redef class FlatText
- private fun first_byte: Int do return 0
+ # First byte of the NativeString
+ protected fun first_byte: Int do return 0
- private fun last_byte: Int do return _bytelen - 1
+ # Last byte of the NativeString
+ protected fun last_byte: Int do return first_byte + _bytelen - 1
# Cache of the latest position (char) explored in the string
- private var position: Int = 0
+ var position: Int = 0
# Cached position (bytes) in the NativeString underlying the String
- private var bytepos: Int = 0
+ var bytepos: Int = 0
# Index of the character `index` in `_items`
- private fun char_to_byte_index(index: Int): Int do
- var ln = length
- assert index >= 0
- assert index < ln
+ fun char_to_byte_index(index: Int): Int do
+ var dpos = index - _position
+ var b = _bytepos
+
+ if dpos == 0 then return b
+ if dpos == 1 then
+ b += _items.length_of_char_at(b)
+ _bytepos = b
+ _position = index
+ return b
+ end
+ if dpos == -1 then
+ b = _items.find_beginning_of_char_at(b - 1)
+ _bytepos = b
+ _position = index
+ return b
+ end
+ var ln = _length
var pos = _position
# Find best insertion point
var delta_begin = index
var ns_i: Int
var my_i: Int
- if min == delta_begin then
- ns_i = first_byte
- my_i = 0
- else if min == delta_cache then
+ if min == delta_cache then
ns_i = _bytepos
my_i = pos
+ else if min == delta_begin then
+ ns_i = first_byte
+ my_i = 0
else
ns_i = its.find_beginning_of_char_at(last_byte)
- my_i = length - 1
+ my_i = _length - 1
end
ns_i = its.char_to_byte_index_cached(index, my_i, ns_i)
return ns_i
end
+ # By escaping `self` to HTML, how many more bytes will be needed ?
+ fun chars_to_html_escape: Int do
+ var its = _items
+ var max = last_byte
+ var pos = first_byte
+ var endlen = 0
+ while pos <= max do
+ var c = its[pos]
+ if c == 0x3Cu8 then
+ endlen += 3
+ else if c == 0x3Eu8 then
+ endlen += 3
+ else if c == 0x26u8 then
+ endlen += 4
+ else if c == 0x22u8 then
+ endlen += 4
+ else if c == 0x27u8 then
+ endlen += 4
+ else if c == 0x2Fu8 then
+ endlen += 4
+ end
+ pos += 1
+ end
+ return endlen
+ end
+
+ redef fun html_escape
+ do
+ var extra = chars_to_html_escape
+ if extra == 0 then return to_s
+ var its = _items
+ var max = last_byte
+ var pos = first_byte
+ var nlen = extra + _bytelen
+ var nits = new NativeString(nlen)
+ var outpos = 0
+ while pos <= max do
+ var c = its[pos]
+ # Special codes:
+ # Some HTML characters are used as meta-data, they need
+ # to be replaced by an HTML-Escaped equivalent
+ #
+ # * 0x3C (<) => <
+ # * 0x3E (>) => >
+ # * 0x26 (&) => &
+ # * 0x22 (") => "
+ # * 0x27 (') => '
+ # * 0x2F (/) => /
+ if c == 0x3Cu8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x6Cu8
+ nits[outpos + 2] = 0x74u8
+ nits[outpos + 3] = 0x3Bu8
+ outpos += 4
+ else if c == 0x3Eu8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x67u8
+ nits[outpos + 2] = 0x74u8
+ nits[outpos + 3] = 0x3Bu8
+ outpos += 4
+ else if c == 0x26u8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x61u8
+ nits[outpos + 2] = 0x6Du8
+ nits[outpos + 3] = 0x70u8
+ nits[outpos + 4] = 0x3Bu8
+ outpos += 5
+ else if c == 0x22u8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x23u8
+ nits[outpos + 2] = 0x33u8
+ nits[outpos + 3] = 0x34u8
+ nits[outpos + 4] = 0x3Bu8
+ outpos += 5
+ else if c == 0x27u8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x23u8
+ nits[outpos + 2] = 0x33u8
+ nits[outpos + 3] = 0x39u8
+ nits[outpos + 4] = 0x3Bu8
+ outpos += 5
+ else if c == 0x2Fu8 then
+ nits[outpos] = 0x26u8
+ nits[outpos + 1] = 0x23u8
+ nits[outpos + 2] = 0x34u8
+ nits[outpos + 3] = 0x37u8
+ nits[outpos + 4] = 0x3Bu8
+ outpos += 5
+ else
+ nits[outpos] = c
+ outpos += 1
+ end
+ pos += 1
+ end
+ var s = new FlatString.with_infos(nits, nlen, 0)
