module string
import math
-intrude import collection # FIXME should be collection::array
+import collection
+intrude import collection::array
`{
#include <stdio.h>
# High-level abstraction for all text representations
abstract class Text
super Comparable
- super StringCapable
redef type OTHER: Text
return self.chars.iterator
end
- # Is 'c' contained in self ?
- #
- # DEPRECATED : Use self.chars.has instead
- fun has(c: Char): Bool
- do
- return self.chars.has(c)
- end
# Gets an Array containing the chars of self
#
# As with substring, a `from` index < 0 will be replaced by 0
fun substring_from(from: Int): SELFTYPE
do
- if from > self.length then return empty
+ if from >= self.length then return empty
if from < 0 then from = 0
return substring(from, length - from)
end
#
# assert "abcd".has_substring("bc",1) == true
# assert "abcd".has_substring("bc",2) == false
+ #
+ # Returns true iff all characters of `str` are presents
+ # at the expected index in `self.`
+ # The first character of `str` being at `pos`, the second
+ # character being at `pos+1` and so on...
+ #
+ # This means that all characters of `str` need to be inside `self`.
+ #
+ # assert "abcd".has_substring("xab", -1) == false
+ # assert "abcd".has_substring("cdx", 2) == false
+ #
+ # And that the empty string is always a valid substring.
+ #
+ # assert "abcd".has_substring("", 2) == true
+ # assert "abcd".has_substring("", 200) == true
fun has_substring(str: String, pos: Int): Bool
do
+ if str.is_empty then return true
+ if pos < 0 or pos + str.length > length then return false
var myiter = self.chars.iterator_from(pos)
var itsiter = str.chars.iterator
while myiter.is_ok and itsiter.is_ok do
var i = 0
var neg = false
- for c in self.chars
- do
+ for j in [0..length[ do
+ var c = chars[j]
var v = c.to_i
if v > base then
if neg then
fun is_numeric: Bool
do
var has_point_or_comma = false
- for i in self.chars
- do
- if not i.is_numeric
- then
- if (i == '.' or i == ',') and not has_point_or_comma
- then
+ for i in [0..length[ do
+ var c = chars[i]
+ if not c.is_numeric then
+ if (c == '.' or c == ',') and not has_point_or_comma then
has_point_or_comma = true
else
return false
return true
end
+ # Returns `true` if the string contains only Hex chars
+ #
+ # assert "048bf".is_hex == true
+ # assert "ABCDEF".is_hex == true
+ # assert "0G".is_hex == false
+ fun is_hex: Bool
+ do
+ for i in [0..length[ do
+ var c = chars[i]
+ if not (c >= 'a' and c <= 'f') and
+ not (c >= 'A' and c <= 'F') and
+ not (c >= '0' and c <= '9') then return false
+ end
+ return true
+ end
+
# Are all letters in `self` upper-case ?
#
# assert "HELLO WORLD".is_upper == true
# assert "Hello World".is_upper == false
fun is_upper: Bool
do
- for char in self.chars do
+ for i in [0..length[ do
+ var char = chars[i]
if char.is_lower then return false
end
return true
# assert "Hello World".is_lower == false
fun is_lower: Bool
do
- for char in self.chars do
+ for i in [0..length[ do
+ var char = chars[i]
if char.is_upper then return false
end
return true
#
# assert " \n\thello \n\t".l_trim == "hello \n\t"
#
- # A whitespace is defined as any character which ascii value is less than or equal to 32
+ # `Char::is_whitespace` determines what is a whitespace.
fun l_trim: SELFTYPE
do
var iter = self.chars.iterator
while iter.is_ok do
- if iter.item.ascii > 32 then break
+ if not iter.item.is_whitespace then break
iter.next
end
if iter.index == length then return self.empty
#
# assert " \n\thello \n\t".r_trim == " \n\thello"
#
- # A whitespace is defined as any character which ascii value is less than or equal to 32
+ # `Char::is_whitespace` determines what is a whitespace.
fun r_trim: SELFTYPE
do
var iter = self.chars.reverse_iterator
while iter.is_ok do
- if iter.item.ascii > 32 then break
+ if not iter.item.is_whitespace then break
iter.next
end
- if iter.index == length then return self.empty
+ if iter.index < 0 then return self.empty
return self.substring(0, iter.index + 1)
end
# Trims trailing and preceding white spaces
- # A whitespace is defined as any character which ascii value is less than or equal to 32
#
# assert " Hello World ! ".trim == "Hello World !"
# assert "\na\nb\tc\t".trim == "a\nb\tc"
+ #
+ # `Char::is_whitespace` determines what is a whitespace.
fun trim: SELFTYPE do return (self.l_trim).r_trim
- # Mangle a string to be a unique string only made of alphanumeric characters
+ # Is the string non-empty but only made of whitespaces?
+ #
+ # assert " \n\t ".is_whitespace == true
+ # assert " hello ".is_whitespace == false
+ # assert "".is_whitespace == false
+ #
+ # `Char::is_whitespace` determines what is a whitespace.
+ fun is_whitespace: Bool
+ do
+ if is_empty then return false
+ for c in self.chars do
+ if not c.is_whitespace then return false
+ end
+ return true
+ end
+
+ # Returns `self` removed from its last line terminator (if any).
