any? + * ['food', 'drink'].any?(/foo/) # => true + * ['food', 'drink'].any?(/bar/) # => false + * [].any?(/foo/) # => false + * [0, 1, 2].any?(1) # => true + * [0, 1, 2].any?(3) # => false + * + * Related: Enumerable#any? */ static VALUE @@ -6278,9 +7905,9 @@ rb_ary_any_p(int argc, VALUE *argv, VALUE ary) if (rb_block_given_p()) { rb_warn("given block not used"); } - for (i = 0; i < RARRAY_LEN(ary); ++i) { - if (RTEST(rb_funcall(argv[0], idEqq, 1, RARRAY_AREF(ary, i)))) return Qtrue; - } + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (RTEST(rb_funcall(argv[0], idEqq, 1, RARRAY_AREF(ary, i)))) return Qtrue; + } } else if (!rb_block_given_p()) { for (i = 0; i < len; ++i) { @@ -6288,19 +7915,43 @@ rb_ary_any_p(int argc, VALUE *argv, VALUE ary) } } else { - for (i = 0; i < RARRAY_LEN(ary); ++i) { - if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) return Qtrue; - } + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) return Qtrue; + } } return Qfalse; } /* * call-seq: - * ary.all? [{|obj| block} ] -> true or false - * ary.all?(pattern) -> true or false + * array.all? -> true or false + * array.all? {|element| ... } -> true or false + * array.all?(obj) -> true or false + * + * Returns +true+ if all elements of +self+ meet a given criterion. + * + * With no block given and no argument, returns +true+ if +self+ contains only truthy elements, + * +false+ otherwise: + * + * [0, 1, :foo].all? # => true + * [0, nil, 2].all? # => false + * [].all? # => true + * + * With a block given and no argument, calls the block with each element in +self+; + * returns +true+ if the block returns only truthy values, +false+ otherwise: + * + * [0, 1, 2].all? { |element| element < 3 } # => true + * [0, 1, 2].all? { |element| element < 2 } # => false * - * See also Enumerable#all? + * If argument +obj+ is given, returns +true+ if <tt>obj.===</tt> every element, +false+ otherwise: + * + * ['food', 'fool', 'foot'].all?(/foo/) # => true + * ['food', 'drink'].all?(/bar/) # => false + * [].all?(/foo/) # => true + * [0, 0, 0].all?(0) # => true + * [0, 1, 2].all?(1) # => false + * + * Related: Enumerable#all? */ static VALUE @@ -6333,10 +7984,34 @@ rb_ary_all_p(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.none? [{|obj| block} ] -> true or false - * ary.none?(pattern) -> true or false + * array.none? -> true or false + * array.none? {|element| ... } -> true or false + * array.none?(obj) -> true or false + * + * Returns +true+ if no element of +self+ meet a given criterion. + * + * With no block given and no argument, returns +true+ if +self+ has no truthy elements, + * +false+ otherwise: + * + * [nil, false].none? # => true + * [nil, 0, false].none? # => false + * [].none? # => true + * + * With a block given and no argument, calls the block with each element in +self+; + * returns +true+ if the block returns no truthy value, +false+ otherwise: * - * See also Enumerable#none? + * [0, 1, 2].none? {|element| element > 3 } # => true + * [0, 1, 2].none? {|element| element > 1 } # => false + * + * If argument +obj+ is given, returns +true+ if <tt>obj.===</tt> no element, +false+ otherwise: + * + * ['food', 'drink'].none?(/bar/) # => true + * ['food', 'drink'].none?(/foo/) # => false + * [].none?(/foo/) # => true + * [0, 1, 2].none?(3) # => true + * [0, 1, 2].none?(1) # => false + * + * Related: Enumerable#none? */ static VALUE @@ -6369,10 +8044,38 @@ rb_ary_none_p(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.one? [{|obj| block} ] -> true or false - * ary.one?(pattern) -> true or false + * array.one? -> true or false + * array.one? {|element| ... } -> true or false + * array.one?