diff options
Diffstat (limited to 'array.c')
| -rw-r--r-- | array.c | 7444 |
1 files changed, 5525 insertions, 1919 deletions
@@ -11,63 +11,87 @@ **********************************************************************/ -#include "ruby/ruby.h" -#include "ruby/util.h" -#include "ruby/st.h" -#include "ruby/encoding.h" +#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/st.h" +#include "ruby/util.h" +#include "transient_heap.h" +#include "builtin.h" -#ifndef ARRAY_DEBUG +#if !ARRAY_DEBUG +# undef NDEBUG # define NDEBUG #endif -#include <assert.h> - -#define numberof(array) (int)(sizeof(array) / sizeof((array)[0])) +#include "ruby_assert.h" VALUE rb_cArray; -static ID id_cmp; +/* for OPTIMIZED_CMP: */ +#define id_cmp idCmp #define ARY_DEFAULT_SIZE 16 #define ARY_MAX_SIZE (LONG_MAX / (int)sizeof(VALUE)) +#define SMALL_ARRAY_LEN 16 -void -rb_mem_clear(register VALUE *mem, register long size) +RBIMPL_ATTR_MAYBE_UNUSED() +static int +should_be_T_ARRAY(VALUE ary) { - while (size--) { - *mem++ = Qnil; - } + return RB_TYPE_P(ary, T_ARRAY); } -static inline void -memfill(register VALUE *mem, register long size, register VALUE val) +RBIMPL_ATTR_MAYBE_UNUSED() +static int +should_not_be_shared_and_embedded(VALUE ary) { - while (size--) { - *mem++ = val; - } + return !FL_TEST((ary), ELTS_SHARED) || !FL_TEST((ary), RARRAY_EMBED_FLAG); } -# define ARY_SHARED_P(ary) \ - (assert(!FL_TEST((ary), ELTS_SHARED) || !FL_TEST((ary), RARRAY_EMBED_FLAG)), \ - FL_TEST((ary),ELTS_SHARED)!=0) -# define ARY_EMBED_P(ary) \ - (assert(!FL_TEST((ary), ELTS_SHARED) || !FL_TEST((ary), RARRAY_EMBED_FLAG)), \ - FL_TEST((ary), RARRAY_EMBED_FLAG)!=0) +#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)), \ + RARRAY(a)->as.heap.aux.capa) + #define ARY_EMBED_PTR(a) (assert(ARY_EMBED_P(a)), RARRAY(a)->as.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))) +#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)) -#define ARY_OWNS_HEAP_P(a) (!FL_TEST((a), ELTS_SHARED|RARRAY_EMBED_FLAG)) #define FL_SET_EMBED(a) do { \ assert(!ARY_SHARED_P(a)); \ - assert(!OBJ_FROZEN(a)); \ FL_SET((a), RARRAY_EMBED_FLAG); \ + RARY_TRANSIENT_UNSET(a); \ + ary_verify(a); \ } while (0) + #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)); \ @@ -116,7 +140,7 @@ memfill(register VALUE *mem, register long size, register VALUE val) } while (0) #define ARY_CAPA(ary) (ARY_EMBED_P(ary) ? RARRAY_EMBED_LEN_MAX : \ - ARY_SHARED_ROOT_P(ary) ? RARRAY_LEN(ary) : RARRAY(ary)->as.heap.aux.capa) + ARY_SHARED_ROOT_P(ary) ? RARRAY_LEN(ary) : ARY_HEAP_CAPA(ary)) #define ARY_SET_CAPA(ary, n) do { \ assert(!ARY_EMBED_P(ary)); \ assert(!ARY_SHARED_P(ary)); \ @@ -124,57 +148,348 @@ memfill(register VALUE *mem, register long size, register VALUE val) RARRAY(ary)->as.heap.aux.capa = (n); \ } while (0) -#define ARY_SHARED(ary) (assert(ARY_SHARED_P(ary)), RARRAY(ary)->as.heap.aux.shared) +#define ARY_SHARED_ROOT(ary) (assert(ARY_SHARED_P(ary)), RARRAY(ary)->as.heap.aux.shared_root) #define ARY_SET_SHARED(ary, value) do { \ - assert(!ARY_EMBED_P(ary)); \ - assert(ARY_SHARED_P(ary)); \ - assert(ARY_SHARED_ROOT_P(value)); \ - RARRAY(ary)->as.heap.aux.shared = (value); \ + const VALUE _ary_ = (ary); \ + const VALUE _value_ = (value); \ + assert(!ARY_EMBED_P(_ary_)); \ + assert(ARY_SHARED_P(_ary_)); \ + assert(ARY_SHARED_ROOT_P(_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) (FL_TEST((ary), RARRAY_SHARED_ROOT_FLAG)) -#define ARY_SHARED_NUM(ary) \ +#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_SET_SHARED_NUM(ary, value) do { \ +#define ARY_SHARED_ROOT_OCCUPIED(ary) (ARY_SHARED_ROOT_REFCNT(ary) == 1) +#define ARY_SET_SHARED_ROOT_REFCNT(ary, value) do { \ assert(ARY_SHARED_ROOT_P(ary)); \ RARRAY(ary)->as.heap.aux.capa = (value); \ } while (0) #define FL_SET_SHARED_ROOT(ary) do { \ assert(!ARY_EMBED_P(ary)); \ + assert(!RARRAY_TRANSIENT_P(ary)); \ FL_SET((ary), RARRAY_SHARED_ROOT_FLAG); \ } while (0) +static inline void +ARY_SET(VALUE a, long i, VALUE v) +{ + assert(!ARY_SHARED_P(a)); + assert(!OBJ_FROZEN(a)); + + RARRAY_ASET(a, i, v); +} +#undef RARRAY_ASET + + +#if ARRAY_DEBUG +#define ary_verify(ary) ary_verify_(ary, __FILE__, __LINE__) + +static VALUE +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; + 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(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); + } + else { +#if 1 + const VALUE *ptr = RARRAY_CONST_PTR_TRANSIENT(ary); + long i, len = RARRAY_LEN(ary); + volatile VALUE v; + if (len > 1) len = 1; /* check only HEAD */ + for (i=0; i<len; i++) { + v = ptr[i]; /* access check */ + } + v = v; +#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(); + + return ary; +} + +void +rb_ary_verify(VALUE ary) +{ + ary_verify(ary); +} +#else +#define ary_verify(ary) ((void)0) +#endif + +VALUE * +rb_ary_ptr_use_start(VALUE ary) +{ +#if ARRAY_DEBUG + FL_SET_RAW(ary, RARRAY_PTR_IN_USE_FLAG); +#endif + return (VALUE *)RARRAY_CONST_PTR_TRANSIENT(ary); +} + +void +rb_ary_ptr_use_end(VALUE ary) +{ +#if ARRAY_DEBUG + FL_UNSET_RAW(ary, RARRAY_PTR_IN_USE_FLAG); +#endif +} + +void +rb_mem_clear(VALUE *mem, long size) +{ + while (size--) { + *mem++ = Qnil; + } +} + +static void +ary_mem_clear(VALUE ary, long beg, long size) +{ + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + rb_mem_clear(ptr + beg, size); + }); +} + +static inline void +memfill(register VALUE *mem, register long size, register VALUE val) +{ + while (size--) { + *mem++ = val; + } +} + +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); + }); +} + +static void +ary_memcpy0(VALUE ary, long beg, long argc, const VALUE *argv, VALUE buff_owner_ary) +{ + assert(!ARY_SHARED_P(buff_owner_ary)); + + if (argc > (int)(128/sizeof(VALUE)) /* is magic number (cache line size) */) { + rb_gc_writebarrier_remember(buff_owner_ary); + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + MEMCPY(ptr+beg, argv, VALUE, argc); + }); + } + else { + int i; + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + for (i=0; i<argc; i++) { + RB_OBJ_WRITE(buff_owner_ary, &ptr[i+beg], argv[i]); + } + }); + } +} + +static void +ary_memcpy(VALUE ary, long beg, long argc, const VALUE *argv) +{ + ary_memcpy0(ary, beg, argc, argv, ary); +} + +static VALUE * +ary_heap_alloc(VALUE ary, size_t capa) +{ + VALUE *ptr = rb_transient_heap_alloc(ary, sizeof(VALUE) * capa); + + if (ptr != NULL) { + RARY_TRANSIENT_SET(ary); + } + else { + RARY_TRANSIENT_UNSET(ary); + ptr = ALLOC_N(VALUE, capa); + } + + return ptr; +} + +static void +ary_heap_free_ptr(VALUE ary, const VALUE *ptr, long size) +{ + if (RARRAY_TRANSIENT_P(ary)) { + /* ignore it */ + } + else { + ruby_sized_xfree((void *)ptr, size); + } +} + +static void +ary_heap_free(VALUE ary) +{ + if (RARRAY_TRANSIENT_P(ary)) { + RARY_TRANSIENT_UNSET(ary); + } + else { + ary_heap_free_ptr(ary, ARY_HEAP_PTR(ary), ARY_HEAP_SIZE(ary)); + } +} + +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); + + if (new_ptr == NULL) { + new_ptr = ALLOC_N(VALUE, new_capa); + RARY_TRANSIENT_UNSET(ary); + } + + MEMCPY(new_ptr, ARY_HEAP_PTR(ary), VALUE, old_capa); + ARY_SET_PTR(ary, new_ptr); + } + } + else { + SIZED_REALLOC_N(RARRAY(ary)->as.heap.ptr, VALUE, new_capa, old_capa); + } + ary_verify(ary); + + return alloc_capa; +} + +#if USE_TRANSIENT_HEAP +static inline void +rb_ary_transient_heap_evacuate_(VALUE ary, int transient, int promote) +{ + if (transient) { + 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)); + assert(!ARY_PTR_USING_P(ary)); + + if (promote) { + new_ptr = ALLOC_N(VALUE, capa); + RARY_TRANSIENT_UNSET(ary); + } + else { + new_ptr = ary_heap_alloc(ary, capa); + } + + MEMCPY(new_ptr, old_ptr, VALUE, capa); + /* do not use ARY_SET_PTR() because they assert !frozen */ + RARRAY(ary)->as.heap.ptr = new_ptr; + } + + ary_verify(ary); +} + +void +rb_ary_transient_heap_evacuate(VALUE ary, int promote) +{ + rb_ary_transient_heap_evacuate_(ary, RARRAY_TRANSIENT_P(ary), promote); +} + +void +rb_ary_detransient(VALUE ary) +{ + assert(RARRAY_TRANSIENT_P(ary)); + rb_ary_transient_heap_evacuate_(ary, TRUE, TRUE); +} +#else +void +rb_ary_detransient(VALUE ary) +{ + /* do nothing */ +} +#endif + static void ary_resize_capa(VALUE ary, long capacity) { assert(RARRAY_LEN(ary) <= capacity); assert(!OBJ_FROZEN(ary)); assert(!ARY_SHARED_P(ary)); + if (capacity > RARRAY_EMBED_LEN_MAX) { + size_t new_capa = capacity; if (ARY_EMBED_P(ary)) { long len = ARY_EMBED_LEN(ary); - VALUE *ptr = ALLOC_N(VALUE, (capacity)); + VALUE *ptr = ary_heap_alloc(ary, capacity); + MEMCPY(ptr, ARY_EMBED_PTR(ary), VALUE, len); FL_UNSET_EMBED(ary); ARY_SET_PTR(ary, ptr); ARY_SET_HEAP_LEN(ary, len); } else { - REALLOC_N(RARRAY(ary)->as.heap.ptr, VALUE, (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)) { - long len = RARRAY_LEN(ary); - VALUE *ptr = RARRAY_PTR(ary); + long len = ARY_HEAP_LEN(ary); + long old_capa = ARY_HEAP_CAPA(ary); + const VALUE *ptr = ARY_HEAP_PTR(ary); + if (len > capacity) len = capacity; - MEMCPY(RARRAY(ary)->as.ary, ptr, VALUE, len); + MEMCPY((VALUE *)RARRAY(ary)->as.ary, ptr, VALUE, len); + ary_heap_free_ptr(ary, ptr, old_capa); + FL_SET_EMBED(ary); ARY_SET_LEN(ary, len); - xfree(ptr); } } + + ary_verify(ary); +} + +static inline void +ary_shrink_capa(VALUE ary) +{ + long capacity = ARY_HEAP_LEN(ary); + long old_capa = ARY_HEAP_CAPA(ary); + assert(!ARY_SHARED_P(ary)); + assert(old_capa >= capacity); + if (old_capa > capacity) ary_heap_realloc(ary, capacity); + + ary_verify(ary); } static void @@ -190,28 +505,26 @@ ary_double_capa(VALUE ary, long min) } new_capa += min; ary_resize_capa(ary, new_capa); + + ary_verify(ary); } static void -rb_ary_decrement_share(VALUE shared) +rb_ary_decrement_share(VALUE shared_root) { - if (shared) { - long num = ARY_SHARED_NUM(shared) - 1; - if (num == 0) { - rb_ary_free(shared); - rb_gc_force_recycle(shared); - } - else if (num > 0) { - ARY_SET_SHARED_NUM(shared, num); - } + if (shared_root) { + long num = ARY_SHARED_ROOT_REFCNT(shared_root) - 1; + if (num > 0) { + ARY_SET_SHARED_ROOT_REFCNT(shared_root, num); + } } } static void rb_ary_unshare(VALUE ary) { - VALUE shared = RARRAY(ary)->as.heap.aux.shared; - rb_ary_decrement_share(shared); + VALUE shared_root = RARRAY(ary)->as.heap.aux.shared_root; + rb_ary_decrement_share(shared_root); FL_UNSET_SHARED(ary); } @@ -224,92 +537,187 @@ rb_ary_unshare_safe(VALUE ary) } static VALUE -rb_ary_increment_share(VALUE shared) +rb_ary_increment_share(VALUE shared_root) { - long num = ARY_SHARED_NUM(shared); + long num = ARY_SHARED_ROOT_REFCNT(shared_root); if (num >= 0) { - ARY_SET_SHARED_NUM(shared, num + 1); + ARY_SET_SHARED_ROOT_REFCNT(shared_root, num + 1); } - return shared; + return shared_root; } static void -rb_ary_set_shared(VALUE ary, VALUE shared) +rb_ary_set_shared(VALUE ary, VALUE shared_root) { - rb_ary_increment_share(shared); + rb_ary_increment_share(shared_root); FL_SET_SHARED(ary); - ARY_SET_SHARED(ary, shared); + RB_DEBUG_COUNTER_INC(obj_ary_shared_create); + ARY_SET_SHARED(ary, shared_root); } static inline void rb_ary_modify_check(VALUE ary) { rb_check_frozen(ary); - if (!OBJ_UNTRUSTED(ary) && rb_safe_level() >= 4) - rb_raise(rb_eSecurityError, "Insecure: can't modify array"); + ary_verify(ary); } void -rb_ary_modify(VALUE ary) +rb_ary_cancel_sharing(VALUE ary) { - rb_ary_modify_check(ary); if (ARY_SHARED_P(ary)) { - long 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) { - VALUE *ptr = ARY_HEAP_PTR(ary); - VALUE shared = ARY_SHARED(ary); + const VALUE *ptr = ARY_HEAP_PTR(ary); FL_UNSET_SHARED(ary); FL_SET_EMBED(ary); - MEMCPY(ARY_EMBED_PTR(ary), ptr, VALUE, len); - rb_ary_decrement_share(shared); + 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); + ARY_SET_PTR(ary, RARRAY_CONST_PTR_TRANSIENT(shared_root)); + ARY_SET_CAPA(ary, shared_len); + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + MEMMOVE(ptr, ptr+shift, VALUE, len); + }); + FL_SET_EMBED(shared_root); + rb_ary_decrement_share(shared_root); + } else { - VALUE *ptr = ALLOC_N(VALUE, len); - MEMCPY(ptr, RARRAY_PTR(ary), VALUE, len); + VALUE *ptr = ary_heap_alloc(ary, len); + MEMCPY(ptr, ARY_HEAP_PTR(ary), VALUE, len); rb_ary_unshare(ary); ARY_SET_CAPA(ary, len); ARY_SET_PTR(ary, ptr); } + + rb_gc_writebarrier_remember(ary); } + ary_verify(ary); } -VALUE -rb_ary_freeze(VALUE ary) +void +rb_ary_modify(VALUE ary) { - return rb_obj_freeze(ary); + rb_ary_modify_check(ary); + rb_ary_cancel_sharing(ary); +} + +static VALUE +ary_ensure_room_for_push(VALUE ary, long add_len) +{ + long old_len = RARRAY_LEN(ary); + long new_len = old_len + add_len; + long capa; + + if (old_len > ARY_MAX_SIZE - add_len) { + rb_raise(rb_eIndexError, "index %ld too big", new_len); + } + if (ARY_SHARED_P(ary)) { + if (new_len > RARRAY_EMBED_LEN_MAX) { + 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); + + 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); + } + ary_verify(ary); + return ary; + } + } + } + ary_verify(ary); + rb_ary_modify(ary); + } + else { + rb_ary_modify_check(ary); + } + capa = ARY_CAPA(ary); + if (new_len > capa) { + ary_double_capa(ary, new_len); + } + + ary_verify(ary); + return ary; } /* * call-seq: - * ary.frozen? -> true or false + * array.freeze -> self + * + * Freezes +self+; returns +self+: + * a = [] + * a.frozen? # => false + * a.freeze + * a.frozen? # => true * - * Return <code>true</code> if this array is frozen (or temporarily frozen - * while being sorted). + * An attempt to modify a frozen \Array raises FrozenError. */ -static VALUE -rb_ary_frozen_p(VALUE ary) +VALUE +rb_ary_freeze(VALUE ary) { - if (OBJ_FROZEN(ary)) return Qtrue; + return rb_obj_freeze(ary); +} + +/* This can be used to take a snapshot of an array (with + e.g. rb_ary_replace) and check later whether the array has been + modified from the snapshot. The snapshot is cheap, though if + something does modify the array it will pay the cost of copying + it. If Array#pop or Array#shift has been called, the array will + be still shared with the snapshot, but the array length will + differ. */ +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; + } return Qfalse; } static VALUE ary_alloc(VALUE klass) { - NEWOBJ(ary, struct RArray); - OBJSETUP(ary, klass, T_ARRAY); - FL_SET_EMBED((VALUE)ary); - ARY_SET_EMBED_LEN((VALUE)ary, 0); - + NEWOBJ_OF(ary, struct RArray, klass, T_ARRAY | RARRAY_EMBED_FLAG | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0)); + /* Created array is: + * FL_SET_EMBED((VALUE)ary); + * ARY_SET_EMBED_LEN((VALUE)ary, 0); + */ return (VALUE)ary; } static VALUE +empty_ary_alloc(VALUE klass) +{ + RUBY_DTRACE_CREATE_HOOK(ARRAY, 0); + return ary_alloc(klass); +} + +static VALUE ary_new(VALUE klass, long capa) { - VALUE ary; + VALUE ary,*ptr; if (capa < 0) { rb_raise(rb_eArgError, "negative array size (or size too big)"); @@ -317,10 +725,14 @@ ary_new(VALUE klass, long capa) if (capa > ARY_MAX_SIZE) { rb_raise(rb_eArgError, "array size too big"); } + + RUBY_DTRACE_CREATE_HOOK(ARRAY, capa); + ary = ary_alloc(klass); if (capa > RARRAY_EMBED_LEN_MAX) { + ptr = ary_heap_alloc(ary, capa); FL_UNSET_EMBED(ary); - ARY_SET_PTR(ary, ALLOC_N(VALUE, capa)); + ARY_SET_PTR(ary, ptr); ARY_SET_CAPA(ary, capa); ARY_SET_HEAP_LEN(ary, 0); } @@ -329,22 +741,19 @@ ary_new(VALUE klass, long capa) } VALUE -rb_ary_new2(long capa) +rb_ary_new_capa(long capa) { return ary_new(rb_cArray, capa); } - VALUE rb_ary_new(void) { return rb_ary_new2(RARRAY_EMBED_LEN_MAX); } -#include <stdarg.h> - VALUE -rb_ary_new3(long n, ...) +(rb_ary_new_from_args)(long n, ...) { va_list ar; VALUE ary; @@ -354,7 +763,7 @@ rb_ary_new3(long n, ...) va_start(ar, n); for (i=0; i<n; i++) { - RARRAY_PTR(ary)[i] = va_arg(ar, VALUE); + ARY_SET(ary, i, va_arg(ar, VALUE)); } va_end(ar); @@ -362,14 +771,72 @@ rb_ary_new3(long n, ...) return ary; } +MJIT_FUNC_EXPORTED VALUE +rb_ary_tmp_new_from_values(VALUE klass, long n, const VALUE *elts) +{ + VALUE ary; + + ary = ary_new(klass, n); + if (n > 0 && elts) { + ary_memcpy(ary, 0, n, elts); + ARY_SET_LEN(ary, n); + } + + return ary; +} + +VALUE +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(rb_execution_context_t *ec, VALUE klass) +{ + RB_EC_NEWOBJ_OF(ec, ary, struct RArray, klass, T_ARRAY | RARRAY_EMBED_FLAG | (RGENGC_WB_PROTECTED_ARRAY ? FL_WB_PROTECTED : 0)); + /* Created array is: + * FL_SET_EMBED((VALUE)ary); + * ARY_SET_EMBED_LEN((VALUE)ary, 0); + */ + 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); + + ary = ec_ary_alloc(ec, klass); + + if (capa > RARRAY_EMBED_LEN_MAX) { + 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); + } + + return ary; +} + VALUE -rb_ary_new4(long n, const VALUE *elts) +rb_ec_ary_new_from_values(rb_execution_context_t *ec, long n, const VALUE *elts) { VALUE ary; - ary = rb_ary_new2(n); + ary = ec_ary_new(ec, rb_cArray, n); if (n > 0 && elts) { - MEMCPY(RARRAY_PTR(ary), elts, VALUE, n); + ary_memcpy(ary, 0, n, elts); ARY_SET_LEN(ary, n); } @@ -379,14 +846,48 @@ rb_ary_new4(long n, const VALUE *elts) VALUE rb_ary_tmp_new(long capa) { - return ary_new(0, capa); + VALUE ary = ary_new(0, capa); + rb_ary_transient_heap_evacuate(ary, TRUE); + return ary; +} + +VALUE +rb_ary_tmp_new_fill(long capa) +{ + VALUE ary = ary_new(0, capa); + ary_memfill(ary, 0, capa, Qnil); + ARY_SET_LEN(ary, capa); + rb_ary_transient_heap_evacuate(ary, TRUE); + return ary; } void rb_ary_free(VALUE ary) { if (ARY_OWNS_HEAP_P(ary)) { - xfree(ARY_HEAP_PTR(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); + } + else { + RB_DEBUG_COUNTER_INC(obj_ary_ptr); + ary_heap_free(ary); + } + } + else { + RB_DEBUG_COUNTER_INC(obj_ary_embed); + } + + if (ARY_SHARED_P(ary)) { + RB_DEBUG_COUNTER_INC(obj_ary_shared); + } + if (ARY_SHARED_ROOT_P(ary) && ARY_SHARED_ROOT_OCCUPIED(ary)) { + RB_DEBUG_COUNTER_INC(obj_ary_shared_root_occupied); } } @@ -394,7 +895,7 @@ RUBY_FUNC_EXPORTED size_t rb_ary_memsize(VALUE ary) { if (ARY_OWNS_HEAP_P(ary)) { - return RARRAY(ary)->as.heap.aux.capa * sizeof(VALUE); + return ARY_CAPA(ary) * sizeof(VALUE); } else { return 0; @@ -406,49 +907,64 @@ ary_discard(VALUE ary) { rb_ary_free(ary); RBASIC(ary)->flags |= RARRAY_EMBED_FLAG; - RBASIC(ary)->flags &= ~RARRAY_EMBED_LEN_MASK; + RBASIC(ary)->flags &= ~(RARRAY_EMBED_LEN_MASK | RARRAY_TRANSIENT_FLAG); } static VALUE ary_make_shared(VALUE ary) { assert(!ARY_EMBED_P(ary)); + ary_verify(ary); + if (ARY_SHARED_P(ary)) { - return ARY_SHARED(ary); + return ARY_SHARED_ROOT(ary); } else if (ARY_SHARED_ROOT_P(ary)) { return ary; } else if (OBJ_FROZEN(ary)) { - ary_resize_capa(ary, ARY_HEAP_LEN(ary)); + rb_ary_transient_heap_evacuate(ary, TRUE); + ary_shrink_capa(ary); FL_SET_SHARED_ROOT(ary); - ARY_SET_SHARED_NUM(ary, 1); + ARY_SET_SHARED_ROOT_REFCNT(ary, 1); return ary; } else { - NEWOBJ(shared, struct RArray); - OBJSETUP(shared, 0, T_ARRAY); - FL_UNSET_EMBED(shared); - - ARY_SET_LEN((VALUE)shared, RARRAY_LEN(ary)); - ARY_SET_PTR((VALUE)shared, RARRAY_PTR(ary)); - FL_SET_SHARED_ROOT(shared); - ARY_SET_SHARED_NUM((VALUE)shared, 1); + 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)); + VALUE vshared = (VALUE)shared; + + rb_ary_transient_heap_evacuate(ary, TRUE); + ptr = ARY_HEAP_PTR(ary); + + FL_UNSET_EMBED(vshared); + ARY_SET_LEN(vshared, capa); + ARY_SET_PTR(vshared, ptr); + ary_mem_clear(vshared, len, capa - len); + FL_SET_SHARED_ROOT(vshared); + ARY_SET_SHARED_ROOT_REFCNT(vshared, 1); FL_SET_SHARED(ary); - ARY_SET_SHARED(ary, (VALUE)shared); - OBJ_FREEZE(shared); - return (VALUE)shared; + RB_DEBUG_COUNTER_INC(obj_ary_shared_create); + ARY_SET_SHARED(ary, vshared); + OBJ_FREEZE(vshared); + + ary_verify(vshared); + ary_verify(ary); + + return vshared; } } - static VALUE ary_make_substitution(VALUE ary) { - if (RARRAY_LEN(ary) <= RARRAY_EMBED_LEN_MAX) { - VALUE subst = rb_ary_new2(RARRAY_LEN(ary)); - MEMCPY(ARY_EMBED_PTR(subst), RARRAY_PTR(ary), VALUE, RARRAY_LEN(ary)); - ARY_SET_EMBED_LEN(subst, RARRAY_LEN(ary)); + long len = RARRAY_LEN(ary); + + if (len <= RARRAY_EMBED_LEN_MAX) { + VALUE subst = rb_ary_new2(len); + ary_memcpy(subst, 0, len, RARRAY_CONST_PTR_TRANSIENT(ary)); + ARY_SET_EMBED_LEN(subst, len); return subst; } else { @@ -462,35 +978,43 @@ rb_assoc_new(VALUE car, VALUE cdr) return rb_ary_new3(2, car, cdr); } -static VALUE -to_ary(VALUE ary) +VALUE +rb_to_array_type(VALUE ary) { - return rb_convert_type(ary, T_ARRAY, "Array", "to_ary"); + return rb_convert_type_with_id(ary, T_ARRAY, "Array", idTo_ary); } +#define to_ary rb_to_array_type VALUE rb_check_array_type(VALUE ary) { - return rb_check_convert_type(ary, T_ARRAY, "Array", "to_ary"); + return rb_check_convert_type_with_id(ary, T_ARRAY, "Array", idTo_ary); +} + +MJIT_FUNC_EXPORTED VALUE +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 +to_ary+ method. Returns the - * converted array or +nil+ if +obj+ cannot be converted for any reason. - * 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 @@ -501,58 +1025,46 @@ rb_ary_s_try_convert(VALUE dummy, VALUE ary) /* * call-seq: - * Array.new(size=0, obj=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 +obj+ are sent, an array is created with - * +size+ copies of +obj+. Take notice that all elements will reference the - * same object +obj+. - * - * The second form creates a copy of the array passed as a parameter (the - * array is generated by calling to_ary on the parameter). - * - * first_array = ["Matz", "Guido"] - * - * second_array = Array.new(first_array) #=> ["Matz", "Guido"] - * - * first_array.equal? second_array #=> false - * - * 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. - * - * Array.new(3){ |index| index ** 2 } - * # => [0, 1, 4] - * - * == Common gotchas - * - * When sending the second parameter, the same object will be used as the - * value for all the array elements: - * - * a = Array.new(2, Hash.new) - * # => [{}, {}] - * - * a[0]['cat'] = 'feline' - * a # => [{"cat"=>"feline"}, {"cat"=>"feline"}] - * - * 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"}, {}] - * + * 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 + * + * Returns a new \Array. + * + * With no block and no arguments, returns a new empty \Array object. + * + * With no block and a single \Array argument +array+, + * returns a new \Array formed from +array+: + * a = Array.new([:foo, 'bar', 2]) + * a.class # => Array + * a # => [:foo, "bar", 2] + * + * With no block and a single \Integer argument +size+, + * returns a new \Array of the given size + * whose elements are all +nil+: + * a = Array.new(3) + * a # => [nil, nil, nil] + * + * With no block and arguments +size+ and +default_value+, + * returns an \Array of the given size; + * each element is that same +default_value+: + * a = Array.new(3, 'x') + * a # => ['x', 'x', 'x'] + * + * 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 = Array.new(3) {|index| "Element #{index}" } + * a # => ["Element 0", "Element 1", "Element 2"] + * + * 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 @@ -563,8 +1075,8 @@ rb_ary_initialize(int argc, VALUE *argv, VALUE ary) rb_ary_modify(ary); if (argc == 0) { - if (ARY_OWNS_HEAP_P(ary) && RARRAY_PTR(ary)) { - xfree(RARRAY_PTR(ary)); + if (ARY_OWNS_HEAP_P(ary) && ARY_HEAP_PTR(ary) != NULL) { + ary_heap_free(ary); } rb_ary_unshare_safe(ary); FL_SET_EMBED(ary); @@ -584,12 +1096,14 @@ rb_ary_initialize(int argc, VALUE *argv, VALUE 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"); } if (len > ARY_MAX_SIZE) { 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()) { @@ -604,27 +1118,26 @@ rb_ary_initialize(int argc, VALUE *argv, VALUE ary) } } else { - memfill(RARRAY_PTR(ary), len, val); + ary_memfill(ary, 0, len, val); ARY_SET_LEN(ary, len); } return ary; } - /* -* Returns a new array populated with the given objects. -* -* Array.