+ #
+ # assert "Hello\n".chomp == "Hello"
+ # assert "Hello".chomp == "Hello"
+ #
+ # assert "\n".chomp == ""
+ # assert "".chomp == ""
+ #
+ # Line terminators are `"\n"`, `"\r\n"` and `"\r"`.
+ # A single line terminator, the last one, is removed.
+ #
+ # assert "\r\n".chomp == ""
+ # assert "\r\n\n".chomp == "\r\n"
+ # assert "\r\n\r\n".chomp == "\r\n"
+ # assert "\r\n\r".chomp == "\r\n"
+ #
+ # Note: unlike with most IO methods like `Reader::read_line`,
+ # a single `\r` is considered here to be a line terminator and will be removed.
+ fun chomp: SELFTYPE
+ do
+ var len = length
+ if len == 0 then return self
+ var l = self.chars.last
+ if l == '\r' then
+ return substring(0, len-1)
+ else if l != '\n' then
+ return self
+ else if len > 1 and self.chars[len-2] == '\r' then
+ return substring(0, len-2)
+ else
+ return substring(0, len-1)
+ end
+ end
+
+ # Justify a self in a space of `length`
+ #
+ # `left` is the space ratio on the left side.
+ # * 0.0 for left-justified (no space at the left)
+ # * 1.0 for right-justified (all spaces at the left)
+ # * 0.5 for centered (half the spaces at the left)
+ #
+ # Examples
+ #
+ # assert "hello".justify(10, 0.0) == "hello "
+ # assert "hello".justify(10, 1.0) == " hello"
+ # assert "hello".justify(10, 0.5) == " hello "
+ #
+ # If `length` is not enough, `self` is returned as is.
+ #
+ # assert "hello".justify(2, 0.0) == "hello"
+ #
+ # REQUIRE: `left >= 0.0 and left <= 1.0`
+ # ENSURE: `self.length <= length implies result.length == length`
+ # ENSURE: `self.length >= length implies result == self`
+ fun justify(length: Int, left: Float): String
+ do
+ var diff = length - self.length
+ if diff <= 0 then return to_s
+ assert left >= 0.0 and left <= 1.0
+ var before = (diff.to_f * left).to_i
+ return " " * before + self + " " * (diff-before)
+ end
+
+ # Mangle a string to be a unique string only made of alphanumeric characters and underscores.
+ #
+ # This method is injective (two different inputs never produce the same
+ # output) and the returned string always respect the following rules:
+ #
+ # * Contains only US-ASCII letters, digits and underscores.
+ # * Never starts with a digit.
+ # * Never ends with an underscore.
+ # * Never contains two contiguous underscores.
+ #
+ # assert "42_is/The answer!".to_cmangle == "_52d2_is_47dThe_32danswer_33d"
+ # assert "__".to_cmangle == "_95d_95d"
+ # assert "__d".to_cmangle == "_95d_d"
+ # assert "_d_".to_cmangle == "_d_95d"
+ # assert "_42".to_cmangle == "_95d42"
+ # assert "foo".to_cmangle == "foo"
+ # assert "".to_cmangle == ""
fun to_cmangle: String
do
+ if is_empty then return ""
var res = new FlatBuffer
var underscore = false
- for c in self.chars do
+ var start = 0
+ var c = chars[0]
+
+ if c >= '0' and c <= '9' then
+ res.add('_')
+ res.append(c.ascii.to_s)
+ res.add('d')
+ start = 1
+ end
+ for i in [start..length[ do
+ c = chars[i]
if (c >= 'a' and c <= 'z') or (c >='A' and c <= 'Z') then
res.add(c)
underscore = false
underscore = false
end
end
+ if underscore then
+ res.append('_'.ascii.to_s)
+ res.add('d')
+ end
return res.to_s
end
#
# assert "abAB12<>&".escape_to_c == "abAB12<>&"
# assert "\n\"'\\".escape_to_c == "\\n\\\"\\'\\\\"
+ #
+ # Most non-printable characters (bellow ASCII 32) are escaped to an octal form `\nnn`.
+ # Three digits are always used to avoid following digits to be interpreted as an element
+ # of the octal sequence.
+ #
+ # assert "{0.ascii}{1.ascii}{8.ascii}{31.ascii}{32.ascii}".escape_to_c == "\\000\\001\\010\\037 "
+ #
+ # The exceptions are the common `\t` and `\n`.
fun escape_to_c: String
do
var b = new FlatBuffer
- for c in self.chars do
+ for i in [0..length[ do
+ var c = chars[i]
if c == '\n' then
b.append("\\n")
+ else if c == '\t' then
+ b.append("\\t")
else if c == '\0' then
- b.append("\\0")
+ b.append("\\000")
else if c == '"' then
b.append("\\\"")
else if c == '\'' then
else if c == '\\' then
b.append("\\\\")
else if c.ascii < 32 then
- b.append("\\{c.ascii.to_base(8, false)}")
+ b.add('\\')
+ var oct = c.ascii.to_base(8, false)
+ # Force 3 octal digits since it is the
+ # maximum allowed in the C specification
+ if oct.length == 1 then
+ b.add('0')
+ b.add('0')
+ else if oct.length == 2 then
+ b.add('0')
+ end
+ b.append(oct)
else
b.add(c)
end
# assert "\n\"'\\\{\}".escape_to_nit == "\\n\\\"\\'\\\\\\\{\\\}"
fun escape_to_nit: String do return escape_more_to_c("\{\}")
+ # Escape to POSIX Shell (sh).