+ return s
+ end
+
# By escaping `self` to C, how many more bytes will be needed ?
#
# This enables a double-optimization in `escape_to_c` since if this
# method returns 0, then `self` does not need escaping and can be
# returned as-is
- protected fun chars_to_escape_to_c: Int do
+ fun chars_to_escape_to_c: Int do
var its = _items
var max = last_byte
var pos = first_byte
end
pos += 1
end
- return nns.to_s_with_length(nlen)
+ return nns.to_s_unsafe(nlen)
+ end
+
+ redef fun [](index) do
+ assert index >= 0 and index < _length
+ return fetch_char_at(index)
end
- redef fun [](index) do return _items.char_at(char_to_byte_index(index))
+ # Gets a `Char` at `index` in `self`
+ #
+ # WARNING: Use at your own risks as no bound-checking is done
+ fun fetch_char_at(index: Int): Char do
+ var i = char_to_byte_index(index)
+ var items = _items
+ var b = items[i]
+ if b & 0x80u8 == 0x00u8 then return b.ascii
+ return items.char_at(i)
+ end
+
+ # If `self` contains only digits and alpha <= 'f', return the corresponding integer.
+ #
+ # assert "ff".to_hex == 255
+ redef fun to_hex(pos, ln) do
+ var res = 0
+ if pos == null then pos = 0
+ if ln == null then ln = length - pos
+ pos = char_to_byte_index(pos)
+ var its = _items
+ var max = pos + ln
+ for i in [pos .. max[ do
+ res <<= 4
+ res += its[i].ascii.from_hex
+ end
+ return res
+ end
end
# Immutable strings of characters.
# Index at which `self` begins in `_items`, inclusively
redef var first_byte is noinit
- # Index at which `self` ends in `_items`, inclusively
- redef var last_byte is noinit
-
redef var chars = new FlatStringCharView(self) is lazy
redef var bytes = new FlatStringByteView(self) is lazy
- redef var length is lazy do
- if _bytelen == 0 then return 0
- return _items.utf8_length(_first_byte, _last_byte)
+ redef var to_cstring is lazy do
+ var blen = _bytelen
+ var new_items = new NativeString(blen + 1)
+ _items.copy_to(new_items, blen, _first_byte, 0)
+ new_items[blen] = 0u8
+ return new_items
end
- redef fun reversed
- do
+ redef fun reversed do
var b = new FlatBuffer.with_capacity(_bytelen + 1)
- for i in [length - 1 .. 0].step(-1) do
- b.add self[i]
+ var i = _length - 1
+ while i >= 0 do
+ b.add self.fetch_char_at(i)
+ i -= 1
end
var s = b.to_s.as(FlatString)
- s.length = self.length
+ s._length = self._length
return s
end
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
assert count >= 0
if from < 0 then
count += from
- if count < 0 then count = 0
+ if count < 0 then return ""
from = 0
end
- if (count + from) > length then count = length - from
+ var ln = _length
+ if (count + from) > ln then count = ln - from
if count <= 0 then return ""
var end_index = from + count - 1
+ return substring_impl(from, count, end_index)
+ end
+
+ private fun substring_impl(from, count, end_index: Int): String do
+ var cache = _position
+ var dfrom = (cache - from).abs
+ var dend = (end_index - from).abs
+
+ var bytefrom: Int
+ var byteto: Int
+ if dfrom < dend then
+ bytefrom = char_to_byte_index(from)
+ byteto = char_to_byte_index(end_index)
+ else
+ byteto = char_to_byte_index(end_index)
+ bytefrom = char_to_byte_index(from)
+ end
- var bytefrom = char_to_byte_index(from)
- var byteto = char_to_byte_index(end_index)
var its = _items
byteto += its.length_of_char_at(byteto) - 1
- var s = new FlatString.full(its, byteto - bytefrom + 1, bytefrom, byteto, count)
+ var s = new FlatString.full(its, byteto - bytefrom + 1, bytefrom, count)
return s
end
do
var outstr = new FlatBuffer.with_capacity(self._bytelen + 1)
- var mylen = length
+ var mylen = _length
var pos = 0