(obj) -> true or false + * + * Returns +true+ if exactly one element of +self+ meets a given criterion. + * + * With no block given and no argument, returns +true+ if +self+ has exactly one truthy element, + * +false+ otherwise: * - * See also Enumerable#one? + * [nil, 0].one? # => true + * [0, 0].one? # => false + * [nil, nil].one? # => false + * [].one? # => false + * + * With a block given and no argument, calls the block with each element in +self+; + * returns +true+ if the block a truthy value for exactly one element, +false+ otherwise: + * + * [0, 1, 2].one? {|element| element > 0 } # => false + * [0, 1, 2].one? {|element| element > 1 } # => true + * [0, 1, 2].one? {|element| element > 2 } # => false + * + * If argument +obj+ is given, returns +true+ if <tt>obj.===</tt> exactly one element, + * +false+ otherwise: + * + * [0, 1, 2].one?(0) # => true + * [0, 0, 1].one?(0) # => false + * [1, 1, 2].one?(0) # => false + * ['food', 'drink'].one?(/bar/) # => false + * ['food', 'drink'].one?(/foo/) # => true + * [].one?(/foo/) # => false + * + * Related: Enumerable#one? */ static VALUE @@ -6414,19 +8117,22 @@ rb_ary_one_p(int argc, VALUE *argv, VALUE ary) } /* - * call-seq: - * ary.dig(idx, ...) -> object + * call-seq: + * array.dig(index, *identifiers) -> object + * + * Finds and returns the object in nested objects + * that is specified by +index+ and +identifiers+. + * The nested objects may be instances of various classes. + * See {Dig Methods}[rdoc-ref:dig_methods.rdoc]. * - * Extracts the nested value specified by the sequence of <i>idx</i> - * objects by calling +dig+ at each step, returning +nil+ if any - * intermediate step is +nil+. + * Examples: * - * a = [[1, [2, 3]]] + * a = [:foo, [:bar, :baz, [:bat, :bam]]] + * a.dig(1) # => [:bar, :baz, [:bat, :bam]] + * a.dig(1, 2) # => [:bat, :bam] + * a.dig(1, 2, 0) # => :bat + * a.dig(1, 2, 3) # => nil * - * a.dig(0, 1, 1) #=> 3 - * a.dig(1, 2, 3) #=> nil - * a.dig(0, 0, 0) #=> TypeError: Integer does not have #dig method - * [42, {foo: :bar}].dig(1, :foo) #=> :bar */ static VALUE @@ -6444,14 +8150,8 @@ finish_exact_sum(long n, VALUE r, VALUE v, int z) { if (n != 0) v = rb_fix_plus(LONG2FIX(n), v); - if (r != Qundef) { - /* r can be an Integer when mathn is loaded */ - if (FIXNUM_P(r)) - v = rb_fix_plus(r, v); - else if (RB_TYPE_P(r, T_BIGNUM)) - v = rb_big_plus(r, v); - else - v = rb_rational_plus(r, v); + if (!UNDEF_P(r)) { + v = rb_rational_plus(r, v); } else if (!n && z) { v = rb_fix_plus(LONG2FIX(0), v); @@ -6461,43 +8161,41 @@ finish_exact_sum(long n, VALUE r, VALUE v, int z) /* * call-seq: - * ary.sum(init=0) -> number - * ary.sum(init=0) {|e| expr } -> number + * array.sum(init = 0) -> object + * array.sum(init = 0) {|element| ... } -> object * - * Returns the sum of elements. - * For example, [e1, e2, e3].sum returns init + e1 + e2 + e3. + * When no block is given, returns the object equivalent to: * - * If a block is given, the block is applied to each element - * before addition. + * sum = init + * array.each {|element| sum += element } + * sum * - * If <i>ary</i> is empty, it returns <i>init</i>. + * For example, <tt>[e1, e2, e3].sum</tt> returns <tt>init + e1 + e2 + e3</tt>. * - * [].sum #=> 0 - * [].sum(0.0) #=> 0.0 - * [1, 2, 3].sum #=> 6 - * [3, 5.5].sum #=> 8.5 - * [2.5, 3.0].sum(0.0) {|e| e * e } #=> 15.25 - * [Object.new].sum #=> TypeError + * Examples: * - * The (arithmetic) mean value of an array can be obtained as follows. + * a = [0, 1, 2, 3] + * a.sum # => 6 + * a.sum(100) # => 106 * - * mean = ary.sum(0.0) / ary.length + * The elements need not be numeric, but must be <tt>+</tt>-compatible + * with each other and with +init+: * - * This method can be used for non-numeric objects by - * explicit <i>init</i> argument. + * a = ['abc', 'def', 'ghi'] + * a.sum('jkl') # => "jklabcdefghi" * - * ["a", "b", "c"].sum("") #=> "abc" - * [[1], [[2]], [3]].sum([]) #=> [1, [2], 3] + * When a block is given, it is called with each element + * and the block's return value (instead of the element itself) is used as the addend: * - * However, Array#join and Array#flatten is faster than Array#sum for - * array of strings and array of arrays. + * a = ['zero', 1, :two] + * s = a.sum('Coerced and concatenated: ') {|element| element.to_s } + * s # => "Coerced and concatenated: zero1two" * - * ["a", "b", "c"].join #=> "abc" - * [[1], [[2]], [3]].flatten(1) #=> [1, [2], 3] + * Notes: * - * - * Array#sum method may not respect method redefinition of "+" methods - * such as Integer#+. + * - Array#join and Array#flatten may be faster than Array#sum + * for an \Array of Strings or an \Array of Arrays. + * - Array#sum method may not respect method redefinition of "+" methods such as Integer#+. * */ @@ -6528,10 +8226,10 @@ rb_ary_sum(int argc, VALUE *argv, VALUE ary) n = 0; } } - else if (RB_TYPE_P(e, T_BIGNUM)) + else if (RB_BIGNUM_TYPE_P(e)) v = rb_big_plus(e, v); else if (RB_TYPE_P(e, T_RATIONAL)) { - if (r == Qundef) + if (UNDEF_P(r)) r = e; else r = rb_rational_plus(r, e); @@ -6548,15 +8246,15 @@ rb_ary_sum(int argc, VALUE *argv, VALUE ary) if (RB_FLOAT_TYPE_P(e)) { /* * Kahan-Babuska balancing compensated summation algorithm - * See http://link.springer.com/article/10.1007/s00607-005-0139-x + * See https://link.springer.com/article/10.1007/s00607-005-0139-x */ double f, c; + double x, t; f = NUM2DBL(v); c = 0.0; goto has_float_value; for (; i < RARRAY_LEN(ary); i++) { - double x, t; e = RARRAY_AREF(ary, i); if (block_given) e = rb_yield(e); @@ -6565,7 +8263,7 @@ rb_ary_sum(int argc, VALUE *argv, VALUE ary) x = RFLOAT_VALUE(e); else if (FIXNUM_P(e)) x = FIX2LONG(e); - else if (RB_TYPE_P(e, T_BIGNUM)) + else if (RB_BIGNUM_TYPE_P(e)) x = rb_big2dbl(e); else if (RB_TYPE_P(e, T_RATIONAL)) x = rb_num2dbl(e); @@ -6618,55 +8316,128 @@ rb_ary_deconstruct(VALUE ary) } /* - * Arrays are ordered, integer-indexed collections of any object. + * An \Array is an ordered, integer-indexed collection of objects, called _elements_. + * Any object (even another array) may be an array element, + * and an array can contain objects of different types. + * + * == \Array Indexes + * + * \Array indexing starts at 0, as in C or Java. + * + * A positive index is an offset from the first element: + * + * - Index 0 indicates the first element. + * - Index 1 indicates the second element. + * - ... + * + * A negative index is an offset, backwards, from the end of the array: + * + * - Index -1 indicates the last element. + * - Index -2 indicates the next-to-last element. + * - ... + * + * A non-negative index is <i>in range</i> if and only if it is smaller than + * the size of the array. For a 3-element array: + * + * - Indexes 0 through 2 are in range. + * - Index 3 is out of range. + * + * A negative index is <i>in range</i> if and only if its absolute value is + * not larger than the size of the array. For a 3-element array: + * + * - Indexes -1 through -3 are in range. + * - Index -4 is out of range. + * + * Although the effective index into an array is always an integer, + * some methods (both within and outside of class \Array) + * accept one or more non-integer arguments that are + * {integer-convertible objects}[rdoc-ref:implicit_conversion.rdoc@Integer-Convertible+Objects]. * - * Array indexing starts at 0, as in C or Java. A negative index is assumed - * to be relative to the end of the array---that is, an index of -1 indicates - * the last element of the array, -2 is the next to last element in the - * array, and so on. * * == Creating Arrays * - * A new array can be created by using the literal constructor - * <code>[]</code>. Arrays can contain different types of objects. For - * example, the array below contains an Integer, a String and a Float: + * You can create an \Array object explicitly with: + * + * - An {array literal}[rdoc-ref:literals.rdoc@Array+Literals]: * - * ary = [1, "two", 3.0] #=> [1, "two", 3.0] + * [1, 'one', :one, [2, 'two', :two]] * - * An array can also be created by explicitly calling Array.new with zero, one - * (the initial size of the Array) or two arguments (the initial size and a - * default object). + * - A {%w or %W: string-array Literal}[rdoc-ref:literals.rdoc@25w+and+-25W-3A+String-Array+Literals]: * - * ary = Array.new #=> [] - * Array.new(3) #=> [nil, nil, nil] - * Array.new(3, true) #=> [true, true, true] + * %w[foo bar baz] # => ["foo", "bar", "baz"] + * %w[1 % *] # => ["1", "%", "*"] * - * Note that the second argument populates the array with references to the - * same object. Therefore, it is only recommended in cases when you need to - * instantiate arrays with natively immutable objects such as Symbols, - * numbers, true or false. + * - A {%i pr %I: symbol-array Literal}[rdoc-ref:literals.rdoc@25i+and+-25I-3A+Symbol-Array+Literals]: * - * To create an array with separate objects a block can be passed instead. - * This method is safe to use with mutable objects such as hashes, strings or - * other arrays: + * %i[foo bar baz] # => [:foo, :bar, :baz] + * %i[1 % *] # => [:"1", :%, :*] * - * Array.new(4) {Hash.new} #=> [{}, {}, {}, {}] - * Array.new(4) {|i| i.to_s } #=> ["0", "1", "2", "3"] + * - \Method Kernel#Array: * - * This is also a quick way to build up multi-dimensional arrays: + * Array(["a", "b"]) # => ["a", "b"] + * Array(1..5) # => [1, 2, 3, 4, 5] + * Array(key: :value) # => [[:key, :value]] + * Array(nil) # => [] + * Array(1) # => [1] + * Array({:a => "a", :b => "b"}) # => [[:a, "a"], [:b, "b"]] * - * empty_table = Array.new(3) {Array.new(3)} - * #=> [[nil, nil, nil], [nil, nil, nil], [nil, nil, nil]] + * - \Method Array.new: * - * An array can also be created by using the Array() method, provided by - * Kernel, which tries to call #to_ary, then #to_a on its argument. + * Array.new # => [] + * Array.new(3) # => [nil, nil, nil] + * Array.new(4) {Hash.new} # => [{}, {}, {}, {}] + * Array.new(3, true) # => [true, true, true] * - * Array({:a => "a", :b => "b"}) #=> [[:a, "a"], [:b, "b"]] + * Note that the last example above populates the array + * with references to the same object. + * This is recommended only in cases where that object is a natively immutable object + * such as a symbol, a numeric, +nil+, +true+, or +false+. + * + * Another way to create an array with various objects, using a block; + * this usage is safe for mutable objects such as hashes, strings or + * other arrays: + * + * Array.new(4) {|i| i.to_s } # => ["0", "1", "2", "3"] + * + * Here is a way to create a multi-dimensional array: + * + * Array.new(3) {Array.new(3)} + * # => [[nil, nil, nil], [nil, nil, nil], [nil, nil, nil]] + * + * A number of Ruby methods, both in the core and in the standard library, + * provide instance method +to_a+, which converts an object to an