+ #
+ # Abort if the text contains a null byte.
+ #
+ # assert "\n\"'\\\{\}0".escape_to_sh == "'\n\"'\\''\\\{\}0'"
+ fun escape_to_sh: String do
+ var b = new FlatBuffer
+ b.chars.add '\''
+ for i in [0..length[ do
+ var c = chars[i]
+ if c == '\'' then
+ b.append("'\\''")
+ else
+ assert without_null_byte: c != '\0'
+ b.add(c)
+ end
+ end
+ b.chars.add '\''
+ return b.to_s
+ end
+
+ # Escape to include in a Makefile
+ #
+ # Unfortunately, some characters are not escapable in Makefile.
+ # These characters are `;`, `|`, `\`, and the non-printable ones.
+ # They will be rendered as `"?{hex}"`.
+ fun escape_to_mk: String do
+ var b = new FlatBuffer
+ for i in [0..length[ do
+ var c = chars[i]
+ if c == '$' then
+ b.append("$$")
+ else if c == ':' or c == ' ' or c == '#' then
+ b.add('\\')
+ b.add(c)
+ else if c.ascii < 32 or c == ';' or c == '|' or c == '\\' or c == '=' then
+ b.append("?{c.ascii.to_base(16, false)}")
+ else
+ b.add(c)
+ end
+ end
+ return b.to_s
+ end
+
# Return a string where Nit escape sequences are transformed.
#
# var s = "\\n"
do
var res = new FlatBuffer.with_capacity(self.length)
var was_slash = false
- for c in chars do
+ for i in [0..length[ do
+ var c = chars[i]
if not was_slash then
if c == '\\' then
was_slash = true
return res.to_s
end
+ # Encode `self` to percent (or URL) encoding
+ #
+ # assert "aBc09-._~".to_percent_encoding == "aBc09-._~"
+ # assert "%()< >".to_percent_encoding == "%25%28%29%3c%20%3e"
+ # assert ".com/post?e=asdf&f=123".to_percent_encoding == ".com%2fpost%3fe%3dasdf%26f%3d123"
+ fun to_percent_encoding: String
+ do
+ var buf = new FlatBuffer
+
+ for i in [0..length[ do
+ var c = chars[i]
+ if (c >= '0' and c <= '9') or
+ (c >= 'a' and c <= 'z') or
+ (c >= 'A' and c <= 'Z') or
+ c == '-' or c == '.' or
+ c == '_' or c == '~'
+ then
+ buf.add c
+ else buf.append "%{c.ascii.to_hex}"
+ end
+
+ return buf.to_s
+ end
+
+ # Decode `self` from percent (or URL) encoding to a clear string
+ #
+ # Replace invalid use of '%' with '?'.
+ #
+ # assert "aBc09-._~".from_percent_encoding == "aBc09-._~"
+ # assert "%25%28%29%3c%20%3e".from_percent_encoding == "%()< >"
+ # assert ".com%2fpost%3fe%3dasdf%26f%3d123".from_percent_encoding == ".com/post?e=asdf&f=123"
+ # assert "%25%28%29%3C%20%3E".from_percent_encoding == "%()< >"
+ # assert "incomplete %".from_percent_encoding == "incomplete ?"
+ # assert "invalid % usage".from_percent_encoding == "invalid ? usage"
+ fun from_percent_encoding: String
+ do
+ var buf = new FlatBuffer
+
+ var i = 0
+ while i < length do
+ var c = chars[i]
+ if c == '%' then
+ if i + 2 >= length then
+ # What follows % has been cut off
+ buf.add '?'
+ else
+ i += 1
+ var hex_s = substring(i, 2)
+ if hex_s.is_hex then
+ var hex_i = hex_s.to_hex
+ buf.add hex_i.ascii
+ i += 1
+ else
+ # What follows a % is not Hex
+ buf.add '?'
+ i -= 1
+ end
+ end
+ else buf.add c
+
+ i += 1
+ end
+
+ return buf.to_s
+ end
+
+ # Escape the characters `<`, `>`, `&`, `"`, `'` and `/` as HTML/XML entity references.
+ #
+ # assert "a&b-<>\"x\"/'".html_escape == "a&b-<>"x"/'"
+ #
+ # SEE: <https://www.owasp.org/index.php/XSS_%28Cross_Site_Scripting%29_Prevention_Cheat_Sheet#RULE_.231_-_HTML_Escape_Before_Inserting_Untrusted_Data_into_HTML_Element_Content>
+ fun html_escape: String
+ do
+ var buf = new FlatBuffer
+
+ for i in [0..length[ do
+ var c = chars[i]
+ if c == '&' then
+ buf.append "&"
+ else if c == '<' then
+ buf.append "<"
+ else if c == '>' then
+ buf.append ">"
+ else if c == '"' then
+ buf.append """
+ else if c == '\'' then
+ buf.append "'"
+ else if c == '/' then
+ buf.append "/"
+ else buf.add c
+ end
+
+ return buf.to_s
+ end
+
# Equality of text
# Two pieces of text are equals if thez have the same characters in the same order.
#
end
end
+ # Escape string used in labels for graphviz
+ #
+ # assert ">><<".escape_to_dot == "\\>\\>\\<\\<"
+ fun escape_to_dot: String
+ do
+ return escape_more_to_c("|\{\}<>")
+ end
+
# Flat representation of self
fun flatten: FlatText is abstract
# djb2 hash algorithm
var h = 5381
- for char in self.chars do
+ for i in [0..length[ do
+ var char = chars[i]
h = h.lshift(5) + h + char.ascii
end
abstract class FlatText
super Text
- private var items: NativeString
+ # Underlying C-String (`char*`)
+ #
+ # Warning : Might be void in some subclasses, be sure to check
+ # if set before using it.
