end
# Immutable strings of characters.
-class FlatString
+abstract class FlatString
super FlatText
super String
redef fun fast_cstring do return _items.fast_cstring(_first_byte)
- redef fun substring_from(from) do
- if from >= self._length then return empty
- if from <= 0 then return self
- var c = char_to_byte_index(from)
- var st = c - _first_byte
- var fln = bytelen - st
- return new FlatString.full(items, fln, c, _length - from)
- end
-
redef fun substring(from, count)
do
if count <= 0 then return ""
#
# `_items` will be used as is, without copy, to retrieve the characters of the string.
# Aliasing issues is the responsibility of the caller.
- private init with_infos(items: NativeString, bytelen, from: Int)
+ private new with_infos(items: NativeString, bytelen, from: Int)
do
- self._items = items
- self._bytelen = bytelen
- _first_byte = from
- _bytepos = from
- _length = _items.utf8_length(_first_byte, bytelen)
+ var len = items.utf8_length(from, bytelen)
+ if bytelen == len then return new ASCIIFlatString.full_data(items, bytelen, from, len)
+ return new UnicodeFlatString.full_data(items, bytelen, from, len)
end
# Low-level creation of a new string with all the data.
#
# `_items` will be used as is, without copy, to retrieve the characters of the string.
# Aliasing issues is the responsibility of the caller.
- private init full(items: NativeString, bytelen, from, length: Int)
+ private new full(items: NativeString, bytelen, from, length: Int)
do
- self._items = items
- self._length = length
- self._bytelen = bytelen
- _first_byte = from
- _bytepos = from
+ if bytelen == length then return new ASCIIFlatString.full_data(items, bytelen, from, length)
+ return new UnicodeFlatString.full_data(items, bytelen, from, length)
end
redef fun ==(other)
return new FlatString.full(ns, new_bytelen, 0, newlen)
end
-
redef fun hash
do
if hash_cache == null then
redef fun substrings do return new FlatSubstringsIter(self)
end
+# Regular Nit UTF-8 strings
+private class UnicodeFlatString
+ super FlatString
+
+ init full_data(items: NativeString, bytelen, from, length: Int) do
+ self._items = items
+ self._length = length
+ self._bytelen = bytelen
+ _first_byte = from
+ _bytepos = from
+ end
+
+ redef fun substring_from(from) do
+ if from >= self._length then return empty
+ if from <= 0 then return self
+ var c = char_to_byte_index(from)
+ var st = c - _first_byte
+ var fln = bytelen - st
+ return new FlatString.full(items, fln, c, _length - from)
+ end
+end
+
+# Special cases of String where all the characters are ASCII-based
+#
+# Optimizes access operations to O(1) complexity.
+private class ASCIIFlatString
+ super FlatString
+
+ init full_data(items: NativeString, bytelen, from, length: Int) do
+ self._items = items
+ self._length = length
+ self._bytelen = bytelen
+ _first_byte = from
+ _bytepos = from
+ end
+
+ redef fun [](idx) do
+ assert idx < _bytelen and idx >= 0
+ return _items[idx + _first_byte].ascii
+ end
+
+ redef fun substring(from, count) do
+ if count <= 0 then return ""
+
+ if from < 0 then
+ count += from
+ if count < 0 then return ""
+ from = 0
+ end
+ var ln = _length
+ if (count + from) > ln then count = ln - from
+ return new ASCIIFlatString.full_data(_items, count, from + _first_byte, count)
+ end
+
+ redef fun reversed do
+ var b = new FlatBuffer.with_capacity(_bytelen + 1)
+ var i = _length - 1
+ while i >= 0 do
+ b.add self[i]
+ i -= 1
+ end
+ var s = b.to_s.as(FlatString)
+ return s
+ end
+
+ redef fun char_to_byte_index(index) do return index + _first_byte
+
+ redef fun substring_impl(from, count, end_index) do
+ return new ASCIIFlatString.full_data(_items, count, from + _first_byte, count)
+ end
+
+ redef fun fetch_char_at(i) do return _items[i + _first_byte].ascii
+end
+
private class FlatStringCharReverseIterator
super IndexedIterator[Char]
redef fun clear do
is_dirty = true
- if written then reset
_bytelen = 0
_length = 0
+ if written then reset
end
redef fun empty do return new Buffer
do
var c = capacity
if cap <= c then return
- while c <= cap do c = c * 2 + 2
+ if c <= 16 then c = 16
+ while c <= cap do c = c * 2
# The COW flag can be set at false here, since
# it does a copy of the current `Buffer`
written = false
var bln = _bytelen
- var a = new NativeString(c+1)
+ var a = new NativeString(c)
if bln > 0 then
var it = _items
if bln > 0 then it.copy_to(a, bln, 0, 0)
init from(s: Text)
do
_items = new NativeString(s.bytelen)
- if s isa FlatText then
- _items = s._items
- else
- for i in substrings do i.as(FlatString)._items.copy_to(_items, i._bytelen, 0, 0)
- end
+ for i in s.substrings do i._items.copy_to(_items, i._bytelen, first_byte, 0)
_bytelen = s.bytelen
_length = s.length
_capacity = _bytelen
- written = true
end
# Create a new empty string with a given capacity.
init with_capacity(cap: Int)
do
assert cap >= 0
- _items = new NativeString(cap + 1)
+ _items = new NativeString(cap)
capacity = cap
_bytelen = 0
end
return to_s_with_length(cstring_length)
end
- # Returns `self` as a String of `length`.
- redef fun to_s_with_length(length): FlatString
+ redef fun to_s_with_length(length)
do
assert length >= 0
return clean_utf8(length)
return new FlatString.with_infos(self, len, 0)
end
- # Returns `self` as a new String.
- redef fun to_s_with_copy: FlatString
+ redef fun to_s_with_copy do return to_s_with_copy_and_length(cstring_length)
+
+ # Get a `String` from `length` bytes at `self` copied into Nit memory
+ fun to_s_with_copy_and_length(length: Int): String
do
- var length = cstring_length
var r = clean_utf8(length)
if r.items != self then return r
var new_self = new NativeString(length + 1)
nitlanguage / nit.git / blobdiff