array. + * + * - ARGF#to_a + * - Array#to_a + * - Enumerable#to_a + * - Hash#to_a + * - MatchData#to_a + * - NilClass#to_a + * - OptionParser#to_a + * - Range#to_a + * - Set#to_a + * - Struct#to_a + * - Time#to_a + * - Benchmark::Tms#to_a + * - CSV::Table#to_a + * - Enumerator::Lazy#to_a + * - Gem::List#to_a + * - Gem::NameTuple#to_a + * - Gem::Platform#to_a + * - Gem::RequestSet::Lockfile::Tokenizer#to_a + * - Gem::SourceList#to_a + * - OpenSSL::X509::Extension#to_a + * - OpenSSL::X509::Name#to_a + * - Racc::ISet#to_a + * - Rinda::RingFinger#to_a + * - Ripper::Lexer::Elem#to_a + * - RubyVM::InstructionSequence#to_a + * - YAML::DBM#to_a * * == Example Usage * * In addition to the methods it mixes in through the Enumerable module, the - * Array class has proprietary methods for accessing, searching and otherwise + * \Array class has proprietary methods for accessing, searching and otherwise * manipulating arrays. * * Some of the more common ones are illustrated below. @@ -6714,7 +8485,7 @@ rb_ary_deconstruct(VALUE ary) * * arr.drop(3) #=> [4, 5, 6] * - * == Obtaining Information about an Array + * == Obtaining Information about an \Array * * Arrays keep track of their own length at all times. To query an array * about the number of elements it contains, use #length, #count or #size. @@ -6752,7 +8523,7 @@ rb_ary_deconstruct(VALUE ary) * arr.insert(3, 'orange', 'pear', 'grapefruit') * #=> [0, 1, 2, "orange", "pear", "grapefruit", "apple", 3, 4, 5, 6] * - * == Removing Items from an Array + * == Removing Items from an \Array * * The method #pop removes the last element in an array and returns it: * @@ -6794,9 +8565,9 @@ rb_ary_deconstruct(VALUE ary) * * == Iterating over Arrays * - * Like all classes that include the Enumerable module, Array has an each + * Like all classes that include the Enumerable module, \Array has an each * method, which defines what elements should be iterated over and how. In - * case of Array's #each, all elements in the Array instance are yielded to + * case of Array's #each, all elements in the \Array instance are yielded to * the supplied block in sequence. * * Note that this operation leaves the array unchanged. @@ -6822,7 +8593,8 @@ rb_ary_deconstruct(VALUE ary) * arr.map! {|a| a**2} #=> [1, 4, 9, 16, 25] * arr #=> [1, 4, 9, 16, 25] * - * == Selecting Items from an Array + * + * == Selecting Items from an \Array * * Elements can be selected from an array according to criteria defined in a * block. The selection can happen in a destructive or a non-destructive @@ -6853,18 +8625,198 @@ rb_ary_deconstruct(VALUE ary) * arr.keep_if {|a| a < 4} #=> [1, 2, 3] * arr #=> [1, 2, 3] * + * == What's Here + * + * First, what's elsewhere. \Class \Array: + * + * - Inherits from {class Object}[rdoc-ref:Object@What-27s+Here]. + * - Includes {module Enumerable}[rdoc-ref:Enumerable@What-27s+Here], + * which provides dozens of additional methods. + * + * Here, class \Array provides methods that are useful for: + * + * - {Creating an Array}[rdoc-ref:Array@Methods+for+Creating+an+Array] + * - {Querying}[rdoc-ref:Array@Methods+for+Querying] + * - {Comparing}[rdoc-ref:Array@Methods+for+Comparing] + * - {Fetching}[rdoc-ref:Array@Methods+for+Fetching] + * - {Assigning}[rdoc-ref:Array@Methods+for+Assigning] + * - {Deleting}[rdoc-ref:Array@Methods+for+Deleting] + * - {Combining}[rdoc-ref:Array@Methods+for+Combining] + * - {Iterating}[rdoc-ref:Array@Methods+for+Iterating] + * - {Converting}[rdoc-ref:Array@Methods+for+Converting] + * - {And more....