[]( 1, 'a', /^A/ ) -* Array[ 1, 'a', /^A/ ] -* [ 1, 'a', /^A/ ] -*/ + * Returns a new array populated with the given objects. + * + * Array.[]( 1, 'a', /^A/) # => [1, "a", /^A/] + * Array[ 1, 'a', /^A/ ] # => [1, "a", /^A/] + * [ 1, 'a', /^A/ ] # => [1, "a", /^A/] + */ static VALUE rb_ary_s_create(int argc, VALUE *argv, VALUE klass) { VALUE ary = ary_new(klass, argc); if (argc > 0 && argv) { - MEMCPY(RARRAY_PTR(ary), argv, VALUE, argc); + ary_memcpy(ary, 0, argc, argv); ARY_SET_LEN(ary, argc); } @@ -634,11 +1147,13 @@ rb_ary_s_create(int argc, VALUE *argv, VALUE klass) void rb_ary_store(VALUE ary, long idx, VALUE val) { + long len = RARRAY_LEN(ary); + if (idx < 0) { - idx += RARRAY_LEN(ary); + idx += len; if (idx < 0) { rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", - idx - RARRAY_LEN(ary), -RARRAY_LEN(ary)); + idx - len, -len); } } else if (idx >= ARY_MAX_SIZE) { @@ -649,15 +1164,14 @@ rb_ary_store(VALUE ary, long idx, VALUE val) if (idx >= ARY_CAPA(ary)) { ary_double_capa(ary, idx); } - if (idx > RARRAY_LEN(ary)) { - rb_mem_clear(RARRAY_PTR(ary) + RARRAY_LEN(ary), - idx-RARRAY_LEN(ary) + 1); + if (idx > len) { + ary_mem_clear(ary, len, idx - len + 1); } - if (idx >= RARRAY_LEN(ary)) { + if (idx >= len) { ARY_SET_LEN(ary, idx + 1); } - RARRAY_PTR(ary)[idx] = val; + ARY_SET(ary, idx, val); } static VALUE @@ -669,7 +1183,7 @@ ary_make_partial(VALUE ary, VALUE klass, long offset, long len) if (len <= RARRAY_EMBED_LEN_MAX) { VALUE result = ary_alloc(klass); - MEMCPY(ARY_EMBED_PTR(result), RARRAY_PTR(ary) + offset, VALUE, len); + ary_memcpy(result, 0, len, RARRAY_CONST_PTR_TRANSIENT(ary) + offset); ARY_SET_EMBED_LEN(result, len); return result; } @@ -678,20 +1192,69 @@ ary_make_partial(VALUE ary, VALUE klass, long offset, long len) FL_UNSET_EMBED(result); shared = ary_make_shared(ary); - ARY_SET_PTR(result, RARRAY_PTR(ary)); + ARY_SET_PTR(result, RARRAY_CONST_PTR_TRANSIENT(ary)); ARY_SET_LEN(result, RARRAY_LEN(ary)); rb_ary_set_shared(result, shared); ARY_INCREASE_PTR(result, offset); ARY_SET_LEN(result, len); + + ary_verify(shared); + ary_verify(result); return result; } } 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) || (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; + } + + long ustep = (step < 0) ? -step : step; + len = (len + ustep - 1) / ustep; + + long i; + long j = offset + ((step > 0) ? 0 : (orig_len - 1)); + VALUE result = ary_new(klass, len); + if (len <= RARRAY_EMBED_LEN_MAX) { + 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 @@ -701,82 +1264,93 @@ enum ary_take_pos_flags }; static VALUE -ary_take_first_or_last(int argc, VALUE *argv, VALUE ary, enum ary_take_pos_flags last) +ary_take_first_or_last(int argc, const VALUE *argv, VALUE ary, enum ary_take_pos_flags last) { - VALUE nv; long n; + long len; long offset = 0; - rb_scan_args(argc, argv, "1", &nv); - n = NUM2LONG(nv); - if (n > RARRAY_LEN(ary)) { - n = RARRAY_LEN(ary); + argc = rb_check_arity(argc, 0, 1); + /* the case optional argument is omitted should be handled in + * callers of this function. if another arity case is added, + * this arity check needs to rewrite. */ + RUBY_ASSERT_ALWAYS(argc == 1); + + n = NUM2LONG(argv[0]); + len = RARRAY_LEN(ary); + if (n > len) { + n = len; } else if (n < 0) { rb_raise(rb_eArgError, "negative array size"); } if (last) { - offset = RARRAY_LEN(ary) - n; + offset = len - n; } return ary_make_partial(ary, rb_cArray, offset, n); } -static VALUE rb_ary_push_1(VALUE ary, VALUE item); - /* * call-seq: - * ary << obj -> ary + * array << object -> 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. - * - * [ 1, 2 ] << "c" << "d" << [ 3, 4 ] - * #=> [ 1, 2, "c", "d", [ 3, 4 ] ] + * Appends +object+ to +self+; returns +self+: + * a = [:foo, 'bar', 2] + * a << :baz # => [:foo, "bar", 2, :baz] * + * 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]] */ VALUE rb_ary_push(VALUE ary, VALUE item) { - rb_ary_modify(ary); - return rb_ary_push_1(ary, 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); + }); + ARY_SET_LEN(ary, idx + 1); + ary_verify(ary); + return ary; } -static VALUE -rb_ary_push_1(VALUE ary, VALUE item) +VALUE +rb_ary_cat(VALUE ary, const VALUE *argv, long len) { - long idx = RARRAY_LEN(ary); - - if (idx >= ARY_CAPA(ary)) { - ary_double_capa(ary, idx); - } - RARRAY_PTR(ary)[idx] = item; - ARY_SET_LEN(ary, idx + 1); + long oldlen = RARRAY_LEN(ary); + VALUE target_ary = ary_ensure_room_for_push(ary, len); + ary_memcpy0(ary, oldlen, len, argv, target_ary); + ARY_SET_LEN(ary, oldlen + len); return ary; } /* * call-seq: - * ary.push(obj, ... ) -> ary + * array.push(*objects) -> self + * + * Appends trailing elements. * - * 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. + * Appends each argument in +objects+ to +self+; returns +self+: + * a = [:foo, 'bar', 2] + * a.push(:baz, :bat) # => [:foo, "bar", 2, :baz, :bat] * - * a = [ "a", "b", "c" ] - * a.push("d", "e", "f") - * #=> ["a", "b", "c", "d", "e", "f"] + * 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 rb_ary_push_m(int argc, VALUE *argv, VALUE ary) { - rb_ary_modify(ary); - while (argc--) { - rb_ary_push_1(ary, *argv++); - } - return ary; + return rb_ary_cat(ary, argv, argc); } VALUE @@ -784,33 +1358,46 @@ rb_ary_pop(VALUE ary) { long n; rb_ary_modify_check(ary); - if (RARRAY_LEN(ary) == 0) return Qnil; + n = RARRAY_LEN(ary); + if (n == 0) return Qnil; if (ARY_OWNS_HEAP_P(ary) && - RARRAY_LEN(ary) * 3 < ARY_CAPA(ary) && + n * 3 < ARY_CAPA(ary) && ARY_CAPA(ary) > ARY_DEFAULT_SIZE) { - ary_resize_capa(ary, RARRAY_LEN(ary) * 2); + ary_resize_capa(ary, n * 2); } - n = RARRAY_LEN(ary)-1; + --n; ARY_SET_LEN(ary, n); - return RARRAY_PTR(ary)[n]; + ary_verify(ary); + return RARRAY_AREF(ary, n); } /* * 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. * - * Removes the last element from +self+ and returns it, or - * <code>nil</code> if the array is empty. + * 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"] * - * 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. + * Returns +nil+ if the array is empty. * - * a = [ "a", "b", "c", "d" ] - * a.pop #=> "d" - * a.pop(2) #=> ["b", "c"] - * a #=> ["a"] + * When a non-negative \Integer argument +n+ is given and is in range, + * removes and returns the last +n+ elements in a new \Array: + * a = [:foo, 'bar', 2] + * a.pop(2) # => ["bar", 2] + * + * 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 @@ -825,6 +1412,7 @@ rb_ary_pop_m(int argc, VALUE *argv, VALUE ary) rb_ary_modify_check(ary); result = ary_take_first_or_last(argc, argv, ary, ARY_TAKE_LAST); ARY_INCREASE_LEN(ary, -RARRAY_LEN(result)); + ary_verify(ary); return result; } @@ -832,49 +1420,64 @@ VALUE rb_ary_shift(VALUE ary) { VALUE top; + long len = RARRAY_LEN(ary); rb_ary_modify_check(ary); - if (RARRAY_LEN(ary) == 0) return Qnil; - top = RARRAY_PTR(ary)[0]; + if (len == 0) return Qnil; + top = RARRAY_AREF(ary, 0); if (!ARY_SHARED_P(ary)) { - if (RARRAY_LEN(ary) < ARY_DEFAULT_SIZE) { - MEMMOVE(RARRAY_PTR(ary), RARRAY_PTR(ary)+1, VALUE, RARRAY_LEN(ary)-1); + 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 */ - RARRAY_PTR(ary)[0] = Qnil; + ARY_SET(ary, 0, Qnil); ary_make_shared(ary); } - else if (ARY_SHARED_NUM(ARY_SHARED(ary)) == 1) { - RARRAY_PTR(ary)[0] = Qnil; + else if (ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))) { + RARRAY_PTR_USE_TRANSIENT(ary, ptr, ptr[0] = Qnil); } ARY_INCREASE_PTR(ary, 1); /* shift ptr */ ARY_INCREASE_LEN(ary, -1); + ary_verify(ary); + 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 the first element of +self+ and removes it (shifting all - * other elements down by one). Returns <code>nil</code> if the array - * is empty. + * Returns +nil+ if +self+ is empty. * - * 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. + * When positive \Integer argument +n+ is given, removes the first +n+ elements; + * returns those elements in a new \Array: + * 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: + * a = [:foo, 'bar', 2] + * a.shift(3) # => [:foo, 'bar', 2] * - * args = [ "-m", "-q", "filename" ] - * args.shift(2) #=> ["-m", "-q"] - * args #=> ["filename"] + * If +n+ is zero, returns a new empty \Array; +self+ is unmodified. + * + * Related: #push, #pop, #unshift. */ static VALUE @@ -890,48 +1493,167 @@ rb_ary_shift_m(int argc, VALUE *argv, VALUE ary) rb_ary_modify_check(ary); result = ary_take_first_or_last(argc, argv, ary, ARY_TAKE_FIRST); n = RARRAY_LEN(result); - if (ARY_SHARED_P(ary)) { - if (ARY_SHARED_NUM(ARY_SHARED(ary)) == 1) { - rb_mem_clear(RARRAY_PTR(ary), n); - } - ARY_INCREASE_PTR(ary, n); + rb_ary_behead(ary,n); + + return result; +} + +static VALUE +behead_shared(VALUE ary, long n) +{ + assert(ARY_SHARED_P(ary)); + rb_ary_modify_check(ary); + if (ARY_SHARED_ROOT_OCCUPIED(ARY_SHARED_ROOT(ary))) { + ary_mem_clear(ary, 0, n); + } + ARY_INCREASE_PTR(ary, n); + ARY_INCREASE_LEN(ary, -n); + ary_verify(ary); + return ary; +} + +static VALUE +behead_transient(VALUE ary, long n) +{ + rb_ary_modify_check(ary); + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + MEMMOVE(ptr, ptr+n, VALUE, RARRAY_LEN(ary)-n); + }); /* WB: no new reference */ + ARY_INCREASE_LEN(ary, -n); + ary_verify(ary); + return ary; +} + +MJIT_FUNC_EXPORTED VALUE +rb_ary_behead(VALUE ary, long n) +{ + if (n <= 0) { + return ary; + } + else if (ARY_SHARED_P(ary)) { + return behead_shared(ary, n); + } + else if (RARRAY_LEN(ary) >= ARY_DEFAULT_SIZE) { + ary_make_shared(ary); + return behead_shared(ary, n); } else { - MEMMOVE(RARRAY_PTR(ary), RARRAY_PTR(ary)+n, VALUE, RARRAY_LEN(ary)-n); + return behead_transient(ary, n); } - ARY_INCREASE_LEN(ary, -n); +} - return result; +static VALUE +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; + + rb_ary_modify(ary); + capa = ARY_CAPA(ary); + if (capa - (capa >> 6) <= new_len) { + ary_double_capa(ary, new_len); + } + + /* use shared array for big "queues" */ + if (new_len > ARY_DEFAULT_SIZE * 4) { + ary_verify(ary); + + /* make a room for unshifted items */ + capa = ARY_CAPA(ary); + ary_make_shared(ary); + + head = sharedp = RARRAY_CONST_PTR_TRANSIENT(ary); + return make_room_for_unshift(ary, head, (void *)sharedp, argc, capa, len); + } + else { + /* sliding items */ + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + MEMMOVE(ptr + argc, ptr, VALUE, len); + }); + + ary_verify(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 + * array.unshift(*objects) -> self + * + * Prepends the given +objects+ to +self+: + * a = [:foo, 'bar', 2] + * a.unshift(:bam, :bat) # => [:bam, :bat, :foo, "bar", 2] * - * Prepends objects to the front of +self+, - * moving other elements upwards. + * Array#prepend is an alias for Array#unshift. * - * a = [ "b", "c", "d" ] - * a.unshift("a") #=> ["a", "b", "c", "d"] - * a.unshift(1, 2) #=> [ 1, 2, "a", "b", "c", "d"] + * Related: #push, #pop, #shift. */ static VALUE rb_ary_unshift_m(int argc, VALUE *argv, VALUE ary) { - long len; + long len = RARRAY_LEN(ary); + VALUE target_ary; - rb_ary_modify(ary); - if (argc == 0) return ary; - if (ARY_CAPA(ary) <= (len = RARRAY_LEN(ary)) + argc) { - ary_double_capa(ary, len + argc); + if (argc == 0) { + rb_ary_modify_check(ary); + return ary; } - /* sliding items */ - MEMMOVE(RARRAY_PTR(ary) + argc, RARRAY_PTR(ary), VALUE, len); - MEMCPY(RARRAY_PTR(ary), argv, VALUE, argc); - ARY_INCREASE_LEN(ary, argc); - + target_ary = ary_ensure_room_for_unshift(ary, argc); + ary_memcpy0(ary, 0, argc, argv, target_ary); + ARY_SET_LEN(ary, len + argc); return ary; } @@ -945,119 +1667,198 @@ rb_ary_unshift(VALUE ary, VALUE item) static inline VALUE rb_ary_elt(VALUE ary, long offset) { - if (RARRAY_LEN(ary) == 0) return Qnil; - if (offset < 0 || RARRAY_LEN(ary) <= offset) { + long len = RARRAY_LEN(ary); + if (len == 0) return Qnil; + if (offset < 0 || len <= offset) { return Qnil; } - return RARRAY_PTR(ary)[offset]; + return RARRAY_AREF(ary, offset); } VALUE rb_ary_entry(VALUE ary, long offset) { - if (offset < 0) { - offset += RARRAY_LEN(ary); - } - return rb_ary_elt(ary, offset); + return rb_ary_entry_internal(ary, 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); - if (beg > RARRAY_LEN(ary)) return Qnil; + if (beg > alen) return Qnil; if (beg < 0 || len < 0) return Qnil; - if (RARRAY_LEN(ary) < len || RARRAY_LEN(ary) < beg + len) { - len = RARRAY_LEN(ary) - beg; + if (alen < len || alen < beg + len) { + 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 _start_ and - * continuing for _length_ elements, or returns a subarray - * specified by _range_. - * Negative indices count backward from the end of the - * array (-1 is the last element). 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[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 -rb_ary_aref(int argc, VALUE *argv, VALUE ary) +rb_ary_aref(int argc, const VALUE *argv, VALUE ary) { - VALUE arg; - long beg, len; - + rb_check_arity(argc, 1, 2); if (argc == 2) { - beg = NUM2LONG(argv[0]); - len = NUM2LONG(argv[1]); - if (beg < 0) { - beg += RARRAY_LEN(ary); - } - return rb_ary_subseq(ary, beg, len); + return rb_ary_aref2(ary, argv[0], argv[1]); } - if (argc != 1) { - rb_scan_args(argc, argv, "11", 0, 0); + return rb_ary_aref1(ary, argv[0]); +} + +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); } - arg = argv[0]; + return rb_ary_subseq(ary, beg, len); +} + +MJIT_FUNC_EXPORTED VALUE +rb_ary_aref1(VALUE ary, VALUE arg) +{ + long beg, len, step; + /* special case - speeding up */ if (FIXNUM_P(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 <code>Array#[]</code>. - * - * a = [ "a", "b", "c", "d", "e" ] - * a.at(0) #=> "a" - * a.at(-1) #=> "e" + * Returns the element at \Integer offset +index+; does not modify +self+. + * a = [:foo, 'bar', 2] + * a.at(0) # => :foo + * a.at(2) # => 2 */ -static VALUE +VALUE rb_ary_at(VALUE ary, VALUE pos) { return rb_ary_entry(ary, NUM2LONG(pos)); @@ -1065,24 +1866,39 @@ 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 <code>nil</code>, and the - * second form returns an empty array. + * 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_PTR(ary)[0]; + return RARRAY_AREF(ary, 0); } else { return ary_take_first_or_last(argc, argv, ary, ARY_TAKE_FIRST); @@ -1091,23 +1907,41 @@ rb_ary_first(int argc, VALUE *argv, VALUE ary) /* * 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 <code>nil</code>. + * When no argument is given, returns the last element: + * 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 \Innteger 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, VALUE *argv, VALUE ary) +rb_ary_last(int argc, const VALUE *argv, VALUE ary) { if (argc == 0) { - if (RARRAY_LEN(ary) == 0) return Qnil; - return RARRAY_PTR(ary)[RARRAY_LEN(ary)-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); @@ -1116,22 +1950,35 @@ rb_ary_last(int argc, VALUE *argv, VALUE ary) /* * call-seq: - * ary.fetch(index) -> obj - * ary.fetch(index, default ) -> obj - * ary.fetch(index) {|index| block } -> obj - * - * Tries to return the element at position <i>index</i>. If the index - * lies outside the array, the first form throws an - * <code>IndexError</code> exception, the second form returns - * <i>default</i>, and the third form returns the value of invoking - * the block, passing in the index. Negative values of <i>index</i> - * 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(4) { |i| i*i } #=> 16 + * 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 @@ -1159,29 +2006,42 @@ rb_ary_fetch(int argc, VALUE *argv, VALUE ary) } return ifnone; } - return RARRAY_PTR(ary)[idx]; + return RARRAY_AREF(ary, idx); } /* * call-seq: - * ary.index(obj) -> int or nil - * ary.index {|item| block} -> int or nil - * ary.index -> an_enumerator + * array.index(object) -> integer or nil + * array.index {|element| ... } -> integer or nil + * array.index -> new_enumerator * - * Returns the index of the first object in +self+ such that the object is - * <code>==</code> to <i>obj</i>. If a block is given instead of an - * argument, returns index of first object for which <em>block</em> is true. - * Returns <code>nil</code> if no match is found. - * See also <code>Array#rindex</code>. + * Returns the index of a specified element. * - * If neither block nor argument is given, an enumerator is returned instead. + * When argument +object+ is given but no block, + * returns the index of the first element +element+ + * for which <tt>object == element</tt>: + * a = [:foo, 'bar', 2, 'bar'] + * a.index('bar') # => 1 * - * a = [ "a", "b", "c" ] - * a.index("b") #=> 1 - * a.index("z") #=> nil - * a.index{|x|x=="b"} #=> 1 + * Returns +nil+ if no such element found. * - * This is an alias of <code>#find_index</code>. + * 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: + * a = [:foo, 'bar', 2, 'bar'] + * a.index {|element| element == 'bar' } # => 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 @@ -1193,69 +2053,85 @@ rb_ary_index(int argc, VALUE *argv, VALUE ary) if (argc == 0) { RETURN_ENUMERATOR(ary, 0, 0); for (i=0; i<RARRAY_LEN(ary); i++) { - if (RTEST(rb_yield(RARRAY_PTR(ary)[i]))) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { return LONG2NUM(i); } } return Qnil; } - rb_scan_args(argc, argv, "1", &val); + rb_check_arity(argc, 0, 1); + val = 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_PTR(ary)[i], val)) + 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 -> an_enumerator - * - * Returns the index of the last object in +self+ - * <code>==</code> to <i>obj</i>. If a block is given instead of an - * argument, returns index of first object for which <em>block</em> is - * true, starting from the last object. - * Returns <code>nil</code> if no match is found. - * See also <code>Array#index</code>. - * - * If neither block nor argument is given, an enumerator is returned instead. - * - * a = [ "a", "b", "b", "b", "c" ] - * a.rindex("b") #=> 3 - * a.rindex("z") #=> nil - * a.rindex { |x| x == "b" } #=> 3 + * array.rindex(object) -> integer or nil + * array.rindex {|element| ... } -> integer or nil + * array.rindex -> new_enumerator + * + * Returns the index of the last element for which <tt>object == element</tt>. + * + * When argument +object+ is given but no block, returns the index of the last such element found: + * a = [:foo, 'bar', 2, 'bar'] + * a.rindex('bar') # => 3 + * + * Returns +nil+ if no such object found. + * + * 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 = [: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 rb_ary_rindex(int argc, VALUE *argv, VALUE ary) { VALUE val; - long i = RARRAY_LEN(ary); + long i = RARRAY_LEN(ary), len; if (argc == 0) { RETURN_ENUMERATOR(ary, 