while pos < mylen do
do
var outstr = new FlatBuffer.with_capacity(self._bytelen + 1)
- var mylen = length
+ var mylen = _length
var pos = 0
while pos < mylen do
#
# `_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, to: Int)
+ private init with_infos(items: NativeString, bytelen, from: Int)
do
self._items = items
self._bytelen = bytelen
_first_byte = from
- _last_byte = to
_bytepos = from
+ _length = _items.utf8_length(_first_byte, bytelen)
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, to, length: Int)
+ private init full(items: NativeString, bytelen, from, length: Int)
do
self._items = items
- self.length = length
+ self._length = length
self._bytelen = bytelen
_first_byte = from
- _last_byte = to
_bytepos = from
end
- redef fun to_cstring do
- if real_items != null then return real_items.as(not null)
- var blen = _bytelen
- var new_items = new NativeString(blen + 1)
- _items.copy_to(new_items, blen, _first_byte, 0)
- new_items[blen] = 0u8
- real_items = new_items
- return new_items
- end
-
redef fun ==(other)
do
- if not other isa FlatString then return super
+ if not other isa FlatText then return super
if self.object_id == other.object_id then return true
if other._bytelen != my_length then return false
var my_index = _first_byte
- var its_index = other._first_byte
+ var its_index = other.first_byte
var last_iteration = my_index + my_length
redef fun <(other)
do
- if not other isa FlatString then return super
+ if not other isa FlatText then return super
if self.object_id == other.object_id then return false
- var my_length = self._bytelen
- var its_length = other._bytelen
+ var myits = _items
+ var itsits = other._items
- var max = if my_length < its_length then my_length else its_length
+ var mbt = _bytelen
+ var obt = other.bytelen
- var myits = self.bytes
- var itsits = other.bytes
+ var minln = if mbt < obt then mbt else obt
+ var mst = _first_byte
+ var ost = other.first_byte
- for i in [0 .. max[ do
- var my_curr_char = myits[i]
- var its_curr_char = itsits[i]
+ for i in [0 .. minln[ do
+ var my_curr_char = myits[mst]
+ var its_curr_char = itsits[ost]
- if my_curr_char != its_curr_char then
- if my_curr_char < its_curr_char then return true
- return false
- end
+ if my_curr_char > its_curr_char then return false
+ if my_curr_char < its_curr_char then return true
+
+ mst += 1
+ ost += 1
end
- return my_length < its_length
+ return mbt < obt
end
redef fun +(o) do
var ns = new NativeString(nlen + 1)
mits.copy_to(ns, mlen, mifrom, 0)
sits.copy_to(ns, slen, sifrom, mlen)
- return new FlatString.full(ns, nlen, 0, nlen - 1, length + o.length)
+ return new FlatString.full(ns, nlen, 0, _length + o.length)
else
abort
end
redef fun *(i) do
var mybtlen = _bytelen
var new_bytelen = mybtlen * i
- var mylen = length
+ var mylen = _length
var newlen = mylen * i
var its = _items
var fb = _first_byte
offset += mybtlen
i -= 1
end
- return new FlatString.full(ns, new_bytelen, 0, new_bytelen - 1, newlen)
+ return new FlatString.full(ns, new_bytelen, 0, newlen)
end
var i = _first_byte
var my_items = _items
- var max = _last_byte
+ var max = last_byte
while i <= max do
h = (h << 5) + h + my_items[i].to_i
var curr_pos: Int
- init with_pos(tgt: FlatString, pos: Int)
- do
- init(tgt, pos)
- end
-
redef fun is_ok do return curr_pos >= 0
redef fun item do return target[curr_pos]
var target: FlatString
- var max: Int
+ var max: Int is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatString, pos: Int)
- do
- init(tgt, tgt.length - 1, pos)
- end
+ init do max = target._length - 1
