X-Git-Url: http://nitlanguage.org

diff --git a/lib/core/bytes.nit b/lib/core/bytes.nit
index 439a621..5b5fceb 100644
--- a/lib/core/bytes.nit
+++ b/lib/core/bytes.nit
@@ -53,7 +53,7 @@ redef class Byte
 	super BytePattern
 
 	# Write self as a string into `ns` at position `pos`
-	private fun add_digest_at(ns: NativeString, pos: Int) do
+	private fun add_digest_at(ns: CString, pos: Int) do
 		var tmp = (0xF0u8 & self) >> 4
 		ns[pos] = if tmp >= 0x0Au8 then tmp + 0x37u8 else tmp + 0x30u8
 		tmp = 0x0Fu8 & self
@@ -146,8 +146,8 @@ class Bytes
 	super AbstractArray[Byte]
 	super BytePattern
 
-	# A NativeString being a char*, it can be used as underlying representation here.
-	var items: NativeString
+	# A CString being a char*, it can be used as underlying representation here.
+	var items: CString
 
 	# Number of bytes in the array
 	redef var length
@@ -163,13 +163,13 @@ class Bytes
 	#     var b = new Bytes.empty
 	#     assert b.to_s == ""
 	init empty do
-		var ns = new NativeString(0)
+		var ns = new CString(0)
 		init(ns, 0, 0)
 	end
 
 	# Init a `Bytes` with capacity `cap`
 	init with_capacity(cap: Int) do
-		var ns = new NativeString(cap)
+		var ns = new CString(cap)
 		init(ns, 0, cap)
 	end
 
@@ -251,10 +251,18 @@ class Bytes
 		return slice(from, length)
 	end
 
-	# Returns self as a hexadecimal digest
+	# Returns self as an hexadecimal digest.
+	#
+	# Also known as plain hexdump or postscript hexdump.
+	#
+	# ~~~
+	# var b = "abcd".to_bytes
+	# assert b.hexdigest == "61626364"
+	# assert b.hexdigest.hexdigest_to_bytes == b
+	# ~~~
 	fun hexdigest: String do
 		var elen = length * 2
-		var ns = new NativeString(elen)
+		var ns = new CString(elen)
 		var i = 0
 		var oi = 0
 		while i < length do
@@ -262,7 +270,101 @@ class Bytes
 			i += 1
 			oi += 2
 		end
-		return new FlatString.full(ns, elen, 0, elen - 1, elen)
+		return new FlatString.full(ns, elen, 0, elen)
+	end
+
+	# Return self as a C hexadecimal digest where bytes are prefixed by `\x`
+	#
+	# The output is compatible with literal stream of bytes for most languages
+	# including C and Nit.
+	#
+	# ~~~
+	# var b = "abcd".to_bytes
+	# assert b.chexdigest == "\\x61\\x62\\x63\\x64"
+	# assert b.chexdigest.unescape_to_bytes == b
+	# ~~~
+	fun chexdigest: String do
+		var elen = length * 4
+		var ns = new CString(elen)
+		var i = 0
+		var oi = 0
+		while i < length do
+			ns[oi] = 0x5Cu8 # b'\\'
+			ns[oi+1] = 0x78u8 # b'x'
+			self[i].add_digest_at(ns, oi+2)
+			i += 1
+			oi += 4
+		end
+		return new FlatString.full(ns, elen, 0, elen)
+	end
+
+
+	# Returns self as a stream of bits (0 and 1)
+	#
+	# ~~~
+	# var b = "abcd".to_bytes
+	# assert b.binarydigest == "01100001011000100110001101100100"
+	# assert b.binarydigest.binarydigest_to_bytes == b
+	# ~~~
+	fun binarydigest: String do
+		var elen = length * 8
+		var ns = new CString(elen)
+		var i = 0
+		var oi = 0
+		while i < length do
+			var c = self[i]
+			var b = 128u8
+			while b > 0u8 do
+				if c & b == 0u8 then
+					ns[oi] = 0x30u8 # b'0'
+				else
+					ns[oi] = 0x31u8 # b'1'
+				end
+				oi += 1
+				b = b >> 1
+			end
+			i += 1
+		end
+		return new FlatString.full(ns, elen, 0, elen)
+	end
+
+	# Interprets `self` as a big-endian positive integer.
+	#
+	# ~~~
+	# var b = "0102".hexdigest_to_bytes
+	# assert b.to_i == 258
+	# ~~~
+	#
+	# Nul bytes on the left are trimmed.
+	# 0 is returned for an empty Bytes object.
+	#
+	# ~~~
+	# assert "01".hexdigest_to_bytes.to_i == 1
+	# assert "0001".hexdigest_to_bytes.to_i == 1
+	#
+	# assert "0000".hexdigest_to_bytes.to_i == 0
+	# assert "00".hexdigest_to_bytes.to_i == 0
+	# assert "".hexdigest_to_bytes.to_i == 0
+	# ~~~
+	#
+	# `Int::to_bytes` is a loosely reverse method.
+	#
+	# ~~~
+	# assert b.to_i.to_bytes == b
+	# assert (b.to_i + 1).to_bytes.hexdigest == "0103"
+	# assert "0001".hexdigest_to_bytes.to_i.to_bytes.hexdigest == "01"
+	# ~~~
+	#
+	# Warning: `Int` might overflow for bytes with more than 60 bits.
+	fun to_i: Int do
+		var res = 0
+		var i = 0
+		while i < length do
+			res *= 256
+			res += self[i].to_i
+			i += 1
+		end
+		return res
 	end
 