0, 0); while (i--) { - if (RTEST(rb_yield(RARRAY_PTR(ary)[i]))) + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) return LONG2NUM(i); - if (i > RARRAY_LEN(ary)) { - i = RARRAY_LEN(ary); + if (i > (len = RARRAY_LEN(ary))) { + i = len; } } return Qnil; } - rb_scan_args(argc, argv, "1", &val); + rb_check_arity(argc, 0, 1); + val = argv[0]; if (rb_block_given_p()) rb_warn("given block not used"); while (i--) { - if (rb_equal(RARRAY_PTR(ary)[i], val)) + VALUE e = RARRAY_AREF(ary, i); + if (rb_equal(e, val)) { return LONG2NUM(i); - if (i > RARRAY_LEN(ary)) { - i = RARRAY_LEN(ary); } + if (i > RARRAY_LEN(ary)) { + break; + } } return Qnil; } @@ -1270,71 +2146,90 @@ rb_ary_to_ary(VALUE obj) } static void -rb_ary_splice(VALUE ary, long beg, long len, VALUE rpl) +rb_ary_splice(VALUE ary, long beg, long len, const VALUE *rptr, long rlen) { - long rlen; + long olen; + long rofs; if (len < 0) rb_raise(rb_eIndexError, "negative length (%ld)", len); + olen = RARRAY_LEN(ary); if (beg < 0) { - beg += RARRAY_LEN(ary); + beg += olen; if (beg < 0) { rb_raise(rb_eIndexError, "index %ld too small for array; minimum: %ld", - beg - RARRAY_LEN(ary), -RARRAY_LEN(ary)); + beg - olen, -olen); } } - if (RARRAY_LEN(ary) < len || RARRAY_LEN(ary) < beg + len) { - len = RARRAY_LEN(ary) - beg; + if (olen < len || olen < beg + len) { + len = olen - beg; } - if (rpl == Qundef) { - rlen = 0; - } - else { - rpl = rb_ary_to_ary(rpl); - rlen = RARRAY_LEN(rpl); + { + const VALUE *optr = RARRAY_CONST_PTR_TRANSIENT(ary); + rofs = (rptr >= optr && rptr < optr + olen) ? rptr - optr : -1; } - rb_ary_modify(ary); - if (beg >= RARRAY_LEN(ary)) { + + 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; - if (len >= ARY_CAPA(ary)) { - ary_double_capa(ary, len); - } - rb_mem_clear(RARRAY_PTR(ary) + RARRAY_LEN(ary), beg - RARRAY_LEN(ary)); + ary_mem_clear(ary, olen, beg - olen); if (rlen > 0) { - MEMCPY(RARRAY_PTR(ary) + beg, RARRAY_PTR(rpl), VALUE, rlen); + if (rofs != -1) rptr = RARRAY_CONST_PTR_TRANSIENT(ary) + rofs; + ary_memcpy0(ary, beg, rlen, rptr, target_ary); } ARY_SET_LEN(ary, len); } else { long alen; - alen = RARRAY_LEN(ary) + rlen - len; + 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) { - MEMMOVE(RARRAY_PTR(ary) + beg + rlen, RARRAY_PTR(ary) + beg + len, - VALUE, RARRAY_LEN(ary) - (beg + len)); + 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) { - MEMMOVE(RARRAY_PTR(ary) + beg, RARRAY_PTR(rpl), VALUE, rlen); + if (rofs != -1) rptr = RARRAY_CONST_PTR_TRANSIENT(ary) + rofs; + /* give up wb-protected ary */ + RB_OBJ_WB_UNPROTECT_FOR(ARRAY, ary); + + /* do not use RARRAY_PTR() because it can causes GC. + * ary can contain T_NONE object because it is not cleared. + */ + RARRAY_PTR_USE_TRANSIENT(ary, ptr, + MEMMOVE(ptr + beg, rptr, VALUE, rlen)); } } } -/*! - * 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. - */ +void +rb_ary_set_len(VALUE ary, long len) +{ + long capa; + + rb_ary_modify_check(ary); + if (ARY_SHARED_P(ary)) { + 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); + } + ARY_SET_LEN(ary, len); +} + VALUE rb_ary_resize(VALUE ary, long len) { @@ -1350,8 +2245,8 @@ rb_ary_resize(VALUE ary, long len) if (len >= ARY_CAPA(ary)) { ary_double_capa(ary, len); } - rb_mem_clear(RARRAY_PTR(ary) + olen, len - olen); - ARY_SET_LEN(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); @@ -1360,45 +2255,133 @@ rb_ary_resize(VALUE ary, long len) VALUE tmp[RARRAY_EMBED_LEN_MAX]; MEMCPY(tmp, ARY_HEAP_PTR(ary), VALUE, len); ary_discard(ary); - MEMCPY(ARY_EMBED_PTR(ary), tmp, VALUE, len); + MEMCPY((VALUE *)ARY_EMBED_PTR(ary), tmp, VALUE, len); /* WB: no new reference */ ARY_SET_EMBED_LEN(ary, len); } else { if (olen > len + ARY_DEFAULT_SIZE) { - REALLOC_N(RARRAY(ary)->as.heap.ptr, VALUE, len); - ARY_SET_CAPA(ary, len); + 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 starting at _start_ and - * continuing for _length_ elements, or replaces a subarray - * specified by _range_. If indices are greater than - * the current capacity of the array, the array grows - * automatically. A negative indices will count backward - * from the end of the array. Inserts elements if _length_ is - * zero. An +IndexError+ is raised if a negative index points - * past the beginning of the array. See also - * <code>Array#push</code>, and <code>Array#unshift</code>. - * - * 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"] + * 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 @@ -1406,43 +2389,55 @@ rb_ary_aset(int argc, VALUE *argv, VALUE ary) { long offset, beg, len; + 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]); - rb_ary_splice(ary, beg, len, argv[2]); - return argv[2]; + return ary_aset_by_rb_ary_splice(ary, beg, len, argv[2]); } - if (argc != 2) { - rb_raise(rb_eArgError, "wrong number of arguments (%d for 2)", argc); - } - rb_ary_modify_check(ary); if (FIXNUM_P(argv[0])) { offset = FIX2LONG(argv[0]); - goto fixnum; + 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 */ - rb_ary_splice(ary, beg, len, argv[1]); - return argv[1]; + 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 - * - * Inserts the given values before the element with the given index - * (which may be negative). - * - * 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"] + * 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: + * a = [:foo, 'bar', 2] + * a.insert(1) + * a.insert(50) + * a.insert(-50) + * a # => [:foo, "bar", 2] + * + * 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] */ static VALUE @@ -1450,77 +2445,129 @@ rb_ary_insert(int argc, VALUE *argv, VALUE ary) { long pos; - if (argc < 1) { - rb_raise(rb_eArgError, "wrong number of arguments (at least 1)"); - } + rb_check_arity(argc, 1, UNLIMITED_ARGUMENTS); rb_ary_modify_check(ary); - if (argc == 1) return ary; pos = NUM2LONG(argv[0]); + if (argc == 1) return ary; if (pos == -1) { pos = RARRAY_LEN(ary); } - if (pos < 0) { + 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++; } - rb_ary_splice(ary, pos, 0, rb_ary_new4(argc - 1, argv + 1)); + rb_ary_splice(ary, pos, 0, argv + 1, argc - 1); return ary; } +static VALUE +rb_ary_length(VALUE ary); + +static VALUE +ary_enum_length(VALUE ary, VALUE args, VALUE eobj) +{ + return rb_ary_length(ary); +} + /* * call-seq: - * ary.each {|item| block } -> ary - * ary.each -> an_enumerator + * array.each {|element| ... } -> self + * array.each -> Enumerator + * + * Iterates over array elements. * - * Calls <i>block</i> once for each element in +self+, passing that - * element as a parameter. + * 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}" } * - * If no block is given, an enumerator is returned instead. + * Output: + * Symbol foo + * String bar + * Integer 2 * - * a = [ "a", "b", "c" ] - * a.each {|x| print x, " -- " } + * 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') } * - * produces: + * Output: + * foo + * bar * - * a -- b -- c -- + * When no block given, returns a new \Enumerator: + * a = [:foo, 'bar', 2] + * e = a.each + * e # => #<Enumerator: [:foo, "bar", 2]:each> + * a1 = e.each {|element| puts "#{element.class} #{element}" } + * + * Output: + * Symbol foo + * String bar + * Integer 2 + * + * Related: #each_index, #reverse_each. */ VALUE -rb_ary_each(VALUE array) +rb_ary_each(VALUE ary) { long i; - volatile VALUE ary = array; - - RETURN_ENUMERATOR(ary, 0, 0); + ary_verify(ary); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); for (i=0; i<RARRAY_LEN(ary); i++) { - rb_yield(RARRAY_PTR(ary)[i]); + rb_yield(RARRAY_AREF(ary, i)); } return ary; } /* * call-seq: - * ary.each_index {|index| block } -> ary - * ary.each_index -> an_enumerator + * array.each_index {|index| ... } -> self + * array.each_index -> Enumerator + * + * Iterates over array indexes. * - * Same as <code>Array#each</code>, but passes the index of the element - * instead of the element itself. + * 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]}" } * - * If no block is given, an enumerator is returned instead. + * Output: + * 0 foo + * 1 bar + * 2 2 * + * Allows the array to be modified during iteration: + * 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: + * 0 + * 1 * - * produces: + * 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]}"} * - * 0 -- 1 -- 2 -- + * Output: + * 0 foo + * 1 bar + * 2 2 + * + * Related: #each, #reverse_each. */ static VALUE rb_ary_each_index(VALUE ary) { long i; - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); for (i=0; i<RARRAY_LEN(ary); i++) { rb_yield(LONG2NUM(i)); @@ -1530,18 +2577,40 @@ rb_ary_each_index(VALUE ary) /* * call-seq: - * ary.reverse_each {|item| block } -> ary - * ary.reverse_each -> an_enumerator - * - * Same as <code>Array#each</code>, but traverses +self+ in reverse - * order. - * - * a = [ "a", "b", "c" ] - * a.reverse_each {|x| print x, " " } - * - * produces: - * - * c b a + * array.reverse_each {|element| ... } -> self + * array.reverse_each -> Enumerator + * + * Iterates backwards over array elements. + * + * When a block given, passes, in reverse order, each element to the block; + * returns +self+: + * a = [:foo, 'bar', 2] + * a.reverse_each {|element| puts "#{element.class} #{element}" } + * + * Output: + * 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 @@ -1549,12 +2618,14 @@ rb_ary_reverse_each(VALUE ary) { long len; - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); len = RARRAY_LEN(ary); while (len--) { - rb_yield(RARRAY_PTR(ary)[len]); - if (RARRAY_LEN(ary) < len) { - len = RARRAY_LEN(ary); + long nlen; + rb_yield(RARRAY_AREF(ary, len)); + nlen = RARRAY_LEN(ary); + if (nlen < len) { + len = nlen; } } return ary; @@ -1562,11 +2633,9 @@ rb_ary_reverse_each(VALUE 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 + * Returns the count of elements in +self+. */ static VALUE @@ -1578,34 +2647,35 @@ rb_ary_length(VALUE ary) /* * call-seq: - * ary.empty? -> true or false + * array.empty? -> true or false * - * Returns <code>true</code> 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 rb_ary_dup(VALUE ary) { - VALUE dup = rb_ary_new2(RARRAY_LEN(ary)); - MEMCPY(RARRAY_PTR(dup), RARRAY_PTR(ary), VALUE, RARRAY_LEN(ary)); - ARY_SET_LEN(dup, RARRAY_LEN(ary)); + long len = RARRAY_LEN(ary); + VALUE dup = rb_ary_new2(len); + ary_memcpy(dup, 0, len, RARRAY_CONST_PTR_TRANSIENT(ary)); + ARY_SET_LEN(dup, len); + + ary_verify(ary); + ary_verify(dup); return dup; } VALUE rb_ary_resurrect(VALUE ary) { - return rb_ary_new4(RARRAY_LEN(ary), RARRAY_PTR(ary)); + return ary_make_partial(ary, rb_cArray, 0, RARRAY_LEN(ary)); } extern VALUE rb_output_fs; @@ -1630,20 +2700,48 @@ recursive_join(VALUE obj, VALUE argp, int recur) return Qnil; } -static void +static long ary_join_0(VALUE ary, VALUE sep, long max, VALUE result) { long i; VALUE val; - if (max > 0) rb_enc_copy(result, RARRAY_PTR(ary)[0]); + if (max > 0) rb_enc_copy(result, RARRAY_AREF(ary, 0)); for (i=0; i<max; i++) { - val = RARRAY_PTR(ary)[i]; + 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); - if (OBJ_TAINTED(val)) OBJ_TAINT(result); - if (OBJ_UNTRUSTED(val)) OBJ_TAINT(result); + } + 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); } } @@ -1656,47 +2754,21 @@ ary_join_1(VALUE obj, VALUE ary, VALUE sep, long i, VALUE result, int *first) if (i > 0 && !NIL_P(sep)) rb_str_buf_append(result, sep); - val = RARRAY_PTR(ary)[i]; - switch (TYPE(val)) { - case T_STRING: - str_join: - rb_str_buf_append(result, val); - *first = FALSE; - break; - case T_ARRAY: - obj = val; - ary_join: - if (val == ary) { - rb_raise(rb_eArgError, "recursive array join"); - } - else { - VALUE args[4]; - - args[0] = val; - args[1] = sep; - args[2] = result; - args[3] = (VALUE)first; - rb_exec_recursive(recursive_join, obj, (VALUE)args); - } - break; - default: - tmp = rb_check_string_type(val); - if (!NIL_P(tmp)) { - val = tmp; - goto str_join; - } - tmp = rb_check_convert_type(val, T_ARRAY, "Array", "to_ary"); - if (!NIL_P(tmp)) { - obj = val; - val = tmp; - goto ary_join; - } - val = rb_obj_as_string(val); - if (*first) { - rb_enc_copy(result, val); - *first = FALSE; - } - goto str_join; + 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); } } } @@ -1705,29 +2777,25 @@ VALUE rb_ary_join(VALUE ary, VALUE sep) { long len = 1, i; - int taint = FALSE; - int untrust = FALSE; VALUE val, tmp, result; if (RARRAY_LEN(ary) == 0) return rb_usascii_str_new(0, 0); - if (OBJ_TAINTED(ary) || OBJ_TAINTED(sep)) taint = TRUE; - if (OBJ_UNTRUSTED(ary) || OBJ_UNTRUSTED(sep)) untrust = TRUE; if (!NIL_P(sep)) { StringValue(sep); len += RSTRING_LEN(sep) * (RARRAY_LEN(ary) - 1); } for (i=0; i<RARRAY_LEN(ary); i++) { - val = RARRAY_PTR(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); + 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()); - if (taint) OBJ_TAINT(result); - if (untrust) OBJ_UNTRUST(result); - ary_join_0(ary, sep, i, result); + i = ary_join_0(ary, sep, i, result); first = i == 0; ary_join_1(ary, ary, sep, i, result, &first); return result; @@ -1736,9 +2804,9 @@ rb_ary_join(VALUE ary, VALUE sep) len += RSTRING_LEN(tmp); } - result = rb_str_buf_new(len); - if (taint) OBJ_TAINT(result); - if (untrust) OBJ_UNTRUST(result); + result = rb_str_new(0, len); + rb_str_set_len(result, 0); + ary_join_0(ary, sep, RARRAY_LEN(ary), result); return result; @@ -1746,22 +2814,38 @@ rb_ary_join(VALUE ary, VALUE sep) /* * call-seq: - * ary.join(sep=$,) -> str - * - * Returns a string created by converting each element of the array to - * a string, separated by <i>sep</i>. - * - * [ "a", "b", "c" ].join #=> "abc" - * [ "a", "b", "c" ].join("-") #=> "a-b-c" + * array.join ->new_string + * array.join(separator = $,) -> new_string + * + * Returns the new \String formed by joining the array elements after conversion. + * For each element +element+ + * - Uses <tt>element.to_s</tt> if +element+ is not a <tt>kind_of?(Array)</tt>. + * - Uses recursive <tt>element.join(separator)</tt> if +element+ is a <tt>kind_of?(Array)</tt>. + * + * With no argument, joins using the output field separator, <tt>$,</tt>: + * a = [:foo, 'bar', 2] + * $, # => nil + * a.join # => "foobar2" + * + * With \string argument +separator+, joins using that separator: + * a = [:foo, 'bar', 2] + * a.join("\n") # => "foo\nbar\n2" + * + * Joins recursively for nested Arrays: + * a = [:foo, [:bar, [:baz, :bat]]] + * a.join # => "foobarbazbat" */ - static VALUE rb_ary_join_m(int argc, VALUE *argv, VALUE ary) { VALUE sep; - rb_scan_args(argc, argv, "01", &sep); - if (NIL_P(sep)) sep = rb_output_fs; + if (rb_check_arity(argc, 0, 1) == 0 || NIL_P(sep = argv[0])) { + sep = rb_output_fs; + if (!NIL_P(sep)) { + rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "$, is set to non-nil value"); + } + } return rb_ary_join(ary, sep); } @@ -1769,33 +2853,31 @@ 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); - int untrust = OBJ_UNTRUSTED(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_PTR(ary)[i]); - if (OBJ_TAINTED(s)) tainted = TRUE; - if (OBJ_UNTRUSTED(s)) untrust = TRUE; + 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); - if (untrust) OBJ_UNTRUST(str); return str; } /* * call-seq: - * ary.to_s -> string - * ary.inspect -> 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 = [:foo, 'bar', 2] + * a.inspect # => "[:foo, \"bar\", 2]" + * + * Array#to_s is an alias for Array#inspect. */ static VALUE @@ -1813,10 +2895,20 @@ rb_ary_to_s(VALUE ary) /* * call-seq: - * ary.to_a -> ary - * - * Returns +self+. If called on a subclass of Array, converts - * the receiver to an Array object. + * to_a -> self or new_array + * + * 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 */ static VALUE @@ -1832,7 +2924,53 @@ rb_ary_to_a(VALUE ary) /* * call-seq: - * ary.to_ary -> ary + * array.to_h -> new_hash + * array.to_h {|item| ... } -> new_hash + * + * Returns a new \Hash formed from +self+. + * + * 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: + * 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"} + */ + +static VALUE +rb_ary_to_h(VALUE ary) +{ + long i; + VALUE hash = rb_hash_new_with_size(RARRAY_LEN(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)); + } + return hash; +} + +/* + * call-seq: + * array.to_ary -> self * * Returns +self+. */ @@ -1844,8 +2982,7 @@ rb_ary_to_ary_m(VALUE ary) } static void -ary_reverse(p1, p2) - VALUE *p1, *p2; +ary_reverse(VALUE *p1, VALUE *p2) { while (p1 < p2) { VALUE tmp = *p1; @@ -1857,26 +2994,26 @@ ary_reverse(p1, p2) VALUE rb_ary_reverse(VALUE ary) { - VALUE *p1, *p2; + VALUE *p2; + long len = RARRAY_LEN(ary); rb_ary_modify(ary); - if (RARRAY_LEN(ary) > 1) { - p1 = RARRAY_PTR(ary); - p2 = p1 + RARRAY_LEN(ary) - 1; /* points last item */ - ary_reverse(p1, p2); + if (len > 1) { + RARRAY_PTR_USE_TRANSIENT(ary, p1, { + p2 = p1 + len - 1; /* points last item */ + ary_reverse(p1, p2); + }); /* WB: no new reference */ } return ary; } /* * call-seq: - * ary.reverse! -> ary - * - * Reverses +self+ in place. + * array.reverse! -> self * - * a = [ "a", "b", "c" ] - * a.reverse! #=> ["c", "b", "a"] - * a #=> ["c", "b", "a"] + * Reverses +self+ in place: + * a = ['foo', 'bar', 'two'] + * a.reverse! # => ["two", "bar", "foo"] */ static VALUE @@ -1887,12 +3024,12 @@ rb_ary_reverse_bang(VALUE ary) /* * call-seq: - * ary.reverse -> new_ary - * - * Returns a new array containing +self+'s elements in reverse order. + * array.reverse -> new_array * - * [ "a", "b", "c" ].reverse #=> ["c", "b", "a"] - * [ 1 ].reverse #=> [1] + * Returns a new \Array with the elements of +self+ in reverse order. + * a = ['foo', 'bar', 'two'] + * a1 = a.reverse + * a1 # => ["two", "bar", "foo"] */ static VALUE @@ -1902,8 +3039,8 @@ rb_ary_reverse_m(VALUE ary) VALUE dup = rb_ary_new2(len); if (len > 0) { - VALUE *p1 = RARRAY_PTR(ary); - VALUE *p2 = RARRAY_PTR(dup) + len - 1; + const VALUE *p1 = RARRAY_CONST_PTR_TRANSIENT(ary); + VALUE *p2 = (VALUE *)RARRAY_CONST_PTR_TRANSIENT(dup) + len - 1; do *p2-- = *p1++; while (--len > 0); } ARY_SET_LEN(dup, RARRAY_LEN(ary)); @@ -1916,92 +3053,147 @@ rotate_count(long cnt, long len) return (cnt < 0) ? (len - (~cnt % len) - 1) : (cnt % len); } +static void +ary_rotate_ptr(VALUE *ptr, long len, long cnt) +{ + 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 rb_ary_rotate(VALUE ary, long cnt) { rb_ary_modify(ary); if (cnt != 0) { - VALUE *ptr = RARRAY_PTR(ary); - long len = RARRAY_LEN(ary); - - if (len > 0 && (cnt = rotate_count(cnt, len)) > 0) { - --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); - return ary; - } + long len = RARRAY_LEN(ary); + if (len > 1 && (cnt = rotate_count(cnt, len)) > 0) { + RARRAY_PTR_USE_TRANSIENT(ary, ptr, ary_rotate_ptr(ptr, len, cnt)); + return ary; + } } - return Qnil; } /* * call-seq: - * ary.rotate!(cnt=1) -> ary - * - * Rotates +self+ in place so that the element at +cnt+ comes first, - * and returns +self+. If +cnt+ is negative then it rotates in - * the opposite direction. - * - * 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"] + * array.rotate! -> self + * array.rotate!