+ private var items: NativeString is noinit
# Real items, used as cache for to_cstring is called
private var real_items: nullable NativeString = null
- redef var length: Int = 0
+ # Returns a char* starting at position `index_from`
+ #
+ # WARNING: If you choose to use this service, be careful of the following.
+ #
+ # Strings and NativeString are *ideally* always allocated through a Garbage Collector.
+ # Since the GC tracks the use of the pointer for the beginning of the char*, it may be
+ # deallocated at any moment, rendering the pointer returned by this function invalid.
+ # Any access to freed memory may very likely cause undefined behaviour or a crash.
+ # (Failure to do so will most certainly result in long and painful debugging hours)
+ #
+ # The only safe use of this pointer is if it is ephemeral (e.g. read in a C function
+ # then immediately return).
+ #
+ # As always, do not modify the content of the String in C code, if this is what you want
+ # copy locally the char* as Nit Strings are immutable.
+ private fun fast_cstring: NativeString is abstract
- init do end
+ redef var length: Int = 0
redef fun output
do
type SELFTYPE: Text
- private var target: SELFTYPE
-
- private init(tgt: SELFTYPE)
- do
- target = tgt
- end
+ var target: SELFTYPE
redef fun is_empty do return target.is_empty
end
+# A `String` holds and manipulates an arbitrary sequence of characters.
+#
+# String objects may be created using literals.
+#
+# assert "Hello World!" isa String
abstract class String
super Text
- redef type SELFTYPE: String
+ redef type SELFTYPE: String is fixed
redef fun to_s do return self
# assert "abc" * 0 == ""
fun *(i: Int): SELFTYPE is abstract
+ # Insert `s` at `pos`.
+ #
+ # assert "helloworld".insert_at(" ", 5) == "hello world"
fun insert_at(s: String, pos: Int): SELFTYPE is abstract
+ redef fun substrings: Iterator[String] is abstract
+
# Returns a reversed version of self
#
# assert "hello".reversed == "olleh"
var new_str = new FlatBuffer.with_capacity(self.length)
var is_first_char = true
- for char in self.chars do
+ for i in [0..length[ do
+ var char = chars[i]
if is_first_char then
new_str.add(char.to_lower)
is_first_char = false
var is_first_char = true
var follows_us = false
- for char in self.chars do
+ for i in [0..length[ do
+ var char = chars[i]
if is_first_char then
new_str.add(char)
is_first_char = false
return new_str.to_s
end
+
+ # Returns a capitalized `self`
+ #
+ # Letters that follow a letter are lowercased
+ # Letters that follow a non-letter are upcased.
+ #
+ # SEE : `Char::is_letter` for the definition of letter.
+ #
+ # assert "jAVASCRIPT".capitalized == "Javascript"
+ # assert "i am root".capitalized == "I Am Root"
+ # assert "ab_c -ab0c ab\nc".capitalized == "Ab_C -Ab0C Ab\nC"
+ fun capitalized: SELFTYPE do
+ if length == 0 then return self
+
+ var buf = new FlatBuffer.with_capacity(length)
+
+ var curr = chars[0].to_upper
+ var prev = curr
+ buf[0] = curr
+
+ for i in [1 .. length[ do
+ prev = curr
+ curr = self[i]
+ if prev.is_letter then
+ buf[i] = curr.to_lower
+ else
+ buf[i] = curr.to_upper
+ end
+ end
+
+ return buf.to_s
+ end
end
private class FlatSubstringsIter
var tgt: nullable FlatText
- init(tgt: FlatText) do self.tgt = tgt
-
redef fun item do
assert is_ok
return tgt.as(not null)
super String
# Index in _items of the start of the string
- private var index_from: Int
+ private var index_from: Int is noinit
# Indes in _items of the last item of the string
- private var index_to: Int
+ private var index_to: Int is noinit
+
+ redef var chars: SequenceRead[Char] = new FlatStringCharView(self) is lazy
- redef var chars: SequenceRead[Char] = new FlatStringCharView(self)
+ redef fun [](index)
+ do
+ # Check that the index (+ index_from) is not larger than indexTo
+ # In other terms, if the index is valid
+ assert index >= 0
+ assert (index + index_from) <= index_to
+ return items[index + index_from]
+ end
################################################
# AbstractString specific methods #
redef fun reversed
do
- var native = calloc_string(self.length + 1)
+ var native = new NativeString(self.length + 1)
var length = self.length
var items = self.items
var pos = 0
return native.to_s_with_length(self.length)
end
+ redef fun fast_cstring do return items.fast_cstring(index_from)
+
redef fun substring(from, count)
do
assert count >= 0
from = 0
end
- var realFrom = index_from + from
+ var new_from = index_from + from
- if (realFrom + count) > index_to then return new FlatString.with_infos(items, index_to - realFrom + 1, realFrom, index_to)