}[rdoc-ref:Array@Other+Methods] + * + * === Methods for Creating an \Array + * + * - ::[]: Returns a new array populated with given objects. + * - ::new: Returns a new array. + * - ::try_convert: Returns a new array created from a given object. + * + * === Methods for Querying + * + * - #length, #size: Returns the count of elements. + * - #include?: Returns whether any element <tt>==</tt> a given object. + * - #empty?: Returns whether there are no elements. + * - #all?: Returns whether all elements meet a given criterion. + * - #any?: Returns whether any element meets a given criterion. + * - #none?: Returns whether no element <tt>==</tt> a given object. + * - #one?: Returns whether exactly one element <tt>==</tt> a given object. + * - #count: Returns the count of elements that meet a given criterion. + * - #find_index, #index: Returns the index of the first element that meets a given criterion. + * - #rindex: Returns the index of the last element that meets a given criterion. + * - #hash: Returns the integer hash code. + * + * === Methods for Comparing + * + * - #<=>: Returns -1, 0, or 1 * as +self+ is less than, equal to, or + * greater than a given object. + * - #==: Returns whether each element in +self+ is <tt>==</tt> to the corresponding element + * in a given object. + * - #eql?: Returns whether each element in +self+ is <tt>eql?</tt> to the corresponding + * element in a given object. + + * === Methods for Fetching + * + * These methods do not modify +self+. + * + * - #[]: Returns one or more elements. + * - #fetch: Returns the element at a given offset. + * - #first: Returns one or more leading elements. + * - #last: Returns one or more trailing elements. + * - #max: Returns one or more maximum-valued elements, + * as determined by <tt><=></tt> or a given block. + * - #min: Returns one or more minimum-valued elements, + * as determined by <tt><=></tt> or a given block. + * - #minmax: Returns the minimum-valued and maximum-valued elements, + * as determined by <tt><=></tt> or a given block. + * - #assoc: Returns the first element that is an array + * whose first element <tt>==</tt> a given object. + * - #rassoc: Returns the first element that is an array + * whose second element <tt>==</tt> a given object. + * - #at: Returns the element at a given offset. + * - #values_at: Returns the elements at given offsets. + * - #dig: Returns the object in nested objects + * that is specified by a given index and additional arguments. + * - #drop: Returns trailing elements as determined by a given index. + * - #take: Returns leading elements as determined by a given index. + * - #drop_while: Returns trailing elements as determined by a given block. + * - #take_while: Returns leading elements as determined by a given block. + * - #slice: Returns consecutive elements as determined by a given argument. + * - #sort: Returns all elements in an order determined by <tt><=></tt> or a given block. + * - #reverse: Returns all elements in reverse order. + * - #compact: Returns an array containing all non-+nil+ elements. + * - #select, #filter: Returns an array containing elements selected by a given block. + * - #uniq: Returns an array containing non-duplicate elements. + * - #rotate: Returns all elements with some rotated from one end to the other. + * - #bsearch: Returns an element selected via a binary search + * as determined by a given block. + * - #bsearch_index: Returns the index of an element selected via a binary search + * as determined by a given block. + * - #sample: Returns one or more random elements. + * - #shuffle: Returns elements in a random order. + * + * === Methods for Assigning + * + * These methods add, replace, or