(count) -> 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: + * 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] */ static VALUE rb_ary_rotate_bang(int argc, VALUE *argv, VALUE ary) { - long n = 1; - - switch (argc) { - case 1: n = NUM2LONG(argv[0]); - case 0: break; - default: rb_scan_args(argc, argv, "01", NULL); - } + long n = (rb_check_arity(argc, 0, 1) ? NUM2LONG(argv[0]) : 1); rb_ary_rotate(ary, n); return ary; } /* * call-seq: - * ary.rotate(cnt=1) -> new_ary - * - * Returns new array by rotating +self+ so that the element at - * +cnt+ in +self+ is the first element of the new array. If +cnt+ - * is negative then it rotates in the opposite direction. - * - * 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"] + * 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: + * a = [:foo, 'bar', 2] + * a1 = a.rotate(2) + * a1 # => [2, :foo, "bar"] + * + * 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] */ static VALUE rb_ary_rotate_m(int argc, VALUE *argv, VALUE ary) { - VALUE rotated, *ptr, *ptr2; - long len, cnt = 1; - - switch (argc) { - case 1: cnt = NUM2LONG(argv[0]); - case 0: break; - default: rb_scan_args(argc, argv, "01", NULL); - } + VALUE rotated; + const VALUE *ptr; + long len; + long cnt = (rb_check_arity(argc, 0, 1) ? NUM2LONG(argv[0]) : 1); len = RARRAY_LEN(ary); rotated = rb_ary_new2(len); if (len > 0) { cnt = rotate_count(cnt, len); - ptr = RARRAY_PTR(ary); - ptr2 = RARRAY_PTR(rotated); + ptr = RARRAY_CONST_PTR_TRANSIENT(ary); len -= cnt; - MEMCPY(ptr2, ptr + cnt, VALUE, len); - MEMCPY(ptr2 + len, ptr, VALUE, cnt); + ary_memcpy(rotated, 0, len, ptr + cnt); + ary_memcpy(rotated, len, cnt, ptr); } ARY_SET_LEN(rotated, RARRAY_LEN(ary)); return rotated; @@ -2009,26 +3201,10 @@ rb_ary_rotate_m(int argc, VALUE *argv, VALUE ary) struct ary_sort_data { VALUE ary; - int opt_methods; - int opt_inited; + VALUE receiver; + struct cmp_opt_data cmp_opt; }; -enum { - sort_opt_Fixnum, - sort_opt_String, - sort_optimizable_count -}; - -#define STRING_P(s) (TYPE(s) == T_STRING && CLASS_OF(s) == rb_cString) - -#define SORT_OPTIMIZABLE_BIT(type) (1U << TOKEN_PASTE(sort_opt_,type)) -#define SORT_OPTIMIZABLE(data, type) \ - (((data)->opt_inited & SORT_OPTIMIZABLE_BIT(type)) ? \ - ((data)->opt_methods & SORT_OPTIMIZABLE_BIT(type)) : \ - (((data)->opt_inited |= SORT_OPTIMIZABLE_BIT(type)), \ - rb_method_basic_definition_p(TOKEN_PASTE(rb_c,type), id_cmp) && \ - ((data)->opt_methods |= SORT_OPTIMIZABLE_BIT(type)))) - static VALUE sort_reentered(VALUE ary) { @@ -2038,17 +3214,29 @@ sort_reentered(VALUE ary) 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) { struct ary_sort_data *data = dummy; VALUE retval = sort_reentered(data->ary); VALUE a = *(const VALUE *)ap, b = *(const VALUE *)bp; + VALUE args[2]; int n; - retval = rb_yield_values(2, a, b); + args[0] = a; + args[1] = b; + retval = rb_yield_values2(2, args); n = rb_cmpint(retval, a, b); - sort_reentered(data->ary); + sort_returned(data); return n; } @@ -2060,36 +3248,59 @@ 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) && SORT_OPTIMIZABLE(data, Fixnum)) { + if (FIXNUM_P(a) && FIXNUM_P(b) && CMP_OPTIMIZABLE(data->cmp_opt, Integer)) { if ((long)a > (long)b) return 1; if ((long)a < (long)b) return -1; return 0; } - if (STRING_P(a) && STRING_P(b) && SORT_OPTIMIZABLE(data, String)) { + if (STRING_P(a) && STRING_P(b) && CMP_OPTIMIZABLE(data->cmp_opt, 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); + } - retval = rb_funcall(a, id_cmp, 1, 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 - * - * Sorts +self+. Comparisons for - * the sort will be done using the <code><=></code> operator or using - * an optional code block. The block implements a comparison between - * <i>a</i> and <i>b</i>, returning -1, 0, or +1. See also - * <code>Enumerable#sort_by</code>. - * - * a = [ "d", "a", "e", "c", "b" ] - * a.sort! #=> ["a", "b", "c", "d", "e"] - * a.sort! {|x,y| y <=> x } #=> ["e", "d", "c", "b", "a"] + * array.sort! -> self + * array.sort! {|a, b| ... } -> self + * + * Returns +self+ with its elements sorted in place. + * + * With no block, compares elements using operator <tt><=></tt> + * (see Comparable): + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a.sort! + * a # => ["a", "b", "c", "d", "e"] + * + * With a block, calls the block with each element pair; + * for each element pair +a+ and +b+, the block should return an integer: + * - Negative when +b+ is to follow +a+. + * - Zero when +a+ and +b+ are equivalent. + * - Positive when +a+ is to follow +b+. + * + * 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"] */ VALUE @@ -2100,29 +3311,29 @@ rb_ary_sort_bang(VALUE ary) if (RARRAY_LEN(ary) > 1) { VALUE tmp = ary_make_substitution(ary); /* only ary refers tmp */ struct ary_sort_data data; - - RBASIC(tmp)->klass = 0; + long len = RARRAY_LEN(ary); + RBASIC_CLEAR_CLASS(tmp); data.ary = tmp; - data.opt_methods = 0; - data.opt_inited = 0; - ruby_qsort(RARRAY_PTR(tmp), RARRAY_LEN(tmp), sizeof(VALUE), - rb_block_given_p()?sort_1:sort_2, &data); - + data.receiver = ary; + data.cmp_opt.opt_methods = 0; + data.cmp_opt.opt_inited = 0; + 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); if (ARY_EMBED_P(tmp)) { - assert(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); - MEMCPY(RARRAY_PTR(ary), ARY_EMBED_PTR(tmp), VALUE, ARY_EMBED_LEN(tmp)); + ary_memcpy(ary, 0, ARY_EMBED_LEN(tmp), ARY_EMBED_PTR(tmp)); ARY_SET_LEN(ary, ARY_EMBED_LEN(tmp)); } else { - assert(!ARY_EMBED_P(tmp)); - if (ARY_HEAP_PTR(ary) == ARY_HEAP_PTR(tmp)) { - assert(!ARY_EMBED_P(ary)); + if (!ARY_EMBED_P(ary) && ARY_HEAP_PTR(ary) == ARY_HEAP_PTR(tmp)) { FL_UNSET_SHARED(ary); - ARY_SET_CAPA(ary, ARY_CAPA(tmp)); + ARY_SET_CAPA(ary, RARRAY_LEN(tmp)); } else { assert(!ARY_SHARED_P(tmp)); @@ -2134,38 +3345,61 @@ rb_ary_sort_bang(VALUE ary) rb_ary_unshare(ary); } else { - xfree(ARY_HEAP_PTR(ary)); + ary_heap_free(ary); } - ARY_SET_PTR(ary, RARRAY_PTR(tmp)); - ARY_SET_HEAP_LEN(ary, RARRAY_LEN(tmp)); - ARY_SET_CAPA(ary, ARY_CAPA(tmp)); + ARY_SET_PTR(ary, ARY_HEAP_PTR(tmp)); + ARY_SET_HEAP_LEN(ary, len); + ARY_SET_CAPA(ary, ARY_HEAP_LEN(tmp)); } /* tmp was lost ownership for the ptr */ FL_UNSET(tmp, FL_FREEZE); FL_SET_EMBED(tmp); ARY_SET_EMBED_LEN(tmp, 0); FL_SET(tmp, FL_FREEZE); - } + } /* tmp will be GC'ed. */ - RBASIC(tmp)->klass = rb_cArray; + RBASIC_SET_CLASS_RAW(tmp, rb_cArray); /* rb_cArray must be marked */ } + ary_verify(ary); return ary; } /* * call-seq: - * ary.sort -> new_ary - * ary.sort {| a,b | block } -> new_ary - * - * Returns a new array created by sorting +self+. Comparisons for - * the sort will be done using the <code><=></code> operator or using - * an optional code block. The block implements a comparison between - * <i>a</i> and <i>b</i>, returning -1, 0, or +1. See also - * <code>Enumerable#sort_by</code>. - * - * a = [ "d", "a", "e", "c", "b" ] - * a.sort #=> ["a", "b", "c", "d", "e"] - * a.sort {|x,y| y <=> x } #=> ["e", "d", "c", "b", "a"] + * array.sort -> new_array + * array.sort {|a, b| ... } -> new_array + * + * Returns a new \Array whose elements are those from +self+, sorted. + * + * With no block, compares elements using operator <tt><=></tt> + * (see Comparable): + * a = 'abcde'.split('').shuffle + * a # => ["e", "b", "d", "a", "c"] + * a1 = a.sort + * a1 # => ["a", "b", "c", "d", "e"] + * + * With a block, calls the block with each element pair; + * for each element pair +a+ and +b+, the block should return an integer: + * - Negative when +b+ is to follow +a+. + * - Zero when +a+ and +b+ are equivalent. + * - Positive when +a+ is to follow +b+. + * + * Example: + * 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"] + * + * 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 @@ -2176,23 +3410,114 @@ rb_ary_sort(VALUE ary) return ary; } +static VALUE rb_ary_bsearch_index(VALUE ary); + +/* + * call-seq: + * 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 +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 index_result; +} + +/* + * call-seq: + * array.bsearch_index {|element| ... } -> integer or nil + * array.bsearch_index -> new_enumerator + * + * Searches +self+ as described at method #bsearch, + * but returns the _index_ of the found element instead of the element itself. + */ + +static VALUE +rb_ary_bsearch_index(VALUE ary) +{ + long low = 0, high = RARRAY_LEN(ary), mid; + int smaller = 0, satisfied = 0; + VALUE v, val; + + 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 (!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); +} + static VALUE -sort_by_i(VALUE i) +sort_by_i(RB_BLOCK_CALL_FUNC_ARGLIST(i, dummy)) { return rb_yield(i); } /* * call-seq: - * ary.sort_by! {| obj | block } -> ary - * ary.sort_by! -> an_enumerator + * array.sort_by! {|element| ... } -> self + * array.sort_by! -> new_enumerator * - * Sorts +self+ in place using a set of keys generated by mapping the - * values in +self+ through the given block. + * Sorts the elements of +self+ in place, + * using an ordering determined by the block; returns self. * - * If no block is given, an enumerator is returned instead. + * Calls the block with each successive element; + * sorts elements based on the values returned from the block. * + * For duplicates returned by the block, the ordering is indeterminate, and may be unstable. + * + * 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!> */ static VALUE @@ -2200,7 +3525,7 @@ rb_ary_sort_by_bang(VALUE ary) { VALUE sorted; - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); rb_ary_modify(ary); sorted = rb_block_call(ary, rb_intern("sort_by"), 0, 0, sort_by_i, 0); rb_ary_replace(ary, sorted); @@ -2210,20 +3535,21 @@ rb_ary_sort_by_bang(VALUE ary) /* * call-seq: - * ary.collect {|item| block } -> new_ary - * ary.map {|item| block } -> new_ary - * ary.collect -> an_enumerator - * ary.map -> an_enumerator + * array.map {|element| ... } -> new_array + * array.map -> new_enumerator * - * Invokes <i>block</i> once for each element of +self+. Creates a - * new array containing the values returned by the block. - * See also <code>Enumerable#collect</code>. + * Calls the block, if given, with each element of +self+; + * returns a new \Array whose elements are the return values from the block: + * a = [:foo, 'bar', 2] + * a1 = a.map {|element| element.class } + * a1 # => [Symbol, String, Integer] * - * If no block is given, an enumerator is returned instead. + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2] + * a1 = a.map + * a1 # => #<Enumerator: [:foo, "bar", 2]:map> * - * a = [ "a", "b", "c", "d" ] - * a.collect {|x| x + "!" } #=> ["a!", "b!", "c!", "d!"] - * a #=> ["a", "b", "c", "d"] + * Array#collect is an alias for Array#map. */ static VALUE @@ -2232,10 +3558,10 @@ rb_ary_collect(VALUE ary) long i; VALUE collect; - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); collect = rb_ary_new2(RARRAY_LEN(ary)); for (i = 0; i < RARRAY_LEN(ary); i++) { - rb_ary_push(collect, rb_yield(RARRAY_PTR(ary)[i])); + rb_ary_push(collect, rb_yield(RARRAY_AREF(ary, i))); } return collect; } @@ -2243,20 +3569,20 @@ rb_ary_collect(VALUE ary) /* * call-seq: - * ary.collect! {|item| block } -> ary - * ary.map! {|item| block } -> ary - * ary.collect -> an_enumerator - * ary.map -> an_enumerator + * array.map! {|element| ... } -> self + * array.map! -> new_enumerator * - * Invokes the block once for each element of +self+, replacing the - * element with the value returned by _block_. - * See also <code>Enumerable#collect</code>. + * Calls the block, if given, with each element; + * replaces the element with the block's return value: + * a = [:foo, 'bar', 2] + * a.map! { |element| element.class } # => [Symbol, String, Integer] * - * If no block is given, an enumerator is returned instead. + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2] + * a1 = a.map! + * a1 # => #<Enumerator: [:foo, "bar", 2]:map!> * - * a = [ "a", "b", "c", "d" ] - * a.collect! {|x| x + "!" } - * a #=> [ "a!", "b!", "c!", "d!" ] + * Array#collect! is an alias for Array#map!. */ static VALUE @@ -2264,16 +3590,16 @@ rb_ary_collect_bang(VALUE ary) { long i; - RETURN_ENUMERATOR(ary, 0, 0); + 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_PTR(ary)[i])); + rb_ary_store(ary, i, rb_yield(RARRAY_AREF(ary, i))); } return ary; } VALUE -rb_get_values_at(VALUE obj, long olen, int argc, VALUE *argv, VALUE (*func) (VALUE, long)) +rb_get_values_at(VALUE obj, long olen, int argc, const VALUE *argv, VALUE (*func) (VALUE, long)) { VALUE result = rb_ary_new2(argc); long beg, len, i, j; @@ -2284,15 +3610,13 @@ rb_get_values_at(VALUE obj, long olen, int argc, VALUE *argv, VALUE (*func) (VAL continue; } /* check if idx is Range */ - switch (rb_range_beg_len(argv[i], &beg, &len, olen, 0)) { - case Qfalse: - break; - case Qnil: - continue; - default: - for (j=0; j<len; j++) { - rb_ary_push(result, (*func)(obj, j+beg)); + 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]))); @@ -2300,42 +3624,100 @@ rb_get_values_at(VALUE obj, long olen, int argc, VALUE *argv, VALUE (*func) (VAL return result; } +static VALUE +append_values_at_single(VALUE result, VALUE ary, long olen, VALUE idx) +{ + long beg, len; + if (FIXNUM_P(idx)) { + beg = FIX2LONG(idx); + } + /* check if idx is Range */ + else if (rb_range_beg_len(idx, &beg, &len, olen, 1)) { + 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; + } + else { + beg = NUM2LONG(idx); + } + return rb_ary_push(result, rb_ary_entry(ary, beg)); +} + /* * call-seq: - * ary.values_at(selector,... ) -> new_ary - * - * Returns an array containing the elements in - * +self+ corresponding to the given selector(s). The selectors - * may be either integer indices or ranges. - * See also <code>Array#select</code>. - * - * a = %w{ a b c d e f } - * a.values_at(1, 3, 5) - * a.values_at(1, 3, 5, 7) - * a.values_at(-1, -3, -5, -7) - * a.values_at(1..3, 2...5) + * 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: + * 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] + * + * Assigns +nil+ for an +index+ that is too large: + * 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] */ static VALUE rb_ary_values_at(int argc, VALUE *argv, VALUE ary) { - return rb_get_values_at(ary, RARRAY_LEN(ary), argc, argv, rb_ary_entry); + 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]); + } + RB_GC_GUARD(ary); + return result; } /* * call-seq: - * ary.select {|item| block } -> new_ary - * ary.select -> an_enumerator + * array.select {|element| ... } -> new_array + * array.select -> new_enumerator * - * Invokes the block passing in successive elements from +self+, - * returning an array containing those elements for which the block - * returns a true value (equivalent to <code>Enumerable#select</code>). + * 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: + * a = [:foo, 'bar', 2, :bam] + * a1 = a.select {|element| element.to_s.start_with?('b') } + * a1 # => ["bar", :bam] * - * If no block is given, an enumerator is returned instead. + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2, :bam] + * a.select # => #<Enumerator: [:foo, "bar", 2, :bam]:select> * - * a = %w{ a b c d e f } - * a.select {|v| v =~ /[aeiou]/} #=> ["a", "e"] + * Array#filter is an alias for Array#select. */ static VALUE @@ -2344,93 +3726,161 @@ rb_ary_select(VALUE ary) VALUE result; long i; - RETURN_ENUMERATOR(ary, 0, 0); + 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_PTR(ary)[i]))) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { rb_ary_push(result, rb_ary_elt(ary, i)); } } return result; } -/* - * call-seq: - * ary.select! {|item| block } -> ary or nil - * ary.select! -> an_enumerator - * - * Invokes the block passing in successive elements from - * +self+, deleting elements for which the block returns a - * false value. It returns +self+ if changes were made, - * otherwise it returns <code>nil</code>. - * See also <code>Array#keep_if</code> - * - * If no block is given, an enumerator is returned instead. - * - */ +struct select_bang_arg { + VALUE ary; + long len[2]; +}; static VALUE -rb_ary_select_bang(VALUE ary) +select_bang_i(VALUE a) { + volatile struct select_bang_arg *arg = (void *)a; + VALUE ary = arg->ary; long i1, i2; - RETURN_ENUMERATOR(ary, 0, 0); - rb_ary_modify(ary); - for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); i1++) { - VALUE v = RARRAY_PTR(ary)[i1]; + 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); } - i2++; + arg->len[1] = ++i2; } + return (i1 == i2) ? Qnil : ary; +} - if (RARRAY_LEN(ary) == i2) return Qnil; - if (i2 < RARRAY_LEN(ary)) - ARY_SET_LEN(ary, i2); +static VALUE +select_bang_ensure(VALUE a) +{ + volatile struct select_bang_arg *arg = (void *)a; + VALUE ary = arg->ary; + long len = RARRAY_LEN(ary); + long i1 = arg->len[0], i2 = arg->len[1]; + + if (i2 < len && i2 < 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); + } return ary; } /* * call-seq: - * ary.keep_if {|item| block } -> ary - * ary.keep_if -> an_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+. * - * Deletes every element of +self+ for which <i>block</i> evaluates - * to false. - * See also <code>Array#select!</code> + * Returns +self+ if any elements were removed: + * a = [:foo, 'bar', 2, :bam] + * a.select! {|element| element.to_s.start_with?('b') } # => ["bar", :bam] + * + * Returns +nil+ if no elements were removed. + * + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2, :bam] + * a.select! # => #<Enumerator: [:foo, "bar", 2, :bam]:select!> + * + * Array#filter! is an alias for Array#select!. + */ + +static VALUE +rb_ary_select_bang(VALUE ary) +{ + struct select_bang_arg args; + + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); + rb_ary_modify(ary); + + args.ary = ary; + args.len[0] = args.len[1] = 0; + return rb_ensure(select_bang_i, (VALUE)&args, select_bang_ensure, (VALUE)&args); +} + +/* + * call-seq: + * array.keep_if {|element| ... } -> self + * array.keep_if -> new_enumeration * - * If no block is given, an enumerator is returned instead. + * Retains those elements for which the block returns a truthy value; + * deletes all other elements; returns +self+: + * 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"] + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2, :bam] + * a.keep_if # => #<Enumerator: [:foo, "bar", 2, :bam]:keep_if> */ static VALUE rb_ary_keep_if(VALUE ary) { - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); rb_ary_select_bang(ary); return ary; } +static void +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); + } + } +} + /* * call-seq: - * ary.delete(obj) -> obj or nil - * ary.delete(obj) { block } -> obj or nil - * - * Deletes items from +self+ that are equal to <i>obj</i>. - * If any items are found, returns <i>obj</i>. If - * the item is not found, returns <code>nil</code>. If the optional - * code block is given, returns the result of <i>block</i> if the item - * is not found. (To remove <code>nil</code> elements and - * get an informative return value, use #compact!) - * - * a = [ "a", "b", "b", "b", "c" ] - * a.delete("b") #=> "b" - * a #=> ["a", "c"] - * a.delete("z") #=> nil - * a.delete("z") { "not found" } #=> "not found" + * array.delete(obj) -> deleted_object + * array.delete(obj) {|nosuch| ... } -> deleted_object or block_return + * + * Removes zero or more elements from +self+; returns +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: + * s1 = 'bar'; s2 = 'bar' + * a = [:foo, s1, 2, s2] + * deleted_obj = a.delete('bar') {|obj| fail 'Cannot happen' } + * a # => [:foo, 2] + * + * If no such elements are found, returns the block's return value: + * a = [:foo, 'bar', 2] + * a.delete(:nosuch) {|obj| "#{obj} not found" } # => "nosuch not found" */ VALUE @@ -2440,7 +3890,7 @@ rb_ary_delete(VALUE ary, VALUE item) long i1, i2; for (i1 = i2 = 0; i1 < RARRAY_LEN(ary); i1++) { - VALUE e = RARRAY_PTR(ary)[i1]; + VALUE e = RARRAY_AREF(ary, i1); if (rb_equal(e, item)) { v = e; @@ -2458,16 +3908,33 @@ rb_ary_delete(VALUE ary, VALUE item) return Qnil; } - rb_ary_modify(ary); - if (RARRAY_LEN(ary) > i2) { - ARY_SET_LEN(ary, i2); - if (i2 * 2 < ARY_CAPA(ary) && - ARY_CAPA(ary) > ARY_DEFAULT_SIZE) { - ary_resize_capa(ary, i2*2); + ary_resize_smaller(ary, i2); + + ary_verify(ary); + return v; +} + +void +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); + + if (e == item) { + continue; + } + if (i1 != i2) { + rb_ary_store(ary, i2, e); } + i2++; + } + if (RARRAY_LEN(ary) == i2) { + return; } - return v; + ary_resize_smaller(ary, i2); } VALUE @@ -2483,26 +3950,34 @@ rb_ary_delete_at(VALUE ary, long pos) } rb_ary_modify(ary); - del = RARRAY_PTR(ary)[pos]; - MEMMOVE(RARRAY_PTR(ary)+pos, RARRAY_PTR(ary)+pos+1, VALUE, - RARRAY_LEN(ary)-pos-1); + del = RARRAY_AREF(ary, pos); + RARRAY_PTR_USE_TRANSIENT(ary, ptr, { + MEMMOVE(ptr+pos, ptr+pos+1, VALUE, len-pos-1); + }); ARY_INCREASE_LEN(ary, -1); - + ary_verify(ary); return del; } /* * 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+. * - * Deletes the element at the specified index, returning that element, - * or <code>nil</code> if the index is out of range. See also - * <code>Array#slice!</code>. + * When +index+ is non-negative, deletes the element at offset +index+: + * a = [:foo, 'bar', 2] + * a.delete_at(1) # => "bar" + * a # => [:foo, 2] * - * a = %w( ant bat cat dog ) - * a.delete_at(2) #=> "cat" - * a #=> ["ant", "bat", "dog"] - * a.delete_at(99) #=> nil + * If index is too large, returns +nil+. + * + * When +index+ is negative, counts backward from the end of the array: + * a = [:foo, 'bar', 2] + * a.delete_at(-2) # => "bar" + * a # => [:foo, 2] + * + * If +index+ is too small (far from zero), returns nil. */ static VALUE @@ -2511,64 +3986,118 @@ 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 with a length) - * or by a range. Returns the deleted object (or objects), or - * <code>nil</code> 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); + rb_check_arity(argc, 1, 2); + arg1 = argv[0]; + 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_PTR(ary)+pos); - RBASIC(arg2)->klass = rb_obj_class(ary); - rb_ary_splice(ary, pos, len, Qundef); - return arg2; + return ary_slice_bang_by_rb_ary_splice(ary, pos, len); } - if (argc != 1) { - /* error report */ - rb_scan_args(argc, argv, "11", NULL, NULL); - } - arg1 = argv[0]; - if (!FIXNUM_P(arg1)) { switch (rb_range_beg_len(arg1, &pos, &len, RARRAY_LEN(ary), 0)) { case Qtrue: /* valid range */ - goto delete_pos_len; + return ary_slice_bang_by_rb_ary_splice(ary, pos, len); case Qnil: /* invalid range */ return Qnil; @@ -2587,68 +4116,83 @@ ary_reject(VALUE orig, VALUE result) long i; for (i = 0; i < RARRAY_LEN(orig); i++) { - VALUE v = RARRAY_PTR(orig)[i]; - if (!RTEST(rb_yield(v))) { - rb_ary_push_1(result, v); + VALUE v = RARRAY_AREF(orig, i); + + if (!RTEST(rb_yield(v))) { + rb_ary_push(result, v); } } return result; } static VALUE -ary_reject_bang(VALUE ary) +reject_bang_i(VALUE a) { - long i; - VALUE result = Qnil; + volatile struct select_bang_arg *arg = (void *)a; + VALUE ary = arg->ary; + long i1, i2; - rb_ary_modify_check(ary); - for (i = 0; i < RARRAY_LEN(ary); ) { - VALUE v = RARRAY_PTR(ary)[i]; - if (RTEST(rb_yield(v))) { - rb_ary_delete_at(ary, i); - result = ary; - } - else { - i++; + 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; } - return result; + return (i1 == i2) ? Qnil : ary; +} + +static VALUE +ary_reject_bang(VALUE ary) +{ + struct select_bang_arg args; + rb_ary_modify_check(ary); + args.ary = ary; + args.len[0] = args.len[1] = 0; + return rb_ensure(reject_bang_i, (VALUE)&args, select_bang_ensure, (VALUE)&args); } /* * call-seq: - * ary.reject! {|item| block } -> ary or nil - * ary.reject! -> an_enumerator + * array.reject! {|element| ... } -> self or nil + * array.reject! -> new_enumerator * - * Equivalent to <code>Array#delete_if</code>, deleting elements from - * +self+ for which the block evaluates to true, but returns - * <code>nil</code> if no changes were made. - * The array is changed instantly every time the block is called and - * not after the iteration is over. - * See also <code>Enumerable#reject</code> and <code>Array#delete_if</code>. + * Removes each element for which the block returns a truthy value. * - * If no block is given, an enumerator is returned instead. + * Returns +self+ if any elements removed: + * a = [:foo, 'bar', 2, 'bat'] + * a.reject! {|element| element.to_s.start_with?