+ if (new_from + count) > index_to then
+ var new_len = index_to - new_from + 1
+ if new_len <= 0 then return empty
+ return new FlatString.with_infos(items, new_len, new_from, index_to)
+ end
- if count == 0 then return empty
+ if count <= 0 then return empty
- var to = realFrom + count - 1
+ var to = new_from + count - 1
- return new FlatString.with_infos(items, to - realFrom + 1, realFrom, to)
+ return new FlatString.with_infos(items, to - new_from + 1, new_from, to)
end
redef fun empty do return "".as(FlatString)
redef fun to_upper
do
- var outstr = calloc_string(self.length + 1)
+ var outstr = new NativeString(self.length + 1)
var out_index = 0
var myitems = self.items
redef fun to_lower
do
- var outstr = calloc_string(self.length + 1)
+ var outstr = new NativeString(self.length + 1)
var out_index = 0
var myitems = self.items
# String Specific Methods #
##################################################
- private init with_infos(items: NativeString, len: Int, from: Int, to: Int)
+ # Low-level creation of a new string with given 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 with_infos(items: NativeString, length: Int, from: Int, to: Int)
do
self.items = items
- length = len
+ self.length = length
index_from = from
index_to = to
end
redef fun to_cstring: NativeString
do
- if real_items != null then return real_items.as(not null)
- if index_from > 0 or index_to != items.cstring_length - 1 then
- var newItems = calloc_string(length + 1)
+ if real_items != null then
+ return real_items.as(not null)
+ else
+ var newItems = new NativeString(length + 1)
self.items.copy_to(newItems, length, index_from, 0)
newItems[length] = '\0'
self.real_items = newItems
return newItems
end
- return items
end
redef fun ==(other)
var total_length = my_length + its_length
- var target_string = calloc_string(my_length + its_length + 1)
+ var target_string = new NativeString(my_length + its_length + 1)
self.items.copy_to(target_string, my_length, index_from, 0)
if s isa FlatString then
s.items.copy_to(target_string, its_length, 0, my_length)
else
var curr_pos = my_length
- for i in s.chars do
- target_string[curr_pos] = i
+ for i in [0..s.length[ do
+ var c = s.chars[i]
+ target_string[curr_pos] = c
curr_pos += 1
end
end
var my_items = self.items
- var target_string = calloc_string((final_length) + 1)
+ var target_string = new NativeString(final_length + 1)
target_string[final_length] = '\0'
curr_pos = pos + tgt.index_from
end
- redef fun is_ok do return curr_pos >= 0
+ redef fun is_ok do return curr_pos >= target.index_from
redef fun item do return target_items[curr_pos]
end
+# A mutable sequence of characters.
abstract class Buffer
super Text
- redef type SELFTYPE: Buffer
+ redef type SELFTYPE: Buffer is fixed
# Specific implementations MUST set this to `true` in order to invalidate caches
protected var is_dirty = true
+ # Copy-On-Write flag
+ #
+ # If the `Buffer` was to_s'd, the next in-place altering
+ # operation will cause the current `Buffer` to be re-allocated.
+ #
+ # The flag will then be set at `false`.
+ protected var written = false
+
# Modifies the char contained at pos `index`
#
# DEPRECATED : Use self.chars.[]= instead
# assert b == "hello world!"
fun lower is abstract
+ # Capitalizes each word in `self`
+ #
+ # Letters that follow a letter are lowercased
+ # Letters that follow a non-letter are upcased.
+ #
+ # SEE: `Char::is_letter` for the definition of a letter.
+ #
+ # var b = new FlatBuffer.from("jAVAsCriPt")
+ # b.capitalize
+ # assert b == "Javascript"
+ # b = new FlatBuffer.from("i am root")
+ # b.capitalize
+ # assert b == "I Am Root"
+ # b = new FlatBuffer.from("ab_c -ab0c ab\nc")
+ # b.capitalize
+ # assert b == "Ab_C -Ab0C Ab\nC"
+ fun capitalize do
+ if length == 0 then return
+ var c = self[0].to_upper
+ self[0] = c
+ var prev = c
+ for i in [1 .. length[ do
+ prev = c
+ c = self[i]
+ if prev.is_letter then
+ self[i] = c.to_lower
+ else
+ self[i] = c.to_upper
+ end
+ end
+ end
+
redef fun hash
do
if is_dirty then hash_cache = null
super FlatText
super Buffer
- redef type SELFTYPE: FlatBuffer
-
- redef var chars: Sequence[Char] = new FlatBufferCharView(self)
+ redef var chars: Sequence[Char] = new FlatBufferCharView(self) is lazy
private var capacity: Int = 0
+ redef fun fast_cstring do return items.fast_cstring(0)
+
redef fun substrings do return new FlatSubstringsIter(self)
+ # Re-copies the `NativeString` into a new one and sets it as the new `Buffer`
+ #
+ # This happens when an operation modifies the current `Buffer` and
+ # the Copy-On-Write flag `written` is set at true.