 	#     var b = new Bytes.with_capacity(1)
@@ -288,6 +390,21 @@ class Bytes
 		length += 1
 	end
 
+	# Adds the UTF-8 representation of `c` to `self`
+	#
+	#     var b = new Bytes.empty
+	#     b.add_char('A')
+	#     b.add_char('キ')
+	#     assert b.hexdigest == "41E382AD"
+	fun add_char(c: Char) do
+		if persisted then regen
+		var cln = c.u8char_len
+		var ln = length
+		enlarge(ln + cln)
+		items.set_char_at(length, c)
+		length += cln
+	end
+
 	#     var b = new Bytes.empty
 	#     b.append([104u8, 101u8, 108u8, 108u8, 111u8])
 	#     assert b.to_s == "hello"
@@ -313,13 +430,13 @@ class Bytes
 
 	# Regenerates the buffer, necessary when it was persisted
 	private fun regen do
-		var nns = new NativeString(capacity)
+		var nns = new CString(capacity)
 		items.copy_to(nns, length, 0, 0)
 		persisted = false
 	end
 
 	# Appends the `ln` first bytes of `ns` to self
-	fun append_ns(ns: NativeString, ln: Int) do
+	fun append_ns(ns: CString, ln: Int) do
 		if persisted then regen
 		var nlen = length + ln
 		if nlen > capacity then enlarge(nlen)
@@ -328,7 +445,7 @@ class Bytes
 	end
 
 	# Appends `ln` bytes from `ns` starting at index `from` to self
-	fun append_ns_from(ns: NativeString, ln, from: Int) do
+	fun append_ns_from(ns: CString, ln, from: Int) do
 		if persisted then regen
 		var nlen = length + ln
 		if nlen > capacity then enlarge(nlen)
@@ -339,7 +456,7 @@ class Bytes
 	# Appends the bytes of `s` to `selftextextt`
 	fun append_text(s: Text) do
 		for i in s.substrings do
-			append_ns(i.fast_cstring, i.bytelen)
+			append_ns(i.fast_cstring, i.byte_length)
 		end
 	end
 
@@ -349,7 +466,7 @@ class Bytes
 		if capacity >= sz then return
 		persisted = false
 		while capacity < sz do capacity = capacity * 2 + 2
-		var ns = new NativeString(capacity)
+		var ns = new CString(capacity)
 		items.copy_to(ns, length, 0, 0)
 		items = ns
 	end
@@ -357,7 +474,7 @@ class Bytes
 	redef fun to_s do
 		persisted = true
 		var b = self
-		var r = b.items.to_s_with_length(length)
+		var r = b.items.to_s_unsafe(length, copy=false)
 		if r != items then persisted = false
 		return r
 	end
@@ -455,15 +572,16 @@ class Bytes
 