('b') } # => [:foo, 2] * + * Returns +nil+ if no elements removed. + * + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2] + * a.reject! # => #<Enumerator: [:foo, "bar", 2]:reject!> */ static VALUE rb_ary_reject_bang(VALUE ary) { - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); + rb_ary_modify(ary); return ary_reject_bang(ary); } /* * call-seq: - * ary.reject {|item| block } -> new_ary - * ary.reject -> an_enumerator - * - * Returns a new array containing the items in +self+ - * for which the block is not true. - * See also <code>Array#delete_if</code> - * - * If no block is given, an enumerator is returned instead. - * + * array.reject {|element| ... } -> new_array + * array.reject -> new_enumerator + * + * Returns a new \Array whose elements are all those from +self+ + * for which the block returns +false+ or +nil+: + * 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> */ static VALUE @@ -2656,7 +4200,7 @@ rb_ary_reject(VALUE ary) { VALUE rejected_ary; - RETURN_ENUMERATOR(ary, 0, 0); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); rejected_ary = rb_ary_new(); ary_reject(ary, rejected_ary); return rejected_ary; @@ -2664,35 +4208,35 @@ rb_ary_reject(VALUE ary) /* * call-seq: - * ary.delete_if {|item| block } -> ary - * ary.delete_if -> an_enumerator + * array.delete_if {|element| ... } -> self + * array.delete_if -> Enumerator * - * Deletes every element of +self+ for which <i>block</i> evaluates - * to true. - * The array is changed instantly every time the block is called and - * not after the iteration is over. - * See also <code>Array#reject!</code> + * Removes each element in +self+ for which the block returns a truthy value; + * returns +self+: + * a = [:foo, 'bar', 2, 'bat'] + * a.delete_if {|element| element.to_s.start_with?('b') } # => [:foo, 2] * - * If no block is given, an enumerator is returned instead. - * - * a = [ "a", "b", "c" ] - * a.delete_if {|x| x >= "b" } #=> ["a"] + * Returns a new \Enumerator if no block given: + * a = [:foo, 'bar', 2] + * a.delete_if # => #<Enumerator: [:foo, "bar", 2]:delete_if> */ static VALUE rb_ary_delete_if(VALUE ary) { - RETURN_ENUMERATOR(ary, 0, 0); + ary_verify(ary); + RETURN_SIZED_ENUMERATOR(ary, 0, 0, ary_enum_length); ary_reject_bang(ary); return ary; } static VALUE -take_i(VALUE val, VALUE *args, int argc, VALUE *argv) +take_i(RB_BLOCK_CALL_FUNC_ARGLIST(val, cbarg)) { - if (args[1]-- == 0) rb_iter_break(); + VALUE *args = (VALUE *)cbarg; if (argc > 1) val = rb_ary_new4(argc, argv); rb_ary_push(args[0], val); + if (--args[1] == 0) rb_iter_break(); return Qnil; } @@ -2702,76 +4246,132 @@ 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; - rb_block_call(obj, rb_intern("each"), 0, 0, take_i, (VALUE)args); + 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)); return result; } /* * call-seq: - * ary.zip(arg, ...) -> new_ary - * ary.zip(arg, ...) {| arr | block } -> nil - * - * Converts any arguments to arrays, then merges elements of - * +self+ with corresponding elements from each argument. This - * generates a sequence of <code>self.size</code> <em>n</em>-element - * arrays, where <em>n</em> is one more that the count of arguments. If - * the size of any argument is less than <code>enumObj.size</code>, - * <code>nil</code> values are supplied. If a block is given, it is - * invoked for each output array, otherwise an array of arrays is - * returned. - * - * 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]] + * 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]] + * + * If any array in +other_arrays+ is larger than +self+, + * its trailing elements are ignored: + * 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]] + * + * When a block is given, calls the block with each of the sub-arrays (formed as above); returns nil + * 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] */ static VALUE rb_ary_zip(int argc, VALUE *argv, VALUE ary) { int i, j; - long len; + long len = RARRAY_LEN(ary); VALUE result = Qnil; - len = RARRAY_LEN(ary); for (i=0; i<argc; i++) { argv[i] = take_items(argv[i], len); } - if (!rb_block_given_p()) { - result = rb_ary_new2(len); - } - for (i=0; i<RARRAY_LEN(ary); i++) { - VALUE tmp = rb_ary_new2(argc+1); + if (rb_block_given_p()) { + int arity = rb_block_arity(); - rb_ary_push(tmp, rb_ary_elt(ary, i)); - for (j=0; j<argc; j++) { - rb_ary_push(tmp, rb_ary_elt(argv[j], i)); - } - if (NIL_P(result)) { - rb_yield(tmp); + 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); + + 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); } } + return result; } /* * call-seq: - * ary.transpose -> new_ary - * - * Assumes that +self+ is an array of arrays and transposes the - * rows and columns. + * array.transpose -> new_array * - * a = [[1,2], [3,4], [5,6]] - * a.transpose #=> [[1, 3, 5], [2, 4, 6]] + * Transposes the rows and columns in an \Array of Arrays; + * the nested Arrays must all be the same size: + * a = [[:a0, :a1], [:b0, :b1], [:c0, :c1]] + * a.transpose # => [[:a0, :b0, :c0], [:a1, :b1, :c1]] */ static VALUE @@ -2804,14 +4404,11 @@ rb_ary_transpose(VALUE ary) /* * call-seq: - * ary.replace(other_ary) -> ary + * array.replace(other_array) -> self * - * Replaces the contents of +self+ with the contents of - * <i>other_ary</i>, truncating or expanding if necessary. - * - * a = [ "a", "b", "c", "d", "e" ] - * a.replace([ "x", "y", "z" ]) #=> ["x", "y", "z"] - * a #=> ["x", "y", "z"] + * 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] */ VALUE @@ -2822,102 +4419,245 @@ rb_ary_replace(VALUE copy, VALUE orig) if (copy == orig) return copy; if (RARRAY_LEN(orig) <= RARRAY_EMBED_LEN_MAX) { - VALUE *ptr; - VALUE shared = 0; + VALUE shared_root = 0; if (ARY_OWNS_HEAP_P(copy)) { - xfree(RARRAY_PTR(copy)); - } + ary_heap_free(copy); + } else if (ARY_SHARED_P(copy)) { - shared = ARY_SHARED(copy); + shared_root = ARY_SHARED_ROOT(copy); FL_UNSET_SHARED(copy); } FL_SET_EMBED(copy); - ptr = RARRAY_PTR(orig); - MEMCPY(RARRAY_PTR(copy), ptr, VALUE, RARRAY_LEN(orig)); - if (shared) { - rb_ary_decrement_share(shared); + 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)); } else { - VALUE shared = ary_make_shared(orig); + VALUE shared_root = ary_make_shared(orig); if (ARY_OWNS_HEAP_P(copy)) { - xfree(RARRAY_PTR(copy)); + ary_heap_free(copy); } else { rb_ary_unshare_safe(copy); } FL_UNSET_EMBED(copy); - ARY_SET_PTR(copy, RARRAY_PTR(orig)); - ARY_SET_LEN(copy, RARRAY_LEN(orig)); - rb_ary_set_shared(copy, shared); + ARY_SET_PTR(copy, ARY_HEAP_PTR(orig)); + ARY_SET_LEN(copy, ARY_HEAP_LEN(orig)); + rb_ary_set_shared(copy, shared_root); } + ary_verify(copy); return copy; } /* * call-seq: - * ary.clear -> ary - * - * Removes all elements from +self+. + * array.clear -> self * - * a = [ "a", "b", "c", "d", "e" ] - * a.clear #=> [ ] + * Removes all elements from +self+: + * a = [:foo, 'bar', 2] + * a.clear # => [] */ VALUE rb_ary_clear(VALUE ary) { rb_ary_modify_check(ary); - ARY_SET_LEN(ary, 0); if (ARY_SHARED_P(ary)) { if (!ARY_EMBED_P(ary)) { rb_ary_unshare(ary); FL_SET_EMBED(ary); + ARY_SET_EMBED_LEN(ary, 0); } } - else if (ARY_DEFAULT_SIZE * 2 < ARY_CAPA(ary)) { - ary_resize_capa(ary, ARY_DEFAULT_SIZE * 2); + else { + ARY_SET_LEN(ary, 0); + if (ARY_DEFAULT_SIZE * 2 < ARY_CAPA(ary)) { + ary_resize_capa(ary, ARY_DEFAULT_SIZE * 2); + } } + ary_verify(ary); return 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 - * - * The first three forms set the selected elements of +self+ (which - * may be the entire array) to <i>obj</i>. A <i>start</i> of - * <code>nil</code> is equivalent to zero. A <i>length</i> of - * <code>nil</code> is equivalent to <i>self.length</i>. The last three - * forms fill the array with the value of the block. The block is - * passed the absolute index of each element to be filled. - * Negative values of <i>start</i> count from the end of the array. - * - * 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] + * 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"] + * + * If <tt>range.last</tt> is negative, counts from the end: + * 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"] + * + * If <tt>range.last</tt> and <tt>range.last</tt> are both negative, + * both count from the end of the array: + * 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"] + * + * 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"] */ static VALUE rb_ary_fill(int argc, VALUE *argv, VALUE ary) { - VALUE item, arg1, arg2; + VALUE item = Qundef, arg1, arg2; long beg = 0, end = 0, len = 0; - VALUE *p, *pend; - int block_p = FALSE; if (rb_block_given_p()) { - block_p = TRUE; rb_scan_args(argc, argv, "02", &arg1, &arg2); argc += 1; /* hackish */ } @@ -2955,96 +4695,123 @@ rb_ary_fill(int argc, VALUE *argv, VALUE ary) if (end >= ARY_CAPA(ary)) { ary_resize_capa(ary, end); } - rb_mem_clear(RARRAY_PTR(ary) + RARRAY_LEN(ary), end - RARRAY_LEN(ary)); + ary_mem_clear(ary, RARRAY_LEN(ary), end - RARRAY_LEN(ary)); ARY_SET_LEN(ary, end); } - if (block_p) { + if (item == Qundef) { VALUE v; long i; for (i=beg; i<end; i++) { v = rb_yield(LONG2NUM(i)); if (i>=RARRAY_LEN(ary)) break; - RARRAY_PTR(ary)[i] = v; + ARY_SET(ary, i, v); } } else { - p = RARRAY_PTR(ary) + beg; - pend = p + len; - while (p < pend) { - *p++ = item; - } + ary_memfill(ary, beg, len, item); } return ary; } /* * call-seq: - * ary + other_ary -> new_ary + * array + other_array -> new_array * - * Concatenation---Returns a new array built by concatenating the - * two arrays together to produce a third array. + * Returns a new \Array containing all elements of +array+ + * followed by all elements of +other_array+: + * a = [0, 1] + [2, 3] + * a # => [0, 1, 2, 3] * - * [ 1, 2, 3 ] + [ 4, 5 ] #=> [ 1, 2, 3, 4, 5 ] + * Related: #concat. */ VALUE rb_ary_plus(VALUE x, VALUE y) { VALUE z; - long len; + long len, xlen, ylen; y = to_ary(y); - len = RARRAY_LEN(x) + RARRAY_LEN(y); + xlen = RARRAY_LEN(x); + ylen = RARRAY_LEN(y); + len = xlen + ylen; z = rb_ary_new2(len); - MEMCPY(RARRAY_PTR(z), RARRAY_PTR(x), VALUE, RARRAY_LEN(x)); - MEMCPY(RARRAY_PTR(z) + RARRAY_LEN(x), RARRAY_PTR(y), VALUE, RARRAY_LEN(y)); + + ary_memcpy(z, 0, xlen, RARRAY_CONST_PTR_TRANSIENT(x)); + ary_memcpy(z, xlen, ylen, RARRAY_CONST_PTR_TRANSIENT(y)); ARY_SET_LEN(z, len); return z; } +static VALUE +ary_append(VALUE x, VALUE y) +{ + long n = RARRAY_LEN(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_ary) -> ary - * - * Appends the elements of <i>other_ary</i> to +self+. + * array.concat(*other_arrays) -> self * - * [ "a", "b" ].concat( ["c", "d"] ) #=> [ "a", "b", "c", "d" ] + * Adds to +array+ all elements from each \Array in +other_arrays+; returns +self+: + * a = [0, 1] + * a.concat([2, 3], [4, 5]) # => [0, 1, 2, 3, 4, 5] */ +static VALUE +rb_ary_concat_multi(int argc, VALUE *argv, VALUE ary) +{ + rb_ary_modify_check(ary); + + if (argc == 1) { + 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); + } + + ary_verify(ary); + return ary; +} VALUE rb_ary_concat(VALUE x, VALUE y) { - rb_ary_modify_check(x); - y = to_ary(y); - if (RARRAY_LEN(y) > 0) { - rb_ary_splice(x, RARRAY_LEN(x), 0, y); - } - return x; + return ary_append(x, to_ary(y)); } - /* * call-seq: - * ary * int -> new_ary - * ary * str -> new_string - * - * Repetition---With a String argument, equivalent to - * self.join(str). Otherwise, returns a new array - * built by concatenating the _int_ copies of +self+. - * + * array * n -> new_array + * array * string_separator -> new_string * - * [ 1, 2, 3 ] * 3 #=> [ 1, 2, 3, 1, 2, 3, 1, 2, 3 ] - * [ 1, 2, 3 ] * "," #=> "1,2,3" + * When non-negative argument \Integer +n+ is given, + * returns a new \Array built by concatenating the +n+ 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>: + * [0, [0, 1], {foo: 0}] * ', ' # => "0, 0, 1, {:foo=>0}" */ static VALUE rb_ary_times(VALUE ary, VALUE times) { - VALUE ary2, tmp, *ptr, *ptr2; + VALUE ary2, tmp; + const VALUE *ptr; long t, len; tmp = rb_check_string_type(times); @@ -3054,7 +4821,7 @@ rb_ary_times(VALUE ary, VALUE times) len = NUM2LONG(times); if (len == 0) { - ary2 = ary_new(rb_obj_class(ary), 0); + ary2 = ary_new(rb_cArray, 0); goto out; } if (len < 0) { @@ -3065,46 +4832,37 @@ rb_ary_times(VALUE ary, VALUE times) } 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_PTR(ary); - ptr2 = RARRAY_PTR(ary2); + ptr = RARRAY_CONST_PTR_TRANSIENT(ary); t = RARRAY_LEN(ary); if (0 < t) { - MEMCPY(ptr2, ptr, VALUE, t); - while (t <= len/2) { - MEMCPY(ptr2+t, ptr2, VALUE, t); + ary_memcpy(ary2, 0, t, ptr); + while (t <= len/2) { + ary_memcpy(ary2, t, t, RARRAY_CONST_PTR_TRANSIENT(ary2)); t *= 2; } if (t < len) { - MEMCPY(ptr2+t, ptr2, VALUE, len-t); + ary_memcpy(ary2, t, len-t, RARRAY_CONST_PTR_TRANSIENT(ary2)); } } out: - OBJ_INFECT(ary2, ary); - return ary2; } /* * call-seq: - * ary.assoc(obj) -> new_ary or nil - * - * Searches through an array whose elements are also arrays - * comparing _obj_ with the first element of each contained array - * using obj.==. - * Returns the first contained array that matches (that - * is, the first associated array), - * or +nil+ if no match is found. - * See also <code>Array#rassoc</code>. - * - * 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 + * array.assoc(obj) -> found_array or nil + * + * Returns the first element in +self+ that is an \Array + * whose first element <tt>==</tt> +obj+: + * a = [{foo: 0}, [2, 4], [4, 5, 6], [4, 5]] + * a.assoc(4) # => [4, 5, 6] + * + * Returns +nil+ if no such element is found. + * + * Related: #rassoc. */ VALUE @@ -3114,9 +4872,9 @@ rb_ary_assoc(VALUE ary, VALUE key) VALUE v; for (i = 0; i < RARRAY_LEN(ary); ++i) { - v = rb_check_array_type(RARRAY_PTR(ary)[i]); + v = rb_check_array_type(RARRAY_AREF(ary, i)); if (!NIL_P(v) && RARRAY_LEN(v) > 0 && - rb_equal(RARRAY_PTR(v)[0], key)) + rb_equal(RARRAY_AREF(v, 0), key)) return v; } return Qnil; @@ -3124,16 +4882,16 @@ rb_ary_assoc(VALUE ary, VALUE key) /* * call-seq: - * ary.rassoc(obj) -> new_ary or nil + * array.rassoc(obj) -> found_array or nil + * + * Returns the first element in +self+ that is an \Array + * whose second element <tt>==</tt> +obj+: + * a = [{foo: 0}, [2, 4], [4, 5, 6], [4, 5]] + * a.rassoc(4) # => [2, 4] * - * Searches through the array whose elements are also arrays. Compares - * _obj_ with the second element of each contained array using - * <code>==</code>. Returns the first contained array that matches. See - * also <code>Array#assoc</code>. + * 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 @@ -3143,10 +4901,10 @@ rb_ary_rassoc(VALUE ary, VALUE value) VALUE v; for (i = 0; i < RARRAY_LEN(ary); ++i) { - v = RARRAY_PTR(ary)[i]; - if (TYPE(v) == T_ARRAY && + v = RARRAY_AREF(ary, i); + if (RB_TYPE_P(v, T_ARRAY) && RARRAY_LEN(v) > 1 && - rb_equal(RARRAY_PTR(v)[1], value)) + rb_equal(RARRAY_AREF(v, 1), value)) return v; } return Qnil; @@ -3155,41 +4913,66 @@ rb_ary_rassoc(VALUE ary, VALUE value) static VALUE recursive_equal(VALUE ary1, VALUE ary2, int recur) { - long i; + long i, len1; + const VALUE *p1, *p2; if (recur) return Qtrue; /* Subtle! */ - for (i=0; i<RARRAY_LEN(ary1); i++) { - if (!rb_equal(rb_ary_elt(ary1, i), rb_ary_elt(ary2, i))) - return Qfalse; + + /* rb_equal() can evacuate ptrs */ + p1 = RARRAY_CONST_PTR(ary1); + p2 = RARRAY_CONST_PTR(ary2); + 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; + p1 = RARRAY_CONST_PTR(ary1) + i; + p2 = RARRAY_CONST_PTR(ary2) + i; + } + else { + return Qfalse; + } + } + p1++; + p2++; } return Qtrue; } /* * call-seq: - * ary == other_ary -> bool + * array == other_array -> true or false * - * 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 the other array. + * 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>: + * 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 rb_ary_equal(VALUE ary1, VALUE ary2) { if (ary1 == ary2) return Qtrue; - if (TYPE(ary2) != T_ARRAY) { - if (!rb_respond_to(ary2, rb_intern("to_ary"))) { + if (!RB_TYPE_P(ary2, T_ARRAY)) { + 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; return rb_exec_recursive_paired(recursive_equal, ary1, ary2, ary2); } @@ -3208,82 +4991,97 @@ recursive_eql(VALUE ary1, VALUE ary2, int recur) /* * call-seq: - * ary.eql?(other) -> true or false + * array.eql? other_array -> true or false * - * Returns <code>true</code> if +self+ and _other_ are the same object, - * or are both arrays with the same content. + * 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#==}[#method-i-3D-3D], + * which compares using method <tt>Object#==</tt>. */ static VALUE rb_ary_eql(VALUE ary1, VALUE ary2) { if (ary1 == ary2) return Qtrue; - if (TYPE(ary2) != T_ARRAY) return Qfalse; + if (!RB_TYPE_P(ary2, T_ARRAY)) return Qfalse; if (RARRAY_LEN(ary1) != RARRAY_LEN(ary2)) return Qfalse; + if (RARRAY_CONST_PTR_TRANSIENT(ary1) == RARRAY_CONST_PTR_TRANSIENT(ary2)) return Qtrue; return rb_exec_recursive_paired(recursive_eql, ary1, ary2, ary2); } +/* + * call-seq: + * array.hash -> integer + * + * Returns the integer hash value for +self+. + * + * 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 + */ + static VALUE -recursive_hash(VALUE ary, VALUE dummy, int recur) +rb_ary_hash(VALUE ary) { long i; st_index_t h; VALUE n; h = rb_hash_start(RARRAY_LEN(ary)); - if (recur) { - h = rb_hash_uint(h, NUM2LONG(rb_hash(rb_cArray))); - } - else { - for (i=0; i<RARRAY_LEN(ary); i++) { - n = rb_hash(RARRAY_PTR(ary)[i]); - h = rb_hash_uint(h, NUM2LONG(n)); - } + 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)); } h = rb_hash_end(h); - return LONG2FIX(h); -} - -/* - * call-seq: - * ary.hash -> fixnum - * - * Compute a hash-code for this array. Two arrays with the same content - * will have the same hash code (and will compare using <code>eql?</code>). - */ - -static VALUE -rb_ary_hash(VALUE ary) -{ - return rb_exec_recursive_outer(recursive_hash, ary, 0); + return ST2FIX(h); } /* * call-seq: - * ary.include?(obj) -> true or false + * array.include?(obj) -> true or false * - * Returns <code>true</code> if the given object is present in - * +self+ (that is, if any object <code>==</code> <i>anObject</i>), - * <code>false</code> otherwise. - * - * a = [ "a", "b", "c" ] - * a.include?("b") #=> true - * a.include?("z") #=> false + * Returns +true+ if for some index +i+ in +self+, <tt>obj == self[i]</tt>; + * otherwise +false+: + * [0, 1, 2].include?(2) # => true + * [0, 1, 2].include?