+ private fun reset do
+ var nns = new NativeString(capacity)
+ items.copy_to(nns, length, 0, 0)
+ items = nns
+ written = false
+ end
+
+ redef fun [](index)
+ do
+ assert index >= 0
+ assert index < length
+ return items[index]
+ end
+
redef fun []=(index, item)
do
is_dirty = true
add(item)
return
end
+ if written then reset
assert index >= 0 and index < length
items[index] = item
end
redef fun clear do
is_dirty = true
+ if written then reset
length = 0
end
var c = capacity
if cap <= c then return
while c <= cap do c = c * 2 + 2
- var a = calloc_string(c+1)
+ # The COW flag can be set at false here, since
+ # it does a copy of the current `Buffer`
+ written = false
+ var a = new NativeString(c+1)
if length > 0 then items.copy_to(a, length, 0, 0)
items = a
capacity = c
redef fun to_s: String
do
- return to_cstring.to_s_with_length(length)
+ written = true
+ if length == 0 then items = new NativeString(1)
+ return new FlatString.with_infos(items, length, 0, length - 1)
end
redef fun to_cstring
do
if is_dirty then
- var new_native = calloc_string(length + 1)
+ var new_native = new NativeString(length + 1)
new_native[length] = '\0'
if length > 0 then items.copy_to(new_native, length, 0, 0)
real_items = new_native
# Create a new empty string.
init do end
+ # Low-level creation a new buffer with given data.
+ #
+ # `items` will be used as is, without copy, to store the characters of the buffer.
+ # Aliasing issues is the responsibility of the caller.
+ #
+ # If `items` is shared, `written` should be set to true after the creation
+ # so that a modification will do a copy-on-write.
+ private init with_infos(items: NativeString, capacity, length: Int)
+ do
+ self.items = items
+ self.length = length
+ self.capacity = capacity
+ end
+
+ # Create a new string copied from `s`.
init from(s: Text)
do
capacity = s.length + 1
length = s.length
- items = calloc_string(capacity)
+ items = new NativeString(capacity)
if s isa FlatString then
s.items.copy_to(items, length, s.index_from, 0)
else if s isa FlatBuffer then
s.items.copy_to(items, length, 0, 0)
else
var curr_pos = 0
- for i in s.chars do
- items[curr_pos] = i
+ for i in [0..s.length[ do
+ var c = s.chars[i]
+ items[curr_pos] = c
curr_pos += 1
end
end
init with_capacity(cap: Int)
do
assert cap >= 0
- # _items = new NativeString.calloc(cap)
- items = calloc_string(cap+1)
+ items = new NativeString(cap+1)
capacity = cap
length = 0
end
s.items.copy_to(items, sl, 0, length)
else
var curr_pos = self.length
- for i in s.chars do
- items[curr_pos] = i
+ for i in [0..s.length[ do
+ var c = s.chars[i]
+ items[curr_pos] = c
curr_pos += 1
end
end
if from < 0 then from = 0
if count > length then count = length
if from < count then
- var r = new FlatBuffer.with_capacity(count - from)
- while from < count do
- r.chars.push(items[from])
- from += 1
- end
+ var len = count - from
+ var r_items = new NativeString(len)
+ items.copy_to(r_items, len, from, 0)
+ var r = new FlatBuffer.with_infos(r_items, len, len)
return r
else
return new FlatBuffer
redef fun reverse
do
- var ns = calloc_string(capacity)
+ written = false
+ var ns = new NativeString(capacity)
var si = length - 1
var ni = 0
var it = items
redef fun upper
do
+ if written then reset
var it = items
var id = length - 1
while id >= 0 do
redef fun lower
do
+ if written then reset
var it = items
var id = length - 1
while id >= 0 do
private class FlatBufferCharView
super BufferCharView
- super StringCapable
redef type SELFTYPE: FlatBuffer
redef fun append(s)
do
- var my_items = target.items
var s_length = s.length
if target.capacity < s.length then enlarge(s_length + target.length)
end
# The class name of the object.
#
- # assert 5.class_name == "Int"
+ # assert 5.class_name == "Int"
fun class_name: String do return native_class_name.to_s
# Developer readable representation of `self`.
redef class Int
# Wrapper of strerror C function
- private fun strerror_ext: NativeString is extern `{
- return strerror(recv);
- `}
+ private fun strerror_ext: NativeString is extern "strerror"
# Returns a string describing error number
fun strerror: String do return strerror_ext.to_s
end
end
+ # C function to calculate the length of the `NativeString` to receive `self`
+ private fun int_to_s_len: Int is extern "native_int_length_str"
+
# C function to convert an nit Int to a NativeString (char*)
- private fun native_int_to_s: NativeString is extern "native_int_to_s"
+ private fun native_int_to_s(nstr: NativeString, strlen: Int) is extern "native_int_to_s"
# return displayable int in base 10 and signed
#
# assert 1.to_s == "1"
# assert (-123).to_s == "-123"
redef fun to_s do
- return native_int_to_s.to_s
+ # Fast case for common numbers
+ if self == 0 then return "0"
+ if self == 1 then return "1"
+
+ var nslen = int_to_s_len
+ var ns = new NativeString(nslen + 1)
+ ns[nslen] = '\0'
+ native_int_to_s(ns, nslen + 1)
+ return ns.to_s_with_length(nslen)
end
# return displayable int in hexadecimal
end
redef class Float
- # Pretty print self, print needoed decimals up to a max of 3.