reorder elements in +self+. + * + * - #[]=: Assigns specified elements with a given object. + * - #push, #append, #<<: Appends trailing elements. + * - #unshift, #prepend: Prepends leading elements. + * - #insert: Inserts given objects at a given offset; does not replace elements. + * - #concat: Appends all elements from given arrays. + * - #fill: Replaces specified elements with specified objects. + * - #replace: Replaces the content of +self+ with the content of a given array. + * - #reverse!: Replaces +self+ with its elements reversed. + * - #rotate!: Replaces +self+ with its elements rotated. + * - #shuffle!: Replaces +self+ with its elements in random order. + * - #sort!: Replaces +self+ with its elements sorted, + * as determined by <tt><=></tt> or a given block. + * - #sort_by!: Replaces +self+ with its elements sorted, as determined by a given block. + * + * === Methods for Deleting + * + * Each of these methods removes elements from +self+: + * + * - #pop: Removes and returns the last element. + * - #shift: Removes and returns the first element. + * - #compact!: Removes all +nil+ elements. + * - #delete: Removes elements equal to a given object. + * - #delete_at: Removes the element at a given offset. + * - #delete_if: Removes elements specified by a given block. + * - #keep_if: Removes elements not specified by a given block. + * - #reject!: Removes elements specified by a given block. + * - #select!, #filter!: Removes elements not specified by a given block. + * - #slice!: Removes and returns a sequence of elements. + * - #uniq!: Removes duplicates. + * + * === Methods for Combining + * + * - #&: Returns an array containing elements found both in +self+ and a given array. + * - #intersection: Returns an array containing elements found both in +self+ + * and in each given array. + * - #+: Returns an array containing all elements of +self+ followed by all elements of a given array. + * - #-: Returns an array containing all elements of +self+ that are not found in a given array. + * - #|: Returns an array containing all elements of +self+ and all elements of a given array, + * duplicates removed. + * - #union: Returns an array containing all elements of +self+ and all elements of given arrays, + * duplicates removed. + * - #difference: Returns an array containing all elements of +self+ that are not found + * in any of the given arrays.. + * - #product: Returns or yields all combinations of elements from +self+ and given arrays. + * + * === Methods for Iterating + * + * - #each: Passes each element to a given block. + * - #reverse_each: Passes each element, in reverse order, to a given block. + * - #each_index: Passes each element index to a given block. + * - #cycle: Calls a given block with each element, then does so again, + * for a specified number of times, or forever. + * - #combination: Calls a given block with combinations of elements of +self+; + * a combination does not use the same element more than once. + * - #permutation: Calls a given block with permutations of elements of +self+; + * a permutation does not use the same element more than once. + * - #repeated_combination: Calls a given block with combinations of elements of +self+; + * a combination may use the same element more than once. + * - #repeated_permutation: Calls a given block with permutations of elements of +self+; + * a permutation may use the same element more than once. + * + * === Methods for Converting + * + * - #map, #collect: Returns an array containing the block return-value for each element. + * - #map!, #collect!: Replaces each