(3) # => false */ VALUE rb_ary_includes(VALUE ary, VALUE item) { long i; + VALUE e; for (i=0; i<RARRAY_LEN(ary); i++) { - if (rb_equal(RARRAY_PTR(ary)[i], item)) { + e = RARRAY_AREF(ary, i); + if (rb_equal(e, item)) { return Qtrue; } } return Qfalse; } +static VALUE +rb_ary_includes_by_eql(VALUE ary, VALUE item) +{ + long i; + VALUE e; + + for (i=0; i<RARRAY_LEN(ary); i++) { + e = RARRAY_AREF(ary, i); + if (rb_eql(item, e)) { + return Qtrue; + } + } + return Qfalse; +} static VALUE recursive_cmp(VALUE ary1, VALUE ary2, int recur) @@ -3296,7 +5094,8 @@ recursive_cmp(VALUE ary1, VALUE ary2, int recur) len = RARRAY_LEN(ary2); } for (i=0; i<len; i++) { - VALUE v = rb_funcall(rb_ary_elt(ary1, i), id_cmp, 1, rb_ary_elt(ary2, 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; } @@ -3306,22 +5105,24 @@ recursive_cmp(VALUE ary1, VALUE ary2, int recur) /* * call-seq: - * ary <=> other_ary -> -1, 0, +1 or nil - * - * Comparison---Returns an integer (-1, 0, - * or +1) if this array is less than, equal to, or greater than - * <i>other_ary</i>. Each object in each array is compared - * (using <=>). If any value isn't - * equal, then that inequality is the return value. If all the - * values found are equal, then the return is based on a - * comparison of the array lengths. Thus, two arrays are - * ``equal'' according to <code>Array#<=></code> 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. - * - * [ "a", "a", "c" ] <=> [ "a", "b", "c" ] #=> -1 - * [ 1, 2, 3, 4, 5, 6 ] <=> [ 1, 2 ] #=> +1 + * 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: + * [0, 1, 2] <=> [0, 1, 3] # => -1 + * + * Returns 1 if any result is 1: + * [0, 1, 2] <=> [0, 1, 1] # => 1 + * + * When all results are zero: + * - Returns -1 if +array+ is smaller than +other_array+: + * [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 */ VALUE @@ -3347,24 +5148,26 @@ ary_add_hash(VALUE hash, VALUE ary) long i; for (i=0; i<RARRAY_LEN(ary); i++) { - rb_hash_aset(hash, RARRAY_PTR(ary)[i], Qtrue); + VALUE elt = RARRAY_AREF(ary, i); + rb_hash_add_new_element(hash, elt, elt); } return hash; } static inline VALUE -ary_tmp_hash_new(void) +ary_tmp_hash_new(VALUE ary) { - VALUE hash = rb_hash_new(); + long size = RARRAY_LEN(ary); + VALUE hash = rb_hash_new_with_size(size); - RBASIC(hash)->klass = 0; + RBASIC_CLEAR_CLASS(hash); return hash; } static VALUE ary_make_hash(VALUE ary) { - VALUE hash = ary_tmp_hash_new(); + VALUE hash = ary_tmp_hash_new(ary); return ary_add_hash(hash, ary); } @@ -3375,9 +5178,7 @@ ary_add_hash_by(VALUE hash, VALUE ary) for (i = 0; i < RARRAY_LEN(ary); ++i) { VALUE v = rb_ary_elt(ary, i), k = rb_yield(v); - if (rb_hash_lookup2(hash, k, Qundef) == Qundef) { - rb_hash_aset(hash, k, v); - } + rb_hash_add_new_element(hash, k, v); } return hash; } @@ -3385,44 +5186,59 @@ ary_add_hash_by(VALUE hash, VALUE ary) static VALUE ary_make_hash_by(VALUE ary) { - VALUE hash = ary_tmp_hash_new(); + VALUE hash = ary_tmp_hash_new(ary); return ary_add_hash_by(hash, ary); } static inline void ary_recycle_hash(VALUE hash) { - if (RHASH(hash)->ntbl) { - st_table *tbl = RHASH(hash)->ntbl; - RHASH(hash)->ntbl = 0; + assert(RBASIC_CLASS(hash) == 0); + if (RHASH_ST_TABLE_P(hash)) { + st_table *tbl = RHASH_ST_TABLE(hash); st_free_table(tbl); + RHASH_ST_CLEAR(hash); } } /* * call-seq: - * ary - other_ary -> new_ary + * array - other_array -> new_array * - * Array Difference---Returns a new array that is a copy of - * the original array, removing any items that also appear in - * <i>other_ary</i>. (If you need set-like behavior, see the - * library class Set.) + * 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: + * [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 ] + * Related: Array#difference. */ static VALUE rb_ary_diff(VALUE ary1, VALUE ary2) { VALUE ary3; - volatile VALUE hash; + VALUE hash; long i; - hash = ary_make_hash(to_ary(ary2)); + 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; + } + + hash = ary_make_hash(ary2); for (i=0; i<RARRAY_LEN(ary1); i++) { - if (st_lookup(RHASH_TBL(hash), RARRAY_PTR(ary1)[i], 0)) continue; + if (rb_hash_stlike_lookup(hash, RARRAY_AREF(ary1, i), NULL)) continue; rb_ary_push(ary3, rb_ary_elt(ary1, i)); } ary_recycle_hash(hash); @@ -3431,12 +5247,70 @@ rb_ary_diff(VALUE ary1, VALUE ary2) /* * call-seq: - * ary & other_ary -> new_ary + * array.difference(*other_arrays) -> new_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: + * [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] + * + * Returns a copy of +self+ if no arguments given. + * + * Related: Array#-. + */ + +static VALUE +rb_ary_difference_multi(int argc, VALUE *argv, VALUE ary) +{ + VALUE ary_diff; + long i, length; + volatile VALUE t0; + bool *is_hash = ALLOCV_N(bool, t0, argc); + ary_diff = rb_ary_new(); + length = RARRAY_LEN(ary); + + for (i = 0; i < argc; i++) { + argv[i] = to_ary(argv[i]); + is_hash[i] = (length > SMALL_ARRAY_LEN && RARRAY_LEN(argv[i]) > SMALL_ARRAY_LEN); + if (is_hash[i]) argv[i] = ary_make_hash(argv[i]); + } + + for (i = 0; i < RARRAY_LEN(ary); i++) { + int j; + VALUE elt = rb_ary_elt(ary, i); + for (j = 0; j < argc; j++) { + if (is_hash[j]) { + if (rb_hash_stlike_lookup(argv[j], RARRAY_AREF(ary, i), NULL)) + break; + } + else { + if (rb_ary_includes_by_eql(argv[j], elt)) break; + } + } + if (j == argc) rb_ary_push(ary_diff, elt); + } + + ALLOCV_END(t0); + + return ary_diff; +} + + +/* + * call-seq: + * array & other_array -> new_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>: + * [0, 1, 2, 3] & [1, 2] # => [1, 2] + * [0, 1, 0, 1] & [0, 1] # => [0, 1] * - * Set Intersection---Returns a new array - * containing elements common to the two arrays, with no duplicates. + * Preserves order from +array+: + * [0, 1, 2] & [3, 2, 1, 0] # => [0, 1, 2] * - * [ 1, 1, 3, 5 ] & [ 1, 2, 3 ] #=> [ 1, 3 ] + * Related: Array#intersection. */ @@ -3448,16 +5322,25 @@ rb_ary_and(VALUE ary1, VALUE ary2) long i; ary2 = to_ary(ary2); - ary3 = rb_ary_new2(RARRAY_LEN(ary1) < RARRAY_LEN(ary2) ? - RARRAY_LEN(ary1) : RARRAY_LEN(ary2)); - hash = ary_make_hash(ary2); + ary3 = rb_ary_new(); + 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; + } - if (RHASH_EMPTY_P(hash)) - return ary3; + hash = ary_make_hash(ary2); for (i=0; i<RARRAY_LEN(ary1); i++) { - vv = (st_data_t)(v = rb_ary_elt(ary1, i)); - if (st_delete(RHASH_TBL(hash), &vv, 0)) { + v = RARRAY_AREF(ary1, i); + vv = (st_data_t)v; + if (rb_hash_stlike_delete(hash, &vv, 0)) { rb_ary_push(ary3, v); } } @@ -3468,40 +5351,554 @@ rb_ary_and(VALUE ary1, VALUE ary2) /* * call-seq: - * ary | other_ary -> new_ary + * array.intersection(*other_arrays) -> new_array * - * Set Union---Returns a new array by joining this array with - * <i>other_ary</i>, removing duplicates. + * 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] * - * [ "a", "b", "c" ] | [ "c", "d", "a" ] - * #=> [ "a", "b", "c", "d" ] + * 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) +{ + if (existing) return ST_STOP; + *key = *value = (VALUE)arg; + return ST_CONTINUE; +} + +static void +rb_ary_union(VALUE ary_union, VALUE ary) +{ + long i; + for (i = 0; i < RARRAY_LEN(ary); i++) { + VALUE elt = rb_ary_elt(ary, i); + if (rb_ary_includes_by_eql(ary_union, elt)) continue; + rb_ary_push(ary_union, elt); + } +} + +static void +rb_ary_union_hash(VALUE hash, VALUE ary2) +{ + long i; + for (i = 0; i < RARRAY_LEN(ary2); i++) { + VALUE elt = RARRAY_AREF(ary2, i); + if (!rb_hash_stlike_update(hash, (st_data_t)elt, ary_hash_orset, (st_data_t)elt)) { + RB_OBJ_WRITTEN(hash, Qundef, elt); + } + } +} + +/* + * call-seq: + * array | other_array -> new_array + * + * Returns the union of +array+ and \Array +other_array+; + * duplicates are removed; order is preserved; + * items are compared using <tt>eql?</tt>: + * [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] + * + * Related: Array#union. */ static VALUE rb_ary_or(VALUE ary1, VALUE ary2) { - VALUE hash, ary3, v; + VALUE hash, ary3; + + ary2 = to_ary(ary2); + if (RARRAY_LEN(ary1) + RARRAY_LEN(ary2) <= SMALL_ARRAY_LEN) { + ary3 = rb_ary_new(); + rb_ary_union(ary3, ary1); + rb_ary_union(ary3, ary2); + return ary3; + } + + hash = ary_make_hash(ary1); + rb_ary_union_hash(hash, ary2); + + ary3 = rb_hash_values(hash); + ary_recycle_hash(hash); + return ary3; +} + +/* + * call-seq: + * array.union(*other_arrays) -> new_array + * + * 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>: + * [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] + * + * Returns a copy of +self+ if no arguments given. + * + * Related: Array#|. + */ + +static VALUE +rb_ary_union_multi(int argc, VALUE *argv, VALUE ary) +{ + int i; + long sum; + VALUE hash, ary_union; + + sum = RARRAY_LEN(ary); + for (i = 0; i < argc; i++) { + argv[i] = to_ary(argv[i]); + sum += RARRAY_LEN(argv[i]); + } + + if (sum <= SMALL_ARRAY_LEN) { + ary_union = rb_ary_new(); + + rb_ary_union(ary_union, ary); + for (i = 0; i < argc; i++) rb_ary_union(ary_union, argv[i]); + + return ary_union; + } + + hash = ary_make_hash(ary); + for (i = 0; i < argc; i++) rb_ary_union_hash(hash, argv[i]); + + ary_union = rb_hash_values(hash); + ary_recycle_hash(hash); + return ary_union; +} + +/* + * call-seq: + * ary.intersect?(other_ary) -> true or false + * + * Returns +true+ if the array and +other_ary+ have at least one element in + * common, otherwise returns +false+. + * + * a = [ 1, 2, 3 ] + * b = [ 3, 4, 5 ] + * c = [ 5, 6, 7 ] + * a.intersect?(b) #=> true + * a.intersect?(c) #=> false + */ + +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); - ary3 = rb_ary_new2(RARRAY_LEN(ary1)+RARRAY_LEN(ary2)); - hash = ary_add_hash(ary_make_hash(ary1), ary2); + if (RARRAY_LEN(ary1) == 0 || RARRAY_LEN(ary2) == 0) return Qfalse; - for (i=0; i<RARRAY_LEN(ary1); i++) { - vv = (st_data_t)(v = rb_ary_elt(ary1, i)); - if (st_delete(RHASH_TBL(hash), &vv, 0)) { - rb_ary_push(ary3, v); + 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"] + */ +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; + + 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(ary2); i++) { - vv = (st_data_t)(v = rb_ary_elt(ary2, i)); - if (st_delete(RHASH_TBL(hash), &vv, 0)) { - rb_ary_push(ary3, v); + else if (n > 0) { + result = RARRAY_AREF(ary, 0); + if (n > 1) { + if (FIXNUM_P(result) && CMP_OPTIMIZABLE(cmp_opt, Integer)) { + return ary_max_opt_fixnum(ary, 1, result); + } + else if (STRING_P(result) && CMP_OPTIMIZABLE(cmp_opt, String)) { + return ary_max_opt_string(ary, 1, result); + } + else if (RB_FLOAT_TYPE_P(result) && CMP_OPTIMIZABLE(cmp_opt, Float)) { + return ary_max_opt_float(ary, 1, result); + } + else { + return ary_max_generic(ary, 1, result); + } + } + } + if (result == Qundef) 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: + * 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+. + * + * 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 minimum value per method <tt><=></tt>: + * [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"] + */ +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; + + 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; + } } } - ary_recycle_hash(hash); - return ary3; + else if (n > 0) { + result = RARRAY_AREF(ary, 0); + if (n > 1) { + if (FIXNUM_P(result) && CMP_OPTIMIZABLE(cmp_opt, Integer)) { + return ary_min_opt_fixnum(ary, 1, result); + } + else if (STRING_P(result) && CMP_OPTIMIZABLE(cmp_opt, String)) { + return ary_min_opt_string(ary, 1, result); + } + else if (RB_FLOAT_TYPE_P(result) && CMP_OPTIMIZABLE(cmp_opt, Float)) { + return ary_min_opt_float(ary, 1, result); + } + else { + return ary_min_generic(ary, 1, result); + } + } + } + if (result == Qundef) return Qnil; + return result; +} + +/* + * call-seq: + * array.minmax -> [min_val, max_val] + * array.minmax {|a, b| ... } -> [min_val, max_val] + * + * 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"] + */ +static VALUE +rb_ary_minmax(VALUE ary) +{ + if (rb_block_given_p()) { + return rb_call_super(0, NULL); + } + return rb_assoc_new(rb_ary_min(0, 0, ary), rb_ary_max(0, 0, ary)); } static int @@ -3513,60 +5910,57 @@ 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+, the first occurrence always being retained; + * returns +self+ if any elements removed, +nil+ otherwise. * - * Removes duplicate elements from +self+. If a block is given, - * it will use the return value of the block for comparison. - * Returns <code>nil</code> if no changes are made (that is, no - * duplicates are found). + * With no block given, identifies and removes elements using method <tt>eql?</tt> + * to compare. * - * a = [ "a", "a", "b", "b", "c" ] - * a.uniq! # => ["a", "b", "c"] + * Returns +self+ if any elements removed: + * a = [0, 0, 1, 1, 2, 2] + * a.uniq! # => [0, 1, 2] * - * b = [ "a", "b", "c" ] - * b.uniq! # => nil + * Returns +nil+ if no elements removed. * - * c = [["student","sam"], ["student","george"], ["teacher","matz"]] - * c.uniq! { |s| s.first } # => [["student", "sam"], ["teacher", "matz"]] + * 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. * + * Returns +self+ if any elements removed: + * 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) { - VALUE hash, v; - long i, j; + VALUE hash; + long hash_size; rb_ary_modify_check(ary); if (RARRAY_LEN(ary) <= 1) return Qnil; - if (rb_block_given_p()) { + if (rb_block_given_p()) hash = ary_make_hash_by(ary); - if (RARRAY_LEN(ary) == (i = RHASH_SIZE(hash))) { - return Qnil; - } - ARY_SET_LEN(ary, 0); - if (ARY_SHARED_P(ary) && !ARY_EMBED_P(ary)) { - rb_ary_unshare(ary); - FL_SET_EMBED(ary); - } - ary_resize_capa(ary, i); - st_foreach(RHASH_TBL(hash), push_value, ary); - } - else { + else hash = ary_make_hash(ary); - if (RARRAY_LEN(ary) == (long)RHASH_SIZE(hash)) { - return Qnil; - } - for (i=j=0; i<RARRAY_LEN(ary); i++) { - st_data_t vv = (st_data_t)(v = rb_ary_elt(ary, i)); - if (st_delete(RHASH_TBL(hash), &vv, 0)) { - rb_ary_store(ary, j++, v); - } - } - ARY_SET_LEN(ary, j); + + hash_size = RHASH_SIZE(hash); + if (RARRAY_LEN(ary) == hash_size) { + 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); } + ary_resize_capa(ary, hash_size); + rb_hash_foreach(hash, push_value, ary); ary_recycle_hash(hash); return ary; @@ -3574,58 +5968,55 @@ 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+. If a block - * is given, it will use the return value of the block for comparison. - * - * a = [ "a", "a", "b", "b", "c" ] - * a.uniq # => ["a", "b", "c"] - * - * b = [["student","sam"], ["student","george"], ["teacher","matz"]] - * b.uniq { |s| s.first } # => [["student", "sam"], ["teacher", "matz"]] - * + * array.uniq -> new_array + * array.uniq {|element| ... } -> new_array + * + * Returns a new \Array containing those elements from +self+ that are not duplicates, + * the first occurrence always being retained. + * + * With no block given, identifies and omits duplicates using method <tt>eql?</tt> + * to compare. + * a = [0, 0, 1, 1, 2, 2] + * a.uniq # => [0, 1, 2] + * + * 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: + * a = ['a', 'aa', 'aaa', 'b', 'bb', 'bbb'] + * a.uniq {|element| element.size } # => ["a", "aa", "aaa"] */ static VALUE rb_ary_uniq(VALUE ary) { - VALUE hash, uniq, v; - long i; + VALUE hash, uniq; - if (RARRAY_LEN(ary) <= 1) - return rb_ary_dup(ary); - if (rb_block_given_p()) { + 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 = ary_new(rb_obj_class(ary), RHASH_SIZE(hash)); - st_foreach(RHASH_TBL(hash), push_value, uniq); + uniq = rb_hash_values(hash); } else { hash = ary_make_hash(ary); - uniq = ary_new(rb_obj_class(ary), RHASH_SIZE(hash)); - for (i=0; i<RARRAY_LEN(ary); i++) { - st_data_t vv = (st_data_t)(v = rb_ary_elt(ary, i)); - if (st_delete(RHASH_TBL(hash), &vv, 0)) { - rb_ary_push(uniq, v); - } - } + uniq = rb_hash_values(hash); + } + if (hash) { + ary_recycle_hash(hash); } - ary_recycle_hash(hash); return uniq; } /* * call-seq: - * ary.compact! -> ary or nil + * array.compact! -> self or nil * - * Removes +nil+ elements from the array. - * Returns +nil+ if no changes were made, otherwise returns - * <i>ary</i>. + * Removes all +nil+ elements from +self+. * - * [ "a", nil, "b", nil, "c" ].compact! #=> [ "a", "b", "c" ] - * [ "a", "b", "c" ].compact! #=> nil + * Returns +self+ if any elements removed, otherwise +nil+. */ static VALUE @@ -3635,33 +6026,29 @@ rb_ary_compact_bang(VALUE ary) long n; rb_ary_modify(ary); - p = t = RARRAY_PTR(ary); + p = t = (VALUE *)RARRAY_CONST_PTR_TRANSIENT(ary); /* WB: no new reference */ end = p + RARRAY_LEN(ary); while (t < end) { if (NIL_P(*t)) t++; else *p++ = *t++; } - n = p - RARRAY_PTR(ary); + n = p - RARRAY_CONST_PTR_TRANSIENT(ary); if (RARRAY_LEN(ary) == n) { return Qnil; } - ARY_SET_LEN(ary, n); - if (n * 2 < ARY_CAPA(ary) && ARY_DEFAULT_SIZE * 2 < ARY_CAPA(ary)) { - ary_resize_capa(ary, n * 2); - } + ary_resize_smaller(ary, n); return ary; } /* * call-seq: - * ary.compact -> new_ary + * array.compact -> new_array * - * Returns a copy of +self+ with all +nil+ elements removed. - * - * [ "a", nil, "b", nil, "c", nil ].compact - * #=> [ "a", "b", "c" ] + * Returns a new \Array containing all non-+nil+ elements from +self+: + * a = [nil, 0, nil, 1, nil, 2, nil] + * a.compact # => [0, 1, 2] */ static VALUE @@ -3674,45 +6061,52 @@ 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 a count of specified elements. + * + * With no argument and no block, returns the count of all elements: + * [0, 1, 2].count # => 3 + * [].count # => 0 * - * Returns the number of elements. If an argument is given, counts - * the number of elements which equals to <i>obj</i>. If a block is - * given, counts the number of elements yielding a true value. + * 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 * - * ary = [1, 2, 4, 2] - * ary.count #=> 4 - * ary.count(2) #=> 2 - * ary.count{|x|x%2==0} #=> 3 + * 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 rb_ary_count(int argc, VALUE *argv, VALUE ary) { - long n = 0; + long i, n = 0; - if (argc == 0) { - VALUE *p, *pend; + if (rb_check_arity(argc, 0, 1) == 0) { + VALUE v; if (!rb_block_given_p()) return LONG2NUM(RARRAY_LEN(ary)); - for (p = RARRAY_PTR(ary), pend = p + RARRAY_LEN(ary); p < pend; p++) { - if (RTEST(rb_yield(*p))) n++; + for (i = 0; i < RARRAY_LEN(ary); i++) { + v = RARRAY_AREF(ary, i); + if (RTEST(rb_yield(v))) n++; } } else { - VALUE obj, *p, *pend; + VALUE obj = argv[0]; - rb_scan_args(argc, argv, "1", &obj); if (rb_block_given_p()) { rb_warn("given block not used"); } - for (p = RARRAY_PTR(ary), pend = p + RARRAY_LEN(ary); p < pend; p++) { - if (rb_equal(*p, obj)) n++; + for (i = 0; i < RARRAY_LEN(ary); i++) { + if (rb_equal(RARRAY_AREF(ary, i), obj)) n++; } } @@ -3720,37 +6114,71 @@ rb_ary_count(int argc, VALUE *argv, VALUE ary) } 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_PTR(ary)[i++]; + 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) || (level >= 0 && RARRAY_LEN(stack) / 2 >= level)) { + 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"); + 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); } - st_insert(memo, id, (st_data_t)Qtrue); rb_ary_push(stack, ary); rb_ary_push(stack, LONG2NUM(i)); ary = tmp; @@ -3760,35 +6188,49 @@ flatten(VALUE ary, int level, int *modified) if (RARRAY_LEN(stack) == 0) { break; } - id = (st_data_t)ary; - st_delete(memo, &id, 0); + 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); } - st_free_table(memo); + if (memo) { + st_clear(memo); + } - RBASIC(result)->klass = rb_class_of(ary); + RBASIC_SET_CLASS(result, rb_cArray); return result; } /* * call-seq: - * ary.flatten! -> ary or nil - * ary.flatten!(level) -> array or nil - * - * Flattens +self+ in place. - * Returns <code>nil</code> if no modifications were made (i.e., - * <i>ary</i> contains no subarrays.) If the optional <i>level</i> - * argument determines the level of recursion to flatten. - * - * 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]] + * 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: + * 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 + * + * 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! # => 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 */ static VALUE @@ -3797,14 +6239,13 @@ rb_ary_flatten_bang(int argc, VALUE *argv, VALUE ary) int mod = 0, level = -1; VALUE result, lv; - rb_scan_args(argc, argv, "01", &lv); + lv = (rb_check_arity(argc, 0, 1) ? argv[0] : Qnil); rb_ary_modify_check(ary); if (!NIL_P(lv)) level = NUM2INT(lv); if (level == 0) return Qnil; - result = flatten(ary, level, &mod); - if (mod == 0) { - ary_discard(result); + result = flatten(ary, level); + if (result == ary) { return Qnil; } if (!(mod = ARY_EMBED_P(result))) rb_obj_freeze(result); @@ -3816,190 +6257,150 @@ 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 this - * array (recursively). That is, for every element that is an array, - * extract its elements into the new array. If the optional - * <i>level</i> 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]] + * 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] */ static VALUE rb_ary_flatten(int argc, VALUE *argv, VALUE ary) { - int mod = 0, level = -1; - VALUE result, lv; + int level = -1; + VALUE result; - rb_scan_args(argc, argv, "01", &lv); - if (!NIL_P(lv)) level = NUM2INT(lv); - if (level == 0) return ary_make_shared_copy(ary); + if (rb_check_arity(argc, 0, 1) && !NIL_P(argv[0])) { + level = NUM2INT(argv[0]); + 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 VALUE sym_random; - -#define RAND_UPTO(max) (long)(rb_random_real(randgen)*(max)) - -/* - * call-seq: - * ary.shuffle! -> ary - * ary.shuffle!