+ # Pretty representation of `self`, with decimals as needed from 1 to a maximum of 3
#
- # assert 12.34.to_s == "12.34"
- # assert (-0120.03450).to_s == "-120.035"
+ # assert 12.34.to_s == "12.34"
+ # assert (-0120.030).to_s == "-120.03"
#
- # see `to_precision` for a different precision.
+ # see `to_precision` for a custom precision.
redef fun to_s do
var str = to_precision( 3 )
if is_inf != 0 or is_nan then return str
return str
end
- # `self` representation with `nb` digits after the '.'.
+ # `String` representation of `self` with the given number of `decimals`
#
- # assert 12.345.to_precision(1) == "12.3"
- # assert 12.345.to_precision(2) == "12.35"
- # assert 12.345.to_precision(3) == "12.345"
- # assert 12.345.to_precision(4) == "12.3450"
- fun to_precision(nb: Int): String
+ # assert 12.345.to_precision(0) == "12"
+ # assert 12.345.to_precision(3) == "12.345"
+ # assert (-12.345).to_precision(3) == "-12.345"
+ # assert (-0.123).to_precision(3) == "-0.123"
+ # assert 0.999.to_precision(2) == "1.00"
+ # assert 0.999.to_precision(4) == "0.9990"
+ fun to_precision(decimals: Int): String
do
if is_nan then return "nan"
return "-inf"
end
- if nb == 0 then return self.to_i.to_s
+ if decimals == 0 then return self.to_i.to_s
var f = self
- for i in [0..nb[ do f = f * 10.0
+ for i in [0..decimals[ do f = f * 10.0
if self > 0.0 then
f = f + 0.5
else
f = f - 0.5
end
var i = f.to_i
- if i == 0 then return "0.0"
- var s = i.to_s
+ if i == 0 then return "0." + "0"*decimals
+
+ # Prepare both parts of the float, before and after the "."
+ var s = i.abs.to_s
var sl = s.length
- if sl > nb then
- var p1 = s.substring(0, s.length-nb)
- var p2 = s.substring(s.length-nb, nb)
- return p1 + "." + p2
+ var p1
+ var p2
+ if sl > decimals then
+ # Has something before the "."
+ p1 = s.substring(0, sl-decimals)
+ p2 = s.substring(sl-decimals, decimals)
else
- return "0." + ("0"*(nb-sl)) + s
+ p1 = "0"
+ p2 = "0"*(decimals-sl) + s
end
- end
-
- # `self` representation with `nb` digits after the '.'.
- #
- # assert 12.345.to_precision_native(1) == "12.3"
- # assert 12.345.to_precision_native(2) == "12.35"
- # assert 12.345.to_precision_native(3) == "12.345"
- # assert 12.345.to_precision_native(4) == "12.3450"
- fun to_precision_native(nb: Int): String import NativeString.to_s `{
- int size;
- char *str;
- size = snprintf(NULL, 0, "%.*f", (int)nb, recv);
- str = malloc(size + 1);
- sprintf(str, "%.*f", (int)nb, recv );
+ if i < 0 then p1 = "-" + p1
- return NativeString_to_s( str );
- `}
+ return p1 + "." + p2
+ end
end
redef class Char
# Returns true if the char is a numerical digit
#
- # assert '0'.is_numeric
- # assert '9'.is_numeric
- # assert not 'a'.is_numeric
- # assert not '?'.is_numeric
+ # assert '0'.is_numeric
+ # assert '9'.is_numeric
+ # assert not 'a'.is_numeric
+ # assert not '?'.is_numeric
fun is_numeric: Bool
do
return self >= '0' and self <= '9'
# Returns true if the char is an alpha digit
#
- # assert 'a'.is_alpha
- # assert 'Z'.is_alpha
- # assert not '0'.is_alpha
- # assert not '?'.is_alpha
+ # assert 'a'.is_alpha
+ # assert 'Z'.is_alpha
+ # assert not '0'.is_alpha
+ # assert not '?'.is_alpha
fun is_alpha: Bool
do
return (self >= 'a' and self <= 'z') or (self >= 'A' and self <= 'Z')
# Returns true if the char is an alpha or a numeric digit
#
- # assert 'a'.is_alphanumeric
- # assert 'Z'.is_alphanumeric
- # assert '0'.is_alphanumeric
- # assert '9'.is_alphanumeric
- # assert not '?'.is_alphanumeric
+ # assert 'a'.is_alphanumeric
+ # assert 'Z'.is_alphanumeric
+ # assert '0'.is_alphanumeric
+ # assert '9'.is_alphanumeric
+ # assert not '?'.is_alphanumeric
fun is_alphanumeric: Bool
do
return self.is_numeric or self.is_alpha
end
redef class Array[E]
+
# Fast implementation
redef fun to_s
do
- var s = new FlatBuffer
+ 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 na = new NativeArray[String](l)
var i = 0
+ var sl = 0
+ var mypos = 0
+ while i < l do
+ var itsi = its[i]
+ if itsi == null then
+ i += 1
+ continue
+ end
+ var tmp = itsi.to_s
+ sl += tmp.length
+ na[mypos] = tmp
+ i += 1
+ mypos += 1
+ end
+ var ns = new NativeString(sl + 1)
+ ns[sl] = '\0'
+ i = 0
+ var off = 0
+ while i < mypos do
+ var tmp = na[i]
+ var tpl = tmp.length