redef fun is_ok do return curr_pos <= max
redef fun [](index) do return target[index]
- redef fun iterator_from(start) do return new FlatStringCharIterator.with_pos(target, start)
+ redef fun iterator_from(start) do return new FlatStringCharIterator(target, start)
- redef fun reverse_iterator_from(start) do return new FlatStringCharReverseIterator.with_pos(target, start)
+ redef fun reverse_iterator_from(start) do return new FlatStringCharReverseIterator(target, start)
end
var target: FlatString
- var target_items: NativeString
+ var target_items: NativeString is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatString, pos: Int)
+ init
do
- init(tgt, tgt._items, pos + tgt._first_byte)
+ var tgt = target
+ target_items = tgt._items
+ curr_pos += tgt._first_byte
end
redef fun is_ok do return curr_pos >= target._first_byte
var target: FlatString
- var target_items: NativeString
+ var target_items: NativeString is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatString, pos: Int)
+ init
do
- init(tgt, tgt._items, pos + tgt._first_byte)
+ var tgt = target
+ target_items = tgt._items
+ curr_pos += tgt._first_byte
end
- redef fun is_ok do return curr_pos <= target._last_byte
+ redef fun is_ok do return curr_pos <= target.last_byte
redef fun item do return target_items[curr_pos]
redef fun [](index)
do
- # Check that the index (+ _first_byte) is not larger than _last_byte
+ # Check that the index (+ _first_byte) is not larger than last_byte
# In other terms, if the index is valid
- assert index >= 0
- var target = self.target
+ var target = _target
+ assert index >= 0 and index < target._bytelen
var ind = index + target._first_byte
- assert ind <= target._last_byte
return target._items[ind]
end
- redef fun iterator_from(start) do return new FlatStringByteIterator.with_pos(target, start)
+ redef fun iterator_from(start) do return new FlatStringByteIterator(target, start)
- redef fun reverse_iterator_from(start) do return new FlatStringByteReverseIterator.with_pos(target, start)
+ redef fun reverse_iterator_from(start) do return new FlatStringByteReverseIterator(target, start)
end
redef var bytes = new FlatBufferByteView(self) is lazy
- redef var length = 0
-
private var char_cache: Int = -1
private var byte_cache: Int = -1
private var capacity = 0
+ # Real items, used as cache for when to_cstring is called
+ private var real_items: NativeString is noinit
+
redef fun fast_cstring do return _items.fast_cstring(0)
redef fun substrings do return new FlatSubstringsIter(self)
redef fun []=(index, item)
do
- assert index >= 0 and index <= length
+ assert index >= 0 and index <= _length
if written then reset
is_dirty = true
- if index == length then
+ if index == _length then
add item
return
end
lshift_bytes(ip + clen, -size_diff)
end
_bytelen += size_diff
- bytepos += size_diff
it.set_char_at(ip, item)
end
enlarge(bt + clen)
_items.set_char_at(bt, c)
_bytelen += clen
- length += 1
+ _length += 1
end
redef fun clear do
is_dirty = true
if written then reset
_bytelen = 0
- length = 0
+ _length = 0
end
redef fun empty do return new Buffer
written = true
var bln = _bytelen
if bln == 0 then _items = new NativeString(1)
- return new FlatString.full(_items, bln, 0, bln - 1, length)
+ return new FlatString.full(_items, bln, 0, _length)
end
redef fun to_cstring
var bln = _bytelen
var new_native = new NativeString(bln + 1)
new_native[bln] = 0u8
- if length > 0 then _items.copy_to(new_native, bln, 0, 0)
+ if _length > 0 then _items.copy_to(new_native, bln, 0, 0)
real_items = new_native
is_dirty = false
end
- return real_items.as(not null)
