mirror of
https://github.com/nim-lang/Nim.git
synced 2026-02-15 15:44:14 +00:00
141 lines
3.5 KiB
Nim
141 lines
3.5 KiB
Nim
#
|
|
#
|
|
# Nim's Runtime Library
|
|
# (c) Copyright 2019 Andreas Rumpf
|
|
#
|
|
# See the file "copying.txt", included in this
|
|
# distribution, for details about the copyright.
|
|
#
|
|
|
|
# An ``include`` file for the different hash set implementations.
|
|
|
|
|
|
template maxHash(t): untyped = high(t.data)
|
|
template dataLen(t): untyped = len(t.data)
|
|
|
|
include hashcommon
|
|
|
|
template initImpl(s: typed, size: int) =
|
|
assert isPowerOfTwo(size)
|
|
when s is OrderedSet:
|
|
s.first = -1
|
|
s.last = -1
|
|
s.counter = 0
|
|
newSeq(s.data, size)
|
|
|
|
template rawInsertImpl() {.dirty.} =
|
|
if data.len == 0:
|
|
initImpl(s, defaultInitialSize)
|
|
data[h].key = key
|
|
data[h].hcode = hc
|
|
|
|
proc rawInsert[A](s: var HashSet[A], data: var KeyValuePairSeq[A], key: A,
|
|
hc: Hash, h: Hash) =
|
|
rawInsertImpl()
|
|
|
|
proc enlarge[A](s: var HashSet[A]) =
|
|
var n: KeyValuePairSeq[A]
|
|
newSeq(n, len(s.data) * growthFactor)
|
|
swap(s.data, n) # n is now old seq
|
|
for i in countup(0, high(n)):
|
|
if isFilled(n[i].hcode):
|
|
var j = -1 - rawGetKnownHC(s, n[i].key, n[i].hcode)
|
|
rawInsert(s, s.data, n[i].key, n[i].hcode, j)
|
|
|
|
template inclImpl() {.dirty.} =
|
|
if s.data.len == 0:
|
|
initImpl(s, defaultInitialSize)
|
|
var hc: Hash
|
|
var index = rawGet(s, key, hc)
|
|
if index < 0:
|
|
if mustRehash(s):
|
|
enlarge(s)
|
|
index = rawGetKnownHC(s, key, hc)
|
|
rawInsert(s, s.data, key, hc, -1 - index)
|
|
inc(s.counter)
|
|
|
|
template containsOrInclImpl() {.dirty.} =
|
|
if s.data.len == 0:
|
|
initImpl(s, defaultInitialSize)
|
|
var hc: Hash
|
|
var index = rawGet(s, key, hc)
|
|
if index >= 0:
|
|
result = true
|
|
else:
|
|
if mustRehash(s):
|
|
enlarge(s)
|
|
index = rawGetKnownHC(s, key, hc)
|
|
rawInsert(s, s.data, key, hc, -1 - index)
|
|
inc(s.counter)
|
|
|
|
proc exclImpl[A](s: var HashSet[A], key: A): bool {.inline.} =
|
|
var hc: Hash
|
|
var i = rawGet(s, key, hc)
|
|
var msk = high(s.data)
|
|
result = true
|
|
|
|
if i >= 0:
|
|
result = false
|
|
dec(s.counter)
|
|
inc(s.countDeleted)
|
|
s.data[i].hcode = deletedMarker
|
|
s.data[i].key = default(type(s.data[i].key))
|
|
|
|
template dollarImpl() {.dirty.} =
|
|
result = "{"
|
|
for key in items(s):
|
|
if result.len > 1: result.add(", ")
|
|
result.addQuoted(key)
|
|
result.add("}")
|
|
|
|
|
|
|
|
# --------------------------- OrderedSet ------------------------------
|
|
|
|
proc rawGet[A](t: OrderedSet[A], key: A, hc: var Hash): int {.inline.} =
|
|
rawGetImpl()
|
|
|
|
proc rawInsert[A](s: var OrderedSet[A], data: var OrderedKeyValuePairSeq[A],
|
|
key: A, hc: Hash, h: Hash) =
|
|
rawInsertImpl()
|
|
data[h].next = -1
|
|
if s.first < 0: s.first = h
|
|
if s.last >= 0: data[s.last].next = h
|
|
s.last = h
|
|
|
|
proc enlarge[A](s: var OrderedSet[A]) =
|
|
var n: OrderedKeyValuePairSeq[A]
|
|
newSeq(n, len(s.data) * growthFactor)
|
|
var h = s.first
|
|
s.first = -1
|
|
s.last = -1
|
|
swap(s.data, n)
|
|
while h >= 0:
|
|
var nxt = n[h].next
|
|
if isFilled(n[h].hcode):
|
|
var j = -1 - rawGetKnownHC(s, n[h].key, n[h].hcode)
|
|
rawInsert(s, s.data, n[h].key, n[h].hcode, j)
|
|
h = nxt
|
|
|
|
proc exclImpl[A](s: var OrderedSet[A], key: A): bool {.inline.} =
|
|
if len(s.data) == 0:
|
|
return true
|
|
var n: OrderedKeyValuePairSeq[A]
|
|
newSeq(n, len(s.data))
|
|
var h = s.first
|
|
s.first = -1
|
|
s.last = -1
|
|
swap(s.data, n)
|
|
let hc = genHash(key)
|
|
result = true
|
|
while h >= 0:
|
|
var nxt = n[h].next
|
|
if isFilled(n[h].hcode):
|
|
if n[h].hcode == hc and n[h].key == key:
|
|
dec s.counter
|
|
result = false
|
|
else:
|
|
var j = -1 - rawGetKnownHC(s, n[h].key, n[h].hcode)
|
|
rawInsert(s, s.data, n[h].key, n[h].hcode, j)
|
|
h = nxt
|