 	# Decode `self` from percent (or URL) encoding to a clear string
 	#
-	# Replace invalid use of '%' with '?'.
+	# Invalid '%' are not decoded.
 	#
 	#     assert "aBc09-._~".to_bytes.from_percent_encoding == "aBc09-._~".to_bytes
 	#     assert "%25%28%29%3c%20%3e".to_bytes.from_percent_encoding == "%()< >".to_bytes
 	#     assert ".com%2fpost%3fe%3dasdf%26f%3d123".to_bytes.from_percent_encoding == ".com/post?e=asdf&f=123".to_bytes
 	#     assert "%25%28%29%3C%20%3E".to_bytes.from_percent_encoding == "%()< >".to_bytes
-	#     assert "incomplete %".to_bytes.from_percent_encoding == "incomplete ?".to_bytes
-	#     assert "invalid % usage".to_bytes.from_percent_encoding == "invalid ? usage".to_bytes
+	#     assert "incomplete %".to_bytes.from_percent_encoding == "incomplete %".to_bytes
+	#     assert "invalid % usage".to_bytes.from_percent_encoding == "invalid % usage".to_bytes
 	#     assert "%c3%a9%e3%81%82%e3%81%84%e3%81%86".to_bytes.from_percent_encoding == "éあいう".to_bytes
+	#     assert "%1 %A %C3%A9A9".to_bytes.from_percent_encoding == "%1 %A éA9".to_bytes
 	fun from_percent_encoding: Bytes do
 		var tmp = new Bytes.with_capacity(length)
 		var pos = 0
@@ -475,14 +593,14 @@ class Bytes
 				continue
 			end
 			if length - pos < 2 then
-				tmp.add '?'.ascii
+				tmp.add '%'.ascii
 				pos += 1
 				continue
 			end
 			var bn = self[pos + 1]
 			var bnn = self[pos + 2]
 			if not bn.is_valid_hexdigit or not bnn.is_valid_hexdigit then
-				tmp.add '?'.ascii
+				tmp.add '%'.ascii
 				pos += 1
 				continue
 			end
@@ -520,7 +638,7 @@ end
 private class BytesIterator
 	super IndexedIterator[Byte]
 
-	var tgt: NativeString
+	var tgt: CString
 
 	redef var index
 
@@ -535,6 +653,58 @@ private class BytesIterator
 	redef fun item do return tgt[index]
 end
 
+redef class Int
+	# A big-endian representation of self.
+	#
+	# ~~~
+	# assert     1.to_bytes.hexdigest ==     "01"
+	# assert   255.to_bytes.hexdigest ==     "FF"
+	# assert   256.to_bytes.hexdigest ==   "0100"
+	# assert 65535.to_bytes.hexdigest ==   "FFFF"
+	# assert 65536.to_bytes.hexdigest == "010000"
+	# ~~~
+	#
+	# For 0, a Bytes object with single nul byte is returned (instead of an empty Bytes object).
+	#
+	# ~~~
+	# assert 0.to_bytes.hexdigest == "00"
+	# ~~~
+	#
+	# `Bytes::to_i` can be used to do the reverse operation.
+	#
+	# ~~~
+	# assert 1234.to_bytes.to_i == 1234
+	# ~~~
+	#
+	# Require self >= 0
+	fun to_bytes: Bytes do
+		if self == 0 then return "\0".to_bytes
+		assert self > 0
+
+		# Compute the len (log256)
+		var len = 1
+		var max = 256
+		while self >= max do
+			len += 1
+			max *= 256
+		end
+
+		# Allocate the buffer
+		var res = new Bytes.with_capacity(len)
+		for i in [0..len[ do res[i] = 0u8
+
+		# Fill it starting with the end
+		var i = len
+		var sum = self
+		while i > 0 do
+			i -= 1
+			res[i] = (sum % 256).to_b
+			sum /= 256
+		end
+		return res
+	end
+end
+
 redef class Text
 	# Returns a mutable copy of `self`'s bytes
 	#
@@ -543,7 +713,7 @@ redef class Text
 	# assert "String".to_bytes == [83u8, 116u8, 114u8, 105u8, 110u8, 103u8]
 	# ~~~
 	fun to_bytes: Bytes do
-		var b = new Bytes.with_capacity(bytelen)
+		var b = new Bytes.with_capacity(byte_length)
 		append_to_bytes b
 		return b
 	end
@@ -561,24 +731,63 @@ redef class Text
 	fun append_to_bytes(b: Bytes) do
 		for s in substrings do
 			var from = if s isa FlatString then s.first_byte else 0
-			b.append_ns_from(s.items, s.bytelen, from)
+			b.append_ns_from(s.items, s.byte_length, from)
 		end
 	end
 