(random: rng) -> ary - * - * Shuffles elements in +self+ in place. - * If +rng+ is given, it will be used as the random number generator. - */ +#define RAND_UPTO(max) (long)rb_random_ulong_limited((randgen), (max)-1) 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 *ptr, opts, *snap_ptr, randgen = rb_cRandom; - long i, snap_len; + long i, len; - if (OPTHASH_GIVEN_P(opts)) { - randgen = rb_hash_lookup2(opts, sym_random, randgen); - } - if (argc > 0) { - rb_raise(rb_eArgError, "wrong number of arguments (%d for 0)", argc); - } rb_ary_modify(ary); - i = RARRAY_LEN(ary); - ptr = RARRAY_PTR(ary); - snap_len = i; - snap_ptr = ptr; - while (i) { - long j = RAND_UPTO(i); - VALUE tmp; - if (snap_len != RARRAY_LEN(ary) || snap_ptr != RARRAY_PTR(ary)) { - rb_raise(rb_eRuntimeError, "modified during shuffle"); + i = len = RARRAY_LEN(ary); + RARRAY_PTR_USE(ary, ptr, { + 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 this array shuffled. - * - * a = [ 1, 2, 3 ] #=> [1, 2, 3] - * a.shuffle #=> [2, 3, 1] - * - * If +rng+ is given, it 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 - * <code>nil</code> and the second form returns an empty array. - * - * If +rng+ is given, it will be used as the random number generator. - */ - - 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, *ptr; - VALUE opts, randgen = rb_cRandom; + VALUE result; long n, len, i, j, k, idx[10]; - double rnds[numberof(idx)]; + long rnds[numberof(idx)]; + long memo_threshold; - if (OPTHASH_GIVEN_P(opts)) { - randgen = rb_hash_lookup2(opts, sym_random, randgen); - } - ptr = RARRAY_PTR(ary); len = RARRAY_LEN(ary); - if (argc == 0) { - if (len == 0) return Qnil; - if (len == 1) { + if (!to_array) { + if (len < 2) i = 0; - } - else { - double x = rb_random_real(randgen); - if ((len = RARRAY_LEN(ary)) == 0) return Qnil; - i = (long)(x * len); - } - return RARRAY_PTR(ary)[i]; + else + i = RAND_UPTO(len); + + return rb_ary_elt(ary, i); } - rb_scan_args(argc, argv, "1", &nv); 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] = rb_random_real(randgen); + rnds[i] = RAND_UPTO(len - i); } } + k = len; len = RARRAY_LEN(ary); - ptr = RARRAY_PTR(ary); + if (len < k && n <= numberof(idx)) { + 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_new2(0); + return rb_ary_new_capa(0); case 1: - i = (long)(rnds[0] * len); - return rb_ary_new4(1, &ptr[i]); + i = rnds[0]; + return rb_ary_new_from_args(1, RARRAY_AREF(ary, i)); case 2: - i = (long)(rnds[0] * len); - j = (long)(rnds[1] * (len-1)); + i = rnds[0]; + j = rnds[1]; if (j >= i) j++; - return rb_ary_new3(2, ptr[i], ptr[j]); + return rb_ary_new_from_args(2, RARRAY_AREF(ary, i), RARRAY_AREF(ary, j)); case 3: - i = (long)(rnds[0] * len); - j = (long)(rnds[1] * (len-1)); - k = (long)(rnds[2] * (len-2)); + 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_new3(3, ptr[i], ptr[j], ptr[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; if (n <= numberof(idx)) { - VALUE *ptr_result; long sorted[numberof(idx)]; - sorted[0] = idx[0] = (long)(rnds[0] * len); + sorted[0] = idx[0] = rnds[0]; for (i=1; i<n; i++) { - k = (long)(rnds[i] * --len); + k = rnds[i]; for (j = 0; j < i; ++j) { if (k < sorted[j]) break; ++k; @@ -4007,161 +6408,332 @@ rb_ary_sample(int argc, VALUE *argv, VALUE ary) memmove(&sorted[j+1], &sorted[j], sizeof(sorted[0])*(i-j)); sorted[j] = idx[i] = k; } - result = rb_ary_new2(n); - ptr_result = RARRAY_PTR(result); - for (i=0; i<n; i++) { - ptr_result[i] = ptr[idx[i]]; - } + 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]); + } + }); + } + else if (n <= memo_threshold / 2) { + 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; + 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); } else { - VALUE *ptr_result; - result = rb_ary_new4(len, ptr); - RBASIC(result)->klass = 0; - ptr_result = RARRAY_PTR(result); + result = rb_ary_dup(ary); + RBASIC_CLEAR_CLASS(result); RB_GC_GUARD(ary); - 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(result)->klass = rb_cArray; + 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); return result; } +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); + } + if (RARRAY_LEN(self) == 0) return INT2FIX(0); + if (NIL_P(n)) return DBL2NUM(HUGE_VAL); + mul = NUM2LONG(n); + if (mul <= 0) return INT2FIX(0); + n = LONG2FIX(mul); + return rb_fix_mul_fix(rb_ary_length(self), n); +} /* * call-seq: - * ary.cycle(n=nil) {|obj| block } -> nil - * ary.cycle(n=nil) -> an_enumerator - * - * Calls <i>block</i> for each element repeatedly _n_ times or - * forever if none or +nil+ is given. If a non-positive number is - * given or the array is empty, does nothing. Returns +nil+ if the - * loop has finished without getting interrupted. - * - * If no block is given, an enumerator is returned instead. - * - * - * 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. - * + * 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] + * + * If +count+ is zero or negative, does not call the block: + * [0, 1].cycle(0) {|element| fail 'Cannot happen' } # => nil + * [0, 1].cycle(-1) {|element| fail 'Cannot happen' } # => nil + * + * When a block is given, and argument is omitted or +nil+, cycles forever: + * # Prints 0 and 1 forever. + * [0, 1].cycle {|element| puts element } + * [0, 1].cycle(nil) {|element| puts element } + * + * 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) { long n, i; - VALUE nv = Qnil; - rb_scan_args(argc, argv, "01", &nv); + rb_check_arity(argc, 0, 1); - RETURN_ENUMERATOR(ary, argc, argv); - if (NIL_P(nv)) { + RETURN_SIZED_ENUMERATOR(ary, argc, argv, rb_ary_cycle_size); + if (argc == 0 || NIL_P(argv[0])) { n = -1; } else { - n = NUM2LONG(nv); + n = NUM2LONG(argv[0]); if (n <= 0) return Qnil; } while (RARRAY_LEN(ary) > 0 && (n < 0 || 0 < n--)) { for (i=0; i<RARRAY_LEN(ary); i++) { - rb_yield(RARRAY_PTR(ary)[i]); + rb_yield(RARRAY_AREF(ary, i)); } } return Qnil; } -#define tmpbuf(n, size) rb_str_tmp_new((n)*(size)) -#define tmpbuf_discard(s) (rb_str_resize((s), 0L), RBASIC(s)->klass = rb_cString) #define tmpary(n) rb_ary_tmp_new(n) -#define tmpary_discard(a) (ary_discard(a), RBASIC(a)->klass = rb_cArray) +#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. + * Return the class of +values+ for reentry check. + */ +static int +yield_indexed_values(const VALUE values, const long r, const long *const p) +{ + const VALUE result = rb_ary_new2(r); + long i; + + for (i = 0; i < r; i++) ARY_SET(result, i, RARRAY_AREF(values, p[i])); + ARY_SET_LEN(result, r); + rb_yield(result); + return !RBASIC(values)->klass; +} /* - * Recursively compute permutations of r elements of the set [0..n-1]. - * When we have a complete permutation of array indexes, copy the values - * at those indexes into a new array and yield that array. + * Compute permutations of +r+ elements of the set <code>[0..n-1]</code>. + * + * When we have a complete permutation of array indices, copy the values + * at those indices into a new array and yield that array. * * n: the size of the set * r: the number of elements in each permutation * p: the array (of size r) that we're filling in - * index: what index we're filling in now * used: an array of booleans: whether a given index is already used * values: the Ruby array that holds the actual values to permute */ static void -permute0(long n, long r, long *p, long index, char *used, VALUE values) +permute0(const long n, const long r, long *const p, char *const used, const VALUE values) { - long i,j; - for (i = 0; i < n; i++) { - if (used[i] == 0) { + 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 */ - used[i] = 1; /* mark index used */ - permute0(n, r, p, index+1, /* recurse */ - used, values); - used[i] = 0; /* index unused */ + p[index] = i = 0; + continue; } - else { - /* We have a complete permutation of array indexes */ - /* Build a ruby array of the corresponding values */ - /* And yield it to the associated block */ - VALUE result = rb_ary_new2(r); - VALUE *result_array = RARRAY_PTR(result); - const VALUE *values_array = RARRAY_PTR(values); - - for (j = 0; j < r; j++) result_array[j] = values_array[p[j]]; - ARY_SET_LEN(result, r); - rb_yield(result); - if (RBASIC(values)->klass) { + 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. + * 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)); + } + } + else { + cnt = LONG2FIX(how_many == 0); + } + return cnt; +} + +static VALUE +binomial_coefficient(long comb, long size) +{ + VALUE r; + long i; + if (comb > size-comb) { + comb = size-comb; + } + if (comb < 0) { + return LONG2FIX(0); + } + else if (comb == 0) { + 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)); + } + return r; +} + +static VALUE +rb_ary_permutation_size(VALUE ary, VALUE args, VALUE eobj) +{ + long n = RARRAY_LEN(ary); + long k = (args && (RARRAY_LEN(args) > 0)) ? NUM2LONG(RARRAY_AREF(args, 0)) : n; + + return descending_factorial(n, k); +} + +/* * call-seq: - * ary.permutation { |p| block } -> ary - * ary.permutation -> an_enumerator - * ary.permutation(n) { |p| block } -> ary - * ary.permutation(n) -> an_enumerator - * - * When invoked with a block, yield all permutations of length <i>n</i> - * of the elements of <i>ary</i>, then return the array itself. - * If <i>n</i> is not specified, yield all permutations of all elements. - * The implementation makes no guarantees about the order in which - * the permutations are yielded. - * - * If no block is given, an enumerator is returned instead. - * - * Examples: - * - * 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 + * 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 +n+ is out of range (negative or larger than <tt>self.size</tt>), + * does not call the block: + * a = [0, 1, 2] + * a.permutation(-1) {|permutation| fail 'Cannot happen' } + * a.permutation(4) {|permutation| fail 'Cannot happen' } + * + * When a block given but no argument, + * behaves the same as <tt>a.permutation(a.size)</tt>: + * 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)> */ static VALUE rb_ary_permutation(int argc, VALUE *argv, VALUE ary) { - VALUE num; long r, n, i; n = RARRAY_LEN(ary); /* Array length */ - RETURN_ENUMERATOR(ary, argc, argv); /* Return enumerator if no block */ - rb_scan_args(argc, argv, "01", &num); - r = NIL_P(num) ? n : NUM2LONG(num); /* Permutation size from argument */ + RETURN_SIZED_ENUMERATOR(ary, argc, argv, rb_ary_permutation_size); /* Return enumerator if no block */ + r = n; + if (rb_check_arity(argc, 0, 1) && !NIL_P(argv[0])) + r = NUM2LONG(argv[0]); /* Permutation size from argument */ if (r < 0 || n < r) { /* no permutations: yield nothing */ @@ -4171,58 +6743,104 @@ rb_ary_permutation(int argc, VALUE *argv, VALUE ary) } 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_PTR(ary)[i])); + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); } } else { /* this is the general case */ - volatile VALUE t0 = tmpbuf(n,sizeof(long)); - long *p = (long*)RSTRING_PTR(t0); - volatile VALUE t1 = tmpbuf(n,sizeof(char)); - char *used = (char*)RSTRING_PTR(t1); + 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(ary0)->klass = 0; + RBASIC_CLEAR_CLASS(ary0); MEMZERO(used, char, n); /* initialize array */ - permute0(n, r, p, 0, used, ary0); /* compute and yield permutations */ - tmpbuf_discard(t0); - tmpbuf_discard(t1); - RBASIC(ary0)->klass = rb_cArray; + permute0(n, r, p, used, ary0); /* compute and yield permutations */ + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } +static void +combinate0(const long len, const long n, long *const stack, const VALUE values) +{ + long lev = 0; + + 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); + } +} + +static VALUE +rb_ary_combination_size(VALUE ary, VALUE args, VALUE eobj) +{ + long n = RARRAY_LEN(ary); + long k = NUM2LONG(RARRAY_AREF(args, 0)); + + return binomial_coefficient(k, n); +} + /* * call-seq: - * ary.combination(n) { |c| block } -> ary - * ary.combination(n) -> an_enumerator - * - * When invoked with a block, yields all combinations of length <i>n</i> - * of elements from <i>ary</i> and then returns <i>ary</i> itself. - * The implementation makes no guarantees about the order in which - * the combinations are yielded. - * - * If no block is given, an enumerator is returned instead. - * - * Examples: - * - * 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 - * + * 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 } + * Output: + * [0, 1] + * [0, 2] + * [1, 2] + * + * Another example: + * a = [0, 1, 2] + * a.combination(3) {|combination| p combination } + * 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)> */ static VALUE rb_ary_combination(VALUE ary, VALUE num) { - long n, i, len; + long i, n, len; n = NUM2LONG(num); - RETURN_ENUMERATOR(ary, 1, &num); + RETURN_SIZED_ENUMERATOR(ary, 1, &num, rb_ary_combination_size); len = RARRAY_LEN(ary); if (n < 0 || len < n) { /* yield nothing */ @@ -4231,108 +6849,135 @@ rb_ary_combination(VALUE ary, VALUE num) rb_yield(rb_ary_new2(0)); } else if (n == 1) { - for (i = 0; i < len; i++) { - rb_yield(rb_ary_new3(1, RARRAY_PTR(ary)[i])); + for (i = 0; i < RARRAY_LEN(ary); i++) { + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); } } else { - volatile VALUE t0 = tmpbuf(n+1, sizeof(long)); - long *stack = (long*)RSTRING_PTR(t0); - volatile VALUE cc = tmpary(n); - VALUE *chosen = RARRAY_PTR(cc); - long lev = 0; - - MEMZERO(stack, long, n); - stack[0] = -1; - for (;;) { - chosen[lev] = RARRAY_PTR(ary)[stack[lev+1]]; - for (lev++; lev < n; lev++) { - chosen[lev] = RARRAY_PTR(ary)[stack[lev+1] = stack[lev]+1]; - } - rb_yield(rb_ary_new4(n, chosen)); - if (RBASIC(t0)->klass) { - rb_raise(rb_eRuntimeError, "combination reentered"); - } - do { - if (lev == 0) goto done; - stack[lev--]++; - } while (stack[lev+1]+n == len+lev+1); - } - done: - tmpbuf_discard(t0); - tmpary_discard(cc); + 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); } return ary; } /* - * Recursively compute repeated permutations of r elements of the set - * [0..n-1]. - * When we have a complete repeated permutation of array indexes, copy the - * values at those indexes into a new array and yield that array. + * Compute repeated permutations of +r+ elements of the set + * <code>[0..n-1]</code>. + * + * When we have a complete repeated permutation of array indices, copy the + * values at those indices into a new array and yield that array. * * n: the size of the set * r: the number of elements in each permutation * p: the array (of size r) that we're filling in - * index: what index we're filling in now * values: the Ruby array that holds the actual values to permute */ static void -rpermute0(long n, long r, long *p, long index, VALUE values) +rpermute0(const long n, const long r, long *const p, const VALUE values) { - long i, j; - for (i = 0; i < n; i++) { - p[index] = i; - if (index < r-1) { /* if not done yet */ - rpermute0(n, r, p, index+1, values); /* recurse */ + long i = 0, index = 0; + + p[index] = i; + for (;;) { + if (++index < r-1) { + p[index] = i = 0; + continue; } - else { - /* We have a complete permutation of array indexes */ - /* Build a ruby array of the corresponding values */ - /* And yield it to the associated block */ - VALUE result = rb_ary_new2(r); - VALUE *result_array = RARRAY_PTR(result); - const VALUE *values_array = RARRAY_PTR(values); - - for (j = 0; j < r; j++) result_array[j] = values_array[p[j]]; - ARY_SET_LEN(result, r); - rb_yield(result); - if (RBASIC(values)->klass) { + 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); } } +static VALUE +rb_ary_repeated_permutation_size(VALUE ary, VALUE args, VALUE eobj) +{ + long n = RARRAY_LEN(ary); + long k = NUM2LONG(RARRAY_AREF(args, 0)); + + if (k < 0) { + return LONG2FIX(0); + } + if (n <= 0) { + 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) -> an_enumerator - * - * When invoked with a block, yield all repeated permutations of length - * <i>n</i> of the elements of <i>ary</i>, then return the array itself. - * The implementation makes no guarantees about the order in which - * the repeated permutations are yielded. - * - * If no block is given, an enumerator is returned instead. - * - * Examples: - * - * 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 + * 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 } + * Output: + * [0, 0] + * [0, 1] + * [0, 2] + * [1, 0] + * [1, 1] + * [1, 2] + * [2, 0] + * [2, 1] + * [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_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]] */ - static VALUE rb_ary_repeated_permutation(VALUE ary, VALUE num) { long r, n, i; n = RARRAY_LEN(ary); /* Array length */ - RETURN_ENUMERATOR(ary, 1, &num); /* Return enumerator if no block */ + RETURN_SIZED_ENUMERATOR(ary, 1, &num, rb_ary_repeated_permutation_size); /* Return Enumerator if no block */ r = NUM2LONG(num); /* Permutation size from argument */ if (r < 0) { @@ -4343,71 +6988,107 @@ rb_ary_repeated_permutation(VALUE ary, VALUE num) } 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_PTR(ary)[i])); + rb_yield(rb_ary_new3(1, RARRAY_AREF(ary, i))); } } else { /* this is the general case */ - volatile VALUE t0 = tmpbuf(r, sizeof(long)); - long *p = (long*)RSTRING_PTR(t0); + volatile VALUE t0; + long *p = ALLOCV_N(long, t0, r); VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - RBASIC(ary0)->klass = 0; + RBASIC_CLEAR_CLASS(ary0); - rpermute0(n, r, p, 0, ary0); /* compute and yield repeated permutations */ - tmpbuf_discard(t0); - RBASIC(ary0)->klass = rb_cArray; + rpermute0(n, r, p, ary0); /* compute and yield repeated permutations */ + ALLOCV_END(t0); + RBASIC_SET_CLASS_RAW(ary0, rb_cArray); } return ary; } static void -rcombinate0(long n, long r, long *p, long index, long rest, VALUE values) +rcombinate0(const long n, const long r, long *const p, const long rest, const VALUE values) { - long j; - if (rest > 0) { - for (; index < n; ++index) { - p[r-rest] = index; - rcombinate0(n, r, p, index, rest-1, values); + long i = 0, index = 0; + + p[index] = i; + for (;;) { + if (++index < r-1) { + p[index] = i; + continue; } - } - else { - VALUE result = rb_ary_new2(r); - VALUE *result_array = RARRAY_PTR(result); - const VALUE *values_array = RARRAY_PTR(values); - - for (j = 0; j < r; ++j) result_array[j] = values_array[p[j]]; - ARY_SET_LEN(result, r); - rb_yield(result); - if (RBASIC(values)->klass) { - rb_raise(rb_eRuntimeError, "repeated combination reentered"); + 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); } } +static VALUE +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 binomial_coefficient(k, n + k - 1); +} + /* * call-seq: - * ary.repeated_combination(n) { |c| block } -> ary - * ary.repeated_combination(n) -> an_enumerator - * - * When invoked with a block, yields all repeated combinations of - * length <i>n</i> of elements from <i>ary</i> and then returns - * <i>ary</i> itself. - * The implementation makes no guarantees about the order in which - * the repeated combinations are yielded. - * - * If no block is given, an enumerator is returned instead. - * - * Examples: - * - * 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 - * + * array.repeated_combination(n) {|combination| ... } -> self + * array.repeated_combination(n) -> new_enumerator + * + * 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. + * + * 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>. + * + * +n+ = 1: + * a = [0, 1, 2] + * a.repeated_combination(1) {|combination| p combination } + * 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]] */ static VALUE @@ -4416,7 +7097,7 @@ rb_ary_repeated_combination(VALUE ary, VALUE num) long n, i, len; n = NUM2LONG(num); /* Combination size from argument */ - RETURN_ENUMERATOR(ary, 1, &num); /* Return enumerator if no block */ + 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 */ @@ -4425,44 +7106,73 @@ rb_ary_repeated_combination(VALUE ary, VALUE num) rb_yield(rb_ary_new2(0)); } else if (n == 1) { - for (i = 0; i < len; i++) { - rb_yield(rb_ary_new3(1, RARRAY_PTR(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 */ } else { - volatile VALUE t0 = tmpbuf(n, sizeof(long)); - long *p = (long*)RSTRING_PTR(t0); + volatile VALUE t0; + long *p = ALLOCV_N(long, t0, n); VALUE ary0 = ary_make_shared_copy(ary); /* private defensive copy of ary */ - RBASIC(ary0)->klass = 0; + RBASIC_CLEAR_CLASS(ary0); - rcombinate0(len, n, p, 0, n, ary0); /* compute and yield repeated combinations */ - tmpbuf_discard(t0); - RBASIC(ary0)->klass = 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 - * - * Returns an array of all combinations of elements from all arrays. - * The length of the returned array is the product of the length - * of +self+ and the argument arrays. - * If given a block, <i>product</i> will yield all combinations - * and return +self+ instead. - * - * - * [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([]) #=> [] + * 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+: + * a.product(a1) {|combination| p combination } + * Output: + * [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] */ static VALUE @@ -4470,15 +7180,14 @@ 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 t1 = tmpbuf(n, sizeof(int)); + volatile VALUE t1 = Qundef; VALUE *arrays = RARRAY_PTR(t0); /* The arrays we're computing the product of */ - int *counters = (int*)RSTRING_PTR(t1); /* The current position in each one */ + int *counters = ALLOCV_N(int, t1, n); /* The current position in each one */ VALUE result = Qnil; /* The array we'll be returning, when no block given */ long i,j; long resultlen = 1; - RBASIC(t0)->klass = 0; - RBASIC(t1)->klass = 0; + RBASIC_CLEAR_CLASS(t0); /* initialize the arrays of arrays */ ARY_SET_LEN(t0, n); @@ -4500,15 +7209,14 @@ rb_ary_product(int argc, VALUE *argv, VALUE ary) 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]), l = resultlen; + long k = RARRAY_LEN(arrays[i]); if (k == 0) { result = rb_ary_new2(0); goto done; } - resultlen *= k; - if (resultlen < k || resultlen < l || resultlen / k != l) { + if (MUL_OVERFLOW_LONG_P(resultlen, k)) rb_raise(rb_eRangeError, "too big to product"); - } + resultlen *= k; } result = rb_ary_new2(resultlen); } @@ -4521,7 +7229,7 @@ rb_ary_product(int argc, VALUE *argv, VALUE ary) } /* put it on the result array */ - if(NIL_P(result)) { + if (NIL_P(result)) { FL_SET(t0, FL_USER5); rb_yield(subarray); if (! FL_TEST(t0, FL_USER5)) { @@ -4550,20 +7258,25 @@ rb_ary_product(int argc, VALUE *argv, VALUE ary) } done: tmpary_discard(t0); - tmpbuf_discard(t1); + ALLOCV_END(t1); return NIL_P(result) ? ary : result; } /* * call-seq: - * ary.take(n) -> new_ary - * - * Returns first n elements from <i>ary</i>. - * - * a = [1, 2, 3, 4, 5, 0] - * a.take(3) #=> [1, 2, 3] - * + * array.take(n) -> new_array + * + * Returns a new \Array containing the first +n+ element of +self+, + * where +n+ is a non-negative \Integer; + * does not modify +self+. + * + * Examples: + * 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] */ static VALUE @@ -4578,17 +7291,22 @@ rb_ary_take(VALUE obj, VALUE n) /* * call-seq: - * ary.take_while {|arr| block } -> new_ary - * ary.take_while -> an_enumerator - * - * Passes elements to the block until the block returns +nil+ or +false+, - * then stops iterating and returns an array of all prior elements. - * - * If no block is given, an enumerator is returned instead. - * - * a = [1, 2, 3, 4, 5, 0] - * a.take_while {|i| i < 3 } #=> [1, 2] - * + * 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+. + * + * 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: + * 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] + * + * With no block given, returns a new \Enumerator: + * [0, 1].take_while # => #<Enumerator: [0, 1]:take_while> */ static VALUE @@ -4598,21 +7316,24 @@ rb_ary_take_while(VALUE ary) RETURN_ENUMERATOR(ary, 0, 0); for (i = 0; i < RARRAY_LEN(ary); i++) { - if (!RTEST(rb_yield(RARRAY_PTR(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 - * - * Drops first n elements from +ary+ and returns the rest of - * the elements in an array. + * array.drop(n) -> new_array * - * a = [1, 2, 3, 4, 5, 0] - * a.drop(3) #=> [4, 5, 0] + * Returns a new \Array containing all but the first +n+ element of +self+, + * where +n+ is a non-negative \Integer; + * does not modify +self+. * + * Examples: + * 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] */ static VALUE @@ -4625,24 +7346,26 @@ rb_ary_drop(VALUE ary, VALUE n) } 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 {|arr| block } -> new_ary - * ary.drop_while -> an_enumerator - * - * 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. - * - * If no block is given, an enumerator is returned instead. + * 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+. * - * a = [1, 2, 3, 4, 5, 0] - * a.drop_while {|i| i < 3 } #=> [3, 4, 5, 0] + * 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: + * a = [0, 1, 2, 3, 4, 5] + * a.drop_while {|element| element < 3 } # => [3, 4, 5] * + * With no block given, returns a new \Enumerator: + * [0, 1].drop_while # => # => #<Enumerator: [0, 1]:drop_while> */ static VALUE @@ -4652,30 +7375,895 @@ rb_ary_drop_while(VALUE ary) RETURN_ENUMERATOR(ary, 0, 0); for (i = 0; i < RARRAY_LEN(ary); i++) { - if (!RTEST(rb_yield(RARRAY_PTR(ary)[i]))) break; + if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) break; } return rb_ary_drop(ary, LONG2FIX(i)); } +/* + * call-seq: + * 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: + * ['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 +rb_ary_any_p(int argc, VALUE *argv, VALUE ary) +{ + long i, len = RARRAY_LEN(ary); + + rb_check_arity(argc, 0, 1); + if (!len) return Qfalse; + if (argc) { + 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; + } + } + else if (!rb_block_given_p()) { + for (i = 0; i < len; ++i) { + if (RTEST(RARRAY_AREF(ary, i))) return Qtrue; + } + } + else { + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) return Qtrue; + } + } + return Qfalse; +} + +/* + * call-seq: + * 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 + * + * 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 +rb_ary_all_p(int argc, VALUE *argv, VALUE ary) +{ + long i, len = RARRAY_LEN(ary); + + rb_check_arity(argc, 0, 1); + if (!len) return Qtrue; + if (argc) { + 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 Qfalse; + } + } + else if (!rb_block_given_p()) { + for (i = 0; i < len; ++i) { + if (!RTEST(RARRAY_AREF(ary, i))) return Qfalse; + } + } + else { + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (!RTEST(rb_yield(RARRAY_AREF(ary, i)))) return Qfalse; + } + } + return Qtrue; +} + +/* + * call-seq: + * 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: + * [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 +rb_ary_none_p(int argc, VALUE *argv, VALUE ary) +{ + long i, len = RARRAY_LEN(ary); + + rb_check_arity(argc, 0, 1); + if (!len) return Qtrue; + if (argc) { + 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 Qfalse; + } + } + else if (!rb_block_given_p()) { + for (i = 0; i < len; ++i) { + if (RTEST(RARRAY_AREF(ary, i))) return Qfalse; + } + } + else { + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) return Qfalse; + } + } + return Qtrue; +} + +/* + * call-seq: + * 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: + * [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 +rb_ary_one_p(int argc, VALUE *argv, VALUE ary) +{ + long i, len = RARRAY_LEN(ary); + VALUE result = Qfalse; + + rb_check_arity(argc, 0, 1); + if (!len) return Qfalse; + if (argc) { + 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)))) { + if (result) return Qfalse; + result = Qtrue; + } + } + } + else if (!rb_block_given_p()) { + for (i = 0; i < len; ++i) { + if (RTEST(RARRAY_AREF(ary, i))) { + if (result) return Qfalse; + result = Qtrue; + } + } + } + else { + for (i = 0; i < RARRAY_LEN(ary); ++i) { + if (RTEST(rb_yield(RARRAY_AREF(ary, i)))) { + if (result) return Qfalse; + result = Qtrue; + } + } + } + return result; +} + +/* + * 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]. + * + * Examples: + * 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 + */ + +static VALUE +rb_ary_dig(int argc, VALUE *argv, VALUE self) +{ + rb_check_arity(argc, 1, UNLIMITED_ARGUMENTS); + self = rb_ary_at(self, *argv); + if (!--argc) return self; + ++argv; + return rb_obj_dig(argc, argv, self, Qnil); +} + +static inline VALUE +finish_exact_sum(long n, VALUE r, VALUE v, int z) +{ + if (n != 0) + v = rb_fix_plus(LONG2FIX(n), v); + if (r != Qundef) { + v = rb_rational_plus(r, v); + } + else if (!n && z) { + v = rb_fix_plus(LONG2FIX(0), v); + } + return v; +} + +/* + * call-seq: + * array.sum(init = 0) -> object + * array.sum(init = 0) {|element| ... } -> object + * + * When no block is given, returns the object equivalent to: + * sum = init + * array.each {|element| sum += element } + * sum + * For example, <tt>[e1, e2, e3].sum</tt> returns <tt>init + e1 + e2 + e3</tt>. + * + * Examples: + * a = [0, 1, 2, 3] + * a.sum # => 6 + * a.sum(100) # => 106 + * + * The elements need not be numeric, but must be <tt>+</tt>-compatible + * with each other and with +init+: + * a = ['abc', 'def', 'ghi'] + * a.sum('jkl') # => "jklabcdefghi" + * + * 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: + * a = ['zero', 1, :two] + * s = a.sum('Coerced and concatenated: ') {|element| element.to_s } + * s # => "Coerced and concatenated: zero1two" + * + * Notes: + * - 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#+. + */ + +static VALUE +rb_ary_sum(int argc, VALUE *argv, VALUE ary) +{ + VALUE e, v, r; + long i, n; + int block_given; + + v = (rb_check_arity(argc, 0, 1) ? argv[0] : LONG2FIX(0)); + + block_given = rb_block_given_p(); + + if (RARRAY_LEN(ary) == 0) + return v; + + n = 0; + r = Qundef; + for (i = 0; i < RARRAY_LEN(ary); i++) { + e = RARRAY_AREF(ary, i); + if (block_given) + e = rb_yield(e); + if (FIXNUM_P(e)) { + n += FIX2LONG(e); /* should not overflow long type */ + if (!FIXABLE(n)) { + v = rb_big_plus(LONG2NUM(n), v); + n = 0; + } + } + else if (RB_BIGNUM_TYPE_P(e)) + v = rb_big_plus(e, v); + else if (RB_TYPE_P(e, T_RATIONAL)) { + if (r == Qundef) + r = e; + else + r = rb_rational_plus(r, e); + } + else + goto not_exact; + } + v = finish_exact_sum(n, r, v, argc!=0); + return v; + not_exact: + v = finish_exact_sum(n, r, v, i!=0); + + if (RB_FLOAT_TYPE_P(e)) { + /* + * Kahan-Babuska balancing compensated summation algorithm + * 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++) { + e = RARRAY_AREF(ary, i); + if (block_given) + e = rb_yield(e); + if (RB_FLOAT_TYPE_P(e)) + has_float_value: + x = RFLOAT_VALUE(e); + else if (FIXNUM_P(e)) + x = FIX2LONG(e); + else if (RB_BIGNUM_TYPE_P(e)) + x = rb_big2dbl(e); + else if (RB_TYPE_P(e, T_RATIONAL)) + x = rb_num2dbl(e); + else + goto not_float; + + if (isnan(f)) continue; + if (isnan(x)) { + f = x; + continue; + } + if (isinf(x)) { + if (isinf(f) && signbit(x) != signbit(f)) + f = NAN; + else + f = x; + continue; + } + if (isinf(f)) continue; + + t = f + x; + if (fabs(f) >= fabs(x)) + c += ((f - t) + x); + else + c += ((x - t) + f); + f = t; + } + f += c; + return DBL2NUM(f); -/* Arrays are ordered, integer-indexed collections of any object. - * 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. + not_float: + v = DBL2NUM(f); + } + + goto has_some_value; + for (; i < RARRAY_LEN(ary); i++) { + e = RARRAY_AREF(ary, i); + if (block_given) + e = rb_yield(e); + has_some_value: + v = rb_funcall(v, idPLUS, 1, e); + } + return v; +} + +static VALUE +rb_ary_deconstruct(VALUE ary) +{ + return ary; +} + +/* + * An \Array is an ordered, integer-indexed collection of objects, + * called _elements_. Any object may be an \Array element. + * + * == \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 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 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. + * + * == Creating Arrays + * + * You can create an \Array object explicitly with: + * + * - An {array literal}[doc/syntax/literals_rdoc.html#label-Array+Literals]. + * + * You can convert certain objects to Arrays with: + * + * - \Method {Array}[Kernel.html#method-i-Array]. + * + * An \Array can contain different types of objects. For + * example, the array below contains an Integer, a String and a Float: + * + * ary = [1, "two", 3.0] #=> [1, "two", 3.0] + * + * An array can also be created by calling Array.new with zero, one + * (the initial size of the Array) or two arguments (the initial size and a + * default object). + * + * ary = Array.new #=> [] + * Array.new(3) #=> [nil, nil, nil] + * Array.new(3, true) #=> [true, true, true] + * + * 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. + * + * 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: + * + * Array.new(4) {Hash.new} #=> [{}, {}, {}, {}] + * Array.new(4) {|i| i.to_s } #=> ["0", "1", "2", "3"] + * + * This is also a quick way to build up multi-dimensional arrays: + * + * empty_table = Array.new(3) {Array.new(3)} + * #=> [[nil, nil, nil], [nil, nil, nil], [nil, nil, nil]] + * + * 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({:a => "a", :b => "b"}) #=> [[:a, "a"], [:b, "b"]] + * + * == 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 + * manipulating arrays. + * + * Some of the more common ones are illustrated below. + * + * == Accessing Elements + * + * Elements in an array can be retrieved using the Array#[] method. It can + * take a single integer argument (a numeric index), a pair of arguments + * (start and length) or a range. Negative indices start counting from the end, + * with -1 being the last element. + * + * arr = [1, 2, 3, 4, 5, 6] + * arr[2] #=> 3 + * arr[100] #=> nil + * arr[-3] #=> 4 + * arr[2, 3] #=> [3, 4, 5] + * arr[1..4] #=> [2, 3, 4, 5] + * arr[1..-3] #=> [2, 3, 4] + * + * Another way to access a particular array element is by using the #at method + * + * arr.at(0) #=> 1 + * + * The #slice method works in an identical manner to Array#[]. + * + * To raise an error for indices outside of the array bounds or else to + * provide a default value when that happens, you can use #fetch. + * + * arr = ['a', 'b', 'c', 'd', 'e', 'f'] + * arr.fetch(100) #=> IndexError: index 100 outside of array bounds: -6...6 + * arr.fetch(100, "oops") #=> "oops" + * + * The special methods #first and #last will return the first and last + * elements of an array, respectively. + * + * arr.first #=> 1 + * arr.last #=> 6 + * + * To return the first +n+ elements of an array, use #take + * + * arr.take(3) #=> [1, 2, 3] + * + * #drop does the opposite of #take, by returning the elements after +n+ + * elements have been dropped: + * + * arr.drop(3) #=> [4, 5, 6] + * + * == 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. + * + * browsers = ['Chrome', 'Firefox', 'Safari', 'Opera', 'IE'] + * browsers.length #=> 5 + * browsers.count #=> 5 + * + * To check whether an array contains any elements at all + * + * browsers.empty? #=> false + * + * To check whether a particular item is included in the array + * + * browsers.include?('Konqueror') #=> false + * + * == Adding Items to Arrays + * + * Items can be added to the end of an array by using either #push or #<< + * + * arr = [1, 2, 3, 4] + * arr.push(5) #=> [1, 2, 3, 4, 5] + * arr << 6 #=> [1, 2, 3, 4, 5, 6] + * + * #unshift will add a new item to the beginning of an array. + * + * arr.unshift(0) #=> [0, 1, 2, 3, 4, 5, 6] + * + * With #insert you can add a new element to an array at any position. + * + * arr.insert(3, 'apple') #=> [0, 1, 2, 'apple', 3, 4, 5, 6] + * + * Using the #insert method, you can also insert multiple values at once: + * + * arr.insert(3, 'orange', 'pear', 'grapefruit') + * #=> [0, 1, 2, "orange", "pear", "grapefruit", "apple", 3, 4, 5, 6] + * + * == Removing Items from an Array + * + * The method #pop removes the last element in an array and returns it: + * + * arr = [1, 2, 3, 4, 5, 6] + * arr.pop #=> 6 + * arr #=> [1, 2, 3, 4, 5] + * + * To retrieve and at the same time remove the first item, use #shift: + * + * arr.shift #=> 1 + * arr #=> [2, 3, 4, 5] + * + * To delete an element at a particular index: + * + * arr.delete_at(2) #=> 4 + * arr #=> [2, 3, 5] + * + * To delete a particular element anywhere in an array, use #delete: + * + * arr = [1, 2, 2, 3] + * arr.delete(2) #=> 2 + * arr #=> [1,3] + * + * A useful method if you need to remove +nil+ values from an array is + * #compact: + * + * arr = ['foo', 0, nil, 'bar', 7, 'baz', nil] + * arr.compact #=> ['foo', 0, 'bar', 7, 'baz'] + * arr #=> ['foo', 0, nil, 'bar', 7, 'baz', nil] + * arr.compact! #=> ['foo', 0, 'bar', 7, 'baz'] + * arr #=> ['foo', 0, 'bar', 7, 'baz'] + * + * Another common need is to remove duplicate elements from an array. + * + * It has the non-destructive #uniq, and destructive method #uniq! + * + * arr = [2, 5, 6, 556, 6, 6, 8, 9, 0, 123, 556] + * arr.uniq #=> [2, 5, 6, 556, 8, 9, 0, 123] + * + * == Iterating over Arrays + * + * 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 + * the supplied block in sequence. + * + * Note that this operation leaves the array unchanged. + * + * arr = [1, 2, 3, 4, 5] + * arr.each {|a| print a -= 10, " "} + * # prints: -9 -8 -7 -6 -5 + * #=> [1, 2, 3, 4, 5] + * + * Another sometimes useful iterator is #reverse_each which will iterate over + * the elements in the array in reverse order. + * + * words = %w[first second third fourth fifth sixth] + * str = "" + * words.reverse_each {|word| str += "#{word} "} + * p str #=> "sixth fifth fourth third second first " + * + * The #map method can be used to create a new array based on the original + * array, but with the values modified by the supplied block: + * + * arr.map {|a| 2*a} #=> [2, 4, 6, 8, 10] + * arr #=> [1, 2, 3, 4, 5] + * arr.map! {|a| a**2} #=> [1, 4, 9, 16, 25] + * arr #=> [1, 4, 9, 16, 25] + * + * == 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 + * manner. While the destructive operations will modify the array they were + * called on, the non-destructive methods usually return a new array with the + * selected elements, but leave the original array unchanged. + * + * === Non-destructive Selection + * + * arr = [1, 2, 3, 4, 5, 6] + * arr.select {|a| a > 3} #=> [4, 5, 6] + * arr.reject {|a| a < 3} #=> [3, 4, 5, 6] + * arr.drop_while {|a| a < 4} #=> [4, 5, 6] + * arr #=> [1, 2, 3, 4, 5, 6] + * + * === Destructive Selection + * + * #select! and #reject! are the corresponding destructive methods to #select + * and #reject + * + * Similar to #select vs. #reject, #delete_if and #keep_if have the exact + * opposite result when supplied with the same block: + * + * arr.delete_if {|a| a < 4} #=> [4, 5, 6] + * arr #=> [4, 5, 6] + * + * arr = [1, 2, 3, 4, 5, 6] + * 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}[Object.html#class-Object-label-What-27s+Here]. + * - Includes {module Enumerable}[Enumerable.html#module-Enumerable-label-What-27s+Here], + * which provides dozens of additional methods. + * + * Here, class \Array provides methods that are useful for: + * + * - {Creating an Array}[#class-Array-label-Methods+for+Creating+an+Array] + * - {Querying}[#class-Array-label-Methods+for+Querying] + * - {Comparing}[#class-Array-label-Methods+for+Comparing] + * - {Fetching}[#class-Array-label-Methods+for+Fetching] + * - {Assigning}[#class-Array-label-Methods+for+Assigning] + * - {Deleting}[#class-Array-label-Methods+for+Deleting] + * - {Combining}[#class-Array-label-Methods+for+Combining] + * - {Iterating}[#class-Array-label-Methods+for+Iterating] + * - {Converting}[#class-Array-label-Methods+for+Converting] + * - {And more....}[#class-Array-label-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 + * {#<=>}[#method-i-3C-3D-3E]:: Returns -1, 0, or 1 + * as +self+ is less than, equal to, or greater than a given object. + * {#==}[#method-i-3D-3D]:: 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. + * #max:: 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 non-+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 + * + * {#&}[#method-i-26]:: 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 containiing all elements of +self+ that are not found in a given array. + * {#|}[#method-i-7C]:: 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, ary_alloc); + rb_define_alloc_func(rb_cArray, empty_ary_alloc); 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); @@ -4684,8 +8272,8 @@ Init_Array(void) rb_define_method(rb_cArray, "inspect", rb_ary_inspect, 0); rb_define_alias(rb_cArray, "to_s", "inspect"); rb_define_method(rb_cArray, "to_a", rb_ary_to_a, 0); + rb_define_method(rb_cArray, "to_h", rb_ary_to_h, 0); rb_define_method(rb_cArray, "to_ary", rb_ary_to_ary_m, 0); - rb_define_method(rb_cArray, "frozen?", rb_ary_frozen_p, 0); rb_define_method(rb_cArray, "==", rb_ary_equal, 1); rb_define_method(rb_cArray, "eql?", rb_ary_eql, 1); @@ -4697,18 +8285,24 @@ Init_Array(void) rb_define_method(rb_cArray, "fetch", rb_ary_fetch, -1); rb_define_method(rb_cArray, "first", rb_ary_first, -1); rb_define_method(rb_cArray, "last", rb_ary_last, -1); - rb_define_method(rb_cArray, "concat", rb_ary_concat, 1); + 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"); rb_define_method(rb_cArray, "pop", rb_ary_pop_m, -1); rb_define_method(rb_cArray, "shift", rb_ary_shift_m, -1); rb_define_method(rb_cArray, "unshift", rb_ary_unshift_m, -1); + rb_define_alias(rb_cArray, "prepend", "unshift"); rb_define_method(rb_cArray, "insert", rb_ary_insert, -1); rb_define_method(rb_cArray, "each", rb_ary_each, 0); 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); @@ -4727,6 +8321,8 @@ Init_Array(void) rb_define_method(rb_cArray, "map!", rb_ary_collect_bang, 0); rb_define_method(rb_cArray, "select", rb_ary_select, 0); rb_define_method(rb_cArray, "select!", rb_ary_select_bang, 0); + rb_define_method(rb_cArray, "filter", rb_ary_select, 0); + rb_define_method(rb_cArray, "filter!", rb_ary_select_bang, 0); rb_define_method(rb_cArray, "keep_if", rb_ary_keep_if, 0); rb_define_method(rb_cArray, "values_at", rb_ary_values_at, -1); rb_define_method(rb_cArray, "delete", rb_ary_delete, 1); @@ -4755,6 +8351,10 @@ Init_Array(void) rb_define_method(rb_cArray, "&", rb_ary_and, 1); rb_define_method(rb_cArray, "|", rb_ary_or, 1); + rb_define_method(rb_cArray, "max", rb_ary_max, -1); + rb_define_method(rb_cArray, "min", rb_ary_min, -1); + rb_define_method(rb_cArray, "minmax", rb_ary_minmax, 0); + rb_define_method(rb_cArray, "uniq", rb_ary_uniq, 0); rb_define_method(rb_cArray, "uniq!", rb_ary_uniq_bang, 0); rb_define_method(rb_cArray, "compact", rb_ary_compact, 0); @@ -4762,9 +8362,6 @@ 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); @@ -4776,7 +8373,16 @@ Init_Array(void) rb_define_method(rb_cArray, "take_while", rb_ary_take_while, 0); rb_define_method(rb_cArray, "drop", rb_ary_drop, 1); rb_define_method(rb_cArray, "drop_while", rb_ary_drop_while, 0); + rb_define_method(rb_cArray, "bsearch", rb_ary_bsearch, 0); + rb_define_method(rb_cArray, "bsearch_index", rb_ary_bsearch_index, 0); + rb_define_method(rb_cArray, "any?", rb_ary_any_p, -1); + rb_define_method(rb_cArray, "all?", rb_ary_all_p, -1); + rb_define_method(rb_cArray, "none?", rb_ary_none_p, -1); + rb_define_method(rb_cArray, "one?", rb_ary_one_p, -1); + rb_define_method(rb_cArray, "dig", rb_ary_dig, -1); + rb_define_method(rb_cArray, "sum", rb_ary_sum, -1); - id_cmp = rb_intern("<=>"); - sym_random = ID2SYM(rb_intern("random")); + rb_define_method(rb_cArray, "deconstruct", rb_ary_deconstruct, 0); } + +#include "array.rbinc" |