+ return real_items
end
# Create a new empty string.
self._items = items
self.capacity = capacity
self._bytelen = bytelen
- self.length = length
+ self._length = length
end
# Create a new string copied from `s`.
for i in substrings do i.as(FlatString)._items.copy_to(_items, i._bytelen, 0, 0)
end
_bytelen = s.bytelen
- length = s.length
+ _length = s.length
_capacity = _bytelen
written = true
end
return
end
_bytelen = nln
- length += s.length
+ _length += s.length
end
# Copies the content of self in `dest`
do
assert count >= 0
if from < 0 then from = 0
- if (from + count) > length then count = length - from
- if count != 0 then
- var its = _items
- var bytefrom = its.char_to_byte_index(from)
- var byteto = its.char_to_byte_index(count + from - 1)
- byteto += its.char_at(byteto).u8char_len - 1
- var byte_length = byteto - bytefrom + 1
- var r_items = new NativeString(byte_length)
- its.copy_to(r_items, byte_length, bytefrom, 0)
- return new FlatBuffer.with_infos(r_items, byte_length, byte_length, count)
- else
- return new Buffer
- end
+ if (from + count) > _length then count = _length - from
+ if count <= 0 then return new Buffer
+ var its = _items
+ var bytefrom = its.char_to_byte_index(from)
+ var byteto = its.char_to_byte_index(count + from - 1)
+ byteto += its.char_at(byteto).u8char_len - 1
+ var byte_length = byteto - bytefrom + 1
+ var r_items = new NativeString(byte_length)
+ its.copy_to(r_items, byte_length, bytefrom, 0)
+ return new FlatBuffer.with_infos(r_items, byte_length, byte_length, count)
end
redef fun reverse
redef fun times(repeats)
do
var bln = _bytelen
- var x = new FlatString.full(_items, bln, 0, bln - 1, length)
+ var x = new FlatString.full(_items, bln, 0, _length)
for i in [1 .. repeats[ do
append(x)
end
redef fun upper
do
if written then reset
- for i in [0 .. length[ do self[i] = self[i].to_upper
+ for i in [0 .. _length[ do self[i] = self[i].to_upper
end
redef fun lower
do
if written then reset
- for i in [0 .. length[ do self[i] = self[i].to_lower
+ for i in [0 .. _length[ do self[i] = self[i].to_lower
end
end
var target: FlatBuffer
- var target_items: NativeString
+ var target_items: NativeString is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatBuffer, pos: Int)
- do
- init(tgt, tgt._items, pos)
- end
+ init do target_items = target._items
redef fun index do return curr_pos
redef fun [](index) do return target._items[index]
- redef fun iterator_from(pos) do return new FlatBufferByteIterator.with_pos(target, pos)
+ redef fun iterator_from(pos) do return new FlatBufferByteIterator(target, pos)
- redef fun reverse_iterator_from(pos) do return new FlatBufferByteReverseIterator.with_pos(target, pos)
+ redef fun reverse_iterator_from(pos) do return new FlatBufferByteReverseIterator(target, pos)
end
var target: FlatBuffer
- var target_items: NativeString
+ var target_items: NativeString is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatBuffer, pos: Int)
- do
- init(tgt, tgt._items, pos)
- end
+ init do target_items = target._items
redef fun index do return curr_pos
var curr_pos: Int
- init with_pos(tgt: FlatBuffer, pos: Int)
- do
- init(tgt, pos)
- end
-
redef fun index do return curr_pos
redef fun is_ok do return curr_pos >= 0
redef fun append(s)
do
var s_length = s.length
- if target.capacity < s.length then enlarge(s_length + target.length)
+ if target.capacity < s.length then enlarge(s_length + target._length)
for i in s do target.add i
end
- redef fun iterator_from(pos) do return new FlatBufferCharIterator.with_pos(target, pos)
+ redef fun iterator_from(pos) do return new FlatBufferCharIterator(target, pos)
- redef fun reverse_iterator_from(pos) do return new FlatBufferCharReverseIterator.with_pos(target, pos)
+ redef fun reverse_iterator_from(pos) do return new FlatBufferCharReverseIterator(target, pos)
end
var target: FlatBuffer
- var max: Int
+ var max: Int is noautoinit
var curr_pos: Int
- init with_pos(tgt: FlatBuffer, pos: Int)
- do
- init(tgt, tgt.length - 1, pos)
- end
+ init do max = target._length - 1
redef fun index do return curr_pos
end
redef fun to_s_full(bytelen, unilen) do
- return new FlatString.full(self, bytelen, 0, bytelen - 1, unilen)
+ return new FlatString.full(self, bytelen, 0, unilen)
+ end
+
+ redef fun to_s_unsafe(len) do
+ if len == null then len = cstring_length
+ return new FlatString.with_infos(self, len, 0)