element with a block return-value. + * - #flatten: Returns an array that is a recursive flattening of +self+. + * - #flatten!: Replaces each nested array in +self+ with the elements from that array. + * - #inspect, #to_s: Returns a new String containing the elements. + * - #join: Returns a newsString containing the elements joined by the field separator. + * - #to_a: Returns +self+ or a new array containing all elements. + * - #to_ary: Returns +self+. + * - #to_h: Returns a new hash formed from the elements. + * - #transpose: Transposes +self+, which must be an array of arrays. + * - #zip: Returns a new array of arrays containing +self+ and given arrays; + * follow the link for details. + * + * === Other Methods + * + * - #*: Returns one of the following: + * + * - With integer argument +n+, a new array that is the concatenation + * of +n+ copies of +self+. + * - With string argument +field_separator+, a new string that is equivalent to + * <tt>join(field_separator)</tt>. + * + * - #abbrev: Returns a hash of unambiguous abbreviations for elements. + * - #pack: Packs the elements into a binary sequence. + * - #sum: Returns a sum of elements according to either <tt>+</tt> or a given block. */ void Init_Array(void) { -#undef rb_intern -#define rb_intern(str) rb_intern_const(str) - rb_cArray = rb_define_class("Array", rb_cObject); rb_include_module(rb_cArray, rb_mEnumerable); rb_define_alloc_func(rb_cArray, empty_ary_alloc); + rb_define_singleton_method(rb_cArray, "new", rb_ary_s_new, -1); rb_define_singleton_method(rb_cArray, "[]", rb_ary_s_create, -1); rb_define_singleton_method(rb_cArray, "try_convert", rb_ary_s_try_convert, 1); rb_define_method(rb_cArray, "initialize", rb_ary_initialize, -1); @@ -6889,6 +8841,8 @@ Init_Array(void) rb_define_method(rb_cArray, "concat", rb_ary_concat_multi, -1); rb_define_method(rb_cArray, "union", rb_ary_union_multi, -1); rb_define_method(rb_cArray, "difference", rb_ary_difference_multi, -1); + rb_define_method(rb_cArray, "intersection", rb_ary_intersection_multi, -1); + rb_define_method(rb_cArray, "intersect?", rb_ary_intersect_p, 1); rb_define_method(rb_cArray, "<<", rb_ary_push, 1); rb_define_method(rb_cArray, "push", rb_ary_push_m, -1); rb_define_alias(rb_cArray, "append", "push"); @@ -6901,7 +8855,7 @@ Init_Array(void) rb_define_method(rb_cArray, "each_index", rb_ary_each_index, 0); rb_define_method(rb_cArray, "reverse_each", rb_ary_reverse_each, 0); rb_define_method(rb_cArray, "length", rb_ary_length, 0); - rb_define_alias(rb_cArray, "size", "length"); + rb_define_method(rb_cArray, "size", rb_ary_length, 0); rb_define_method(rb_cArray, "empty?", rb_ary_empty_p, 0); rb_define_method(rb_cArray, "find_index", rb_ary_index, -1); rb_define_method(rb_cArray, "index", rb_ary_index, -1); @@ -6961,9 +8915,6 @@ Init_Array(void) rb_define_method(rb_cArray, "flatten", rb_ary_flatten, -1); rb_define_method(rb_cArray, "flatten!", rb_ary_flatten_bang, -1); rb_define_method(rb_cArray, "count", rb_ary_count, -1); - rb_define_method(rb_cArray, "shuffle!", rb_ary_shuffle_bang, -1); - rb_define_method(rb_cArray, "shuffle", rb_ary_shuffle, -1); - rb_define_method(rb_cArray, "sample", rb_ary_sample, -1); rb_define_method(rb_cArray, "cycle", rb_ary_cycle, -1); rb_define_method(rb_cArray, "permutation", rb_ary_permutation, -1); rb_define_method(rb_cArray, "combination", rb_ary_combination, 1); @@ -6985,6 +8936,6 @@ Init_Array(void) rb_define_method(rb_cArray, "sum", rb_ary_sum, -1); rb_define_method(rb_cArray, "deconstruct", rb_ary_deconstruct, 0); - - id_random = rb_intern("random"); } + +#include "array.rbinc" |