 	# Returns a new `Bytes` instance with the digest as content
 	#
 	#     assert "0B1F4D".hexdigest_to_bytes == [0x0Bu8, 0x1Fu8, 0x4Du8]
+	#     assert "0B1F4D".hexdigest_to_bytes.hexdigest == "0B1F4D"
+	#
+	# Characters that are not hexadecimal digits are ignored.
+	#
+	#     assert "z0B1 F4\nD".hexdigest_to_bytes.hexdigest == "0B1F4D"
+	#     assert "\\x0b1 \\xf4d".hexdigest_to_bytes.hexdigest == "0B1F4D"
 	#
-	# REQUIRE: `self` is a valid hexdigest and hexdigest.length % 2 == 0
+	# When the number of hexadecimal digit is not even, then a leading 0 is
+	# implicitly considered to fill the left byte (the most significant one).
+	#
+	#     assert "1".hexdigest_to_bytes.hexdigest == "01"
+	#     assert "FFF".hexdigest_to_bytes.hexdigest == "0FFF"
+	#
+	# `Bytes::hexdigest` is a loosely reverse method since its
+	# results contain only pairs of uppercase hexadecimal digits.
+	#
+	#     assert "ABCD".hexdigest_to_bytes.hexdigest == "ABCD"
+	#     assert "a b c".hexdigest_to_bytes.hexdigest == "0ABC"
 	fun hexdigest_to_bytes: Bytes do
 		var b = bytes
+		var max = byte_length
+
+		var dlength = 0 # Number of hex digits
 		var pos = 0
-		var max = bytelen
-		var ret = new Bytes.with_capacity(max / 2)
 		while pos < max do
-			ret.add((b[pos].hexdigit_to_byteval << 4) |
-			b[pos + 1].hexdigit_to_byteval)
-			pos += 2
+			var c = b[pos]
+			if c.is_valid_hexdigit then dlength += 1
+			pos += 1
+		end
+
+		# Allocate the result buffer
+		var ret = new Bytes.with_capacity((dlength+1) / 2)
+
+		var i = (dlength+1) % 2 # current hex digit (1=high, 0=low)
+		var byte = 0u8 # current accumulated byte value
+
+		pos = 0
+		while pos < max do
+			var c = b[pos]
+			if c.is_valid_hexdigit then
+				byte = byte << 4 | c.hexdigit_to_byteval
+				i -= 1
+				if i < 0 then
+					# Last digit known: store and restart
+					ret.add byte
+					i = 1
+					byte = 0u8
+				end
+			end
+			pos += 1
 		end
 		return ret
 	end
@@ -587,30 +796,163 @@ redef class Text
 	#
 	#     assert "&lt;STRING&#47;&rt;".hexdigest == "266C743B535452494E47262334373B2672743B"
 	fun hexdigest: String do
-		var ln = bytelen
-		var outns = new NativeString(ln * 2)
+		var ln = byte_length
+		var outns = new CString(ln * 2)
 		var oi = 0
 		for i in [0 .. ln[ do
 			bytes[i].add_digest_at(outns, oi)
 			oi += 2
 		end
-		return new FlatString.with_infos(outns, ln * 2, 0, ln * 2 - 1)
+		return new FlatString.with_infos(outns, ln * 2, 0)
+	end
+
+	# Return a `Bytes` instance where Nit escape sequences are transformed.
+	#
+	#     assert "B\\n\\x41\\u0103D3".unescape_to_bytes.hexdigest == "420A41F0908F93"
+	#
+	# `Bytes::chexdigest` is a loosely reverse methods since its result is only made