+ if tmp isa FlatString then
+ tmp.items.copy_to(ns, tpl, tmp.index_from, off)
+ off += tpl
+ else
+ for j in tmp.substrings do
+ var s = j.as(FlatString)
+ var slen = s.length
+ s.items.copy_to(ns, slen, s.index_from, off)
+ off += slen
+ end
+ end
+ i += 1
+ end
+ return ns.to_s_with_length(sl)
+ end
+end
+
+redef class NativeArray[E]
+ # Join all the elements using `to_s`
+ #
+ # REQUIRE: `self isa NativeArray[String]`
+ # REQUIRE: all elements are initialized
+ fun native_to_s: String
+ do
+ assert self isa NativeArray[String]
var l = length
+ var na = self
+ var i = 0
+ var sl = 0
+ var mypos = 0
while i < l do
- var e = self[i]
- if e != null then s.append(e.to_s)
+ sl += na[i].length
i += 1
+ mypos += 1
end
- return s.to_s
+ var ns = new NativeString(sl + 1)
+ ns[sl] = '\0'
+ i = 0
+ var off = 0
+ while i < mypos do
+ var tmp = na[i]
+ var tpl = tmp.length
+ if tmp isa FlatString then
+ tmp.items.copy_to(ns, tpl, tmp.index_from, off)
+ off += tpl
+ else
+ for j in tmp.substrings do
+ var s = j.as(FlatString)
+ var slen = s.length
+ s.items.copy_to(ns, slen, s.index_from, off)
+ off += slen
+ end
+ end
+ i += 1
+ end
+ return ns.to_s_with_length(sl)
end
end
var i = iterator
var k = i.key
var e = i.item
- s.append("{k}{couple_sep}{e or else "<null>"}")
+ s.append("{k or else "<null>"}{couple_sep}{e or else "<null>"}")
# Concat other items
i.next
s.append(sep)
k = i.key
e = i.item
- s.append("{k}{couple_sep}{e or else "<null>"}")
+ s.append("{k or else "<null>"}{couple_sep}{e or else "<null>"}")
i.next
end
return s.to_s
# Native strings are simple C char *
extern class NativeString `{ char* `}
- super StringCapable
+ # Creates a new NativeString with a capacity of `length`
+ new(length: Int) is intern
+ # Returns a char* starting at `index`.
+ #
+ # WARNING: Unsafe for extern code, use only for temporary
+ # pointer manipulation purposes (e.g. write to file or such)
+ fun fast_cstring(index: Int): NativeString is intern
+
+ # Get char at `index`.
fun [](index: Int): Char is intern
+
+ # Set char `item` at index.
fun []=(index: Int, item: Char) is intern
+
+ # Copy `self` to `dest`.
fun copy_to(dest: NativeString, length: Int, from: Int, to: Int) is intern
# Position of the first nul character.
while self[l] != '\0' do l += 1
return l
end
+
+ # Parse `self` as an Int.
fun atoi: Int is intern
+
+ # Parse `self` as a Float.
fun atof: Float is extern "atof"
redef fun to_s
return to_s_with_length(cstring_length)
end
+ # Returns `self` as a String of `length`.
fun to_s_with_length(length: Int): FlatString
do
assert length >= 0
- return new FlatString.with_infos(self, length, 0, length - 1)
+ var str = new FlatString.with_infos(self, length, 0, length - 1)
+ return str
end
+ # Returns `self` as a new String.
fun to_s_with_copy: FlatString
do
var length = cstring_length
- var new_self = calloc_string(length + 1)
+ var new_self = new NativeString(length + 1)
copy_to(new_self, length, 0, 0)
- return new FlatString.with_infos(new_self, length, 0, length - 1)
+ var str = new FlatString.with_infos(new_self, length, 0, length - 1)
+ new_self[length] = '\0'
+ str.real_items = new_self
+ return str
end
end
-# StringCapable objects can create native strings
-interface StringCapable
- protected fun calloc_string(size: Int): NativeString is intern
-end
-
redef class Sys
- var _args_cache: nullable Sequence[String]
+ private var args_cache: nullable Sequence[String]
# The arguments of the program as given by the OS
fun program_args: Sequence[String]
#
# Note: it caching is not usefull, see `alpha_comparator`
class CachedAlphaComparator
- super Comparator[Object]
+ super Comparator
+ redef type COMPARED: Object
private var cache = new HashMap[Object, String]
# see `alpha_comparator`
private class AlphaComparator
- super Comparator[Object]
+ super Comparator
redef fun compare(a, b) do return a.to_s <=> b.to_s
end
# var a = [1, 2, 3, 10, 20]
# alpha_comparator.sort(a)
# assert a == [1, 10, 2, 20, 3]
-fun alpha_comparator: Comparator[Object] do return once new AlphaComparator
+fun alpha_comparator: Comparator do return once new AlphaComparator
# The arguments of the program as given by the OS
fun args: Sequence[String]
nitlanguage / nit.git / blobdiff