end
# Returns `self` as a new String.
if r.items != self then return r
var new_self = new NativeString(length + 1)
copy_to(new_self, length, 0, 0)
- var str = new FlatString.with_infos(new_self, length, 0, length - 1)
+ var str = new FlatString.with_infos(new_self, length, 0)
new_self[length] = 0u8
- str.real_items = new_self
+ str.to_cstring = new_self
return str
end
var end_length = len
var pos = 0
var chr_ln = 0
- while pos < len do
+ var rem = len
+ while rem > 0 do
+ while rem >= 4 do
+ var i = fetch_4_chars(pos)
+ if i & 0x80808080 != 0 then break
+ pos += 4
+ chr_ln += 4
+ rem -= 4
+ end
+ if rem == 0 then break
var b = self[pos]
+ if b & 0x80u8 == 0x00u8 then
+ pos += 1
+ chr_ln += 1
+ rem -= 1
+ continue
+ end
var nxst = length_of_char_at(pos)
var ok_st: Bool
if nxst == 1 then
replacements.add pos
end_length += 2
pos += 1
+ rem -= 1
chr_ln += 1
continue
end
var ok_c: Bool
var c = char_at(pos)
- var cp = c.ascii
+ var cp = c.code_point
if nxst == 1 then
ok_c = cp >= 0 and cp <= 0x7F
else if nxst == 2 then
end_length += 2
pos += 1
chr_ln += 1
+ rem -= 1
continue
end
- pos += c.u8char_len
+ var clen = c.u8char_len
+ pos += clen
+ rem -= clen
chr_ln += 1
end
var ret = self
end
copy_to(ret, len - old_repl, old_repl, off)
end
- return new FlatString.full(ret, end_length, 0, end_length - 1, chr_ln)
+ return new FlatString.full(ret, end_length, 0, chr_ln)
end
# Sets the next bytes at position `pos` to the value of `c`, encoded in UTF-8
var ns = new NativeString(nslen + 1)
ns[nslen] = 0u8
native_int_to_s(ns, nslen + 1)
- return new FlatString.full(ns, nslen, 0, nslen - 1, nslen)
+ return new FlatString.full(ns, nslen, 0, nslen)
end
end
# Fast implementation
redef fun plain_to_s
do
- var l = length
+ var l = _length
if l == 0 then return ""
- if l == 1 then if self[0] == null then return "" else return self[0].to_s
- var its = _items
+ var its = _items.as(not null)
+ var first = its[0]
+ if l == 1 then if first == null then return "" else return first.to_s
var na = new NativeArray[String](l)
var i = 0
var sl = 0
end
i += 1
end
- return new FlatString.with_infos(ns, sl, 0, sl - 1)
+ return new FlatString.with_infos(ns, sl, 0)
end
end
end
i += 1
end
- return new FlatString.with_infos(ns, sl, 0, sl - 1)
+ return new FlatString.with_infos(ns, sl, 0)
end
end