+	# of `"\x??"` escape sequences.
+	#
+	#     assert "\\x41\\x42\\x43".unescape_to_bytes.chexdigest == "\\x41\\x42\\x43"
+	#     assert "B\\n\\x41\\u0103D3".unescape_to_bytes.chexdigest == "\\x42\\x0A\\x41\\xF0\\x90\\x8F\\x93"
+	fun unescape_to_bytes: Bytes do
+		var res = new Bytes.with_capacity(self.byte_length)
+		var was_slash = false
+		var i = 0
+		while i < length do
+			var c = self[i]
+			if not was_slash then
+				if c == '\\' then
+					was_slash = true
+				else
+					res.add_char(c)
+				end
+				i += 1
+				continue
+			end
+			was_slash = false
+			if c == 'n' then
+				res.add_char('\n')
+			else if c == 'r' then
+				res.add_char('\r')
+			else if c == 't' then
+				res.add_char('\t')
+			else if c == '0' then
+				res.add_char('\0')
+			else if c == 'x' or c == 'X' then
+				var hx = substring(i + 1, 2)
+				if hx.is_hex then
+					res.add(hx.to_hex.to_b)
+				else
+					res.add_char(c)
+				end
+				i += 2
+			else if c == 'u' or c == 'U' then
+				var hx = substring(i + 1, 6)
+				if hx.is_hex then
+					res.add_char(hx.to_hex.code_point)
+				else
+					res.add_char(c)
+				end
+				i += 6
+			else
+				res.add_char(c)
+			end
+			i += 1
+		end
+		return res
+	end
+
+	# Return a `Bytes` by reading 0 and 1.
+	#
+	#     assert "1010101100001101".binarydigest_to_bytes.hexdigest == "AB0D"
+	#
+	# Note that characters that are neither 0 or 1 are just ignored.
+	#
+	#     assert "a1B01 010\n1100あ001101".binarydigest_to_bytes.hexdigest == "AB0D"
+	#     assert "hello".binarydigest_to_bytes.is_empty
+	#
+	# When the number of bits is not divisible by 8, then leading 0 are
+	# implicitly considered to fill the left byte (the most significant one).
+	#
+	#     assert "1".binarydigest_to_bytes.hexdigest == "01"
+	#     assert "1111111".binarydigest_to_bytes.hexdigest == "7F"
+	#     assert "1000110100".binarydigest_to_bytes.hexdigest == "0234"
+	#
+	# `Bytes::binarydigest` is a loosely reverse method since its
+	# results contain only 1 and 0 by blocks of 8.
+	#
+	#     assert "1010101100001101".binarydigest_to_bytes.binarydigest == "1010101100001101"
+	#     assert "1".binarydigest_to_bytes.binarydigest == "00000001"
+	fun binarydigest_to_bytes: Bytes
+	do
+		var b = bytes
+		var max = byte_length
+
+		# Count bits
+		var bitlen = 0
+		var pos = 0
+		while pos < max do
+			var c = b[pos]
+			pos += 1
+			if c == 0x30u8 or c == 0x31u8 then bitlen += 1 # b'0' or b'1'
+		end
+
+		# Allocate (and take care of the padding)
+		var ret = new Bytes.with_capacity((bitlen+7) / 8)
+
+		var i = (bitlen+7) % 8 # current bit (7th=128, 0th=1)
+		var byte = 0u8 # current accumulated byte value
+
+		pos = 0
+		while pos < max do
+			var c = b[pos]
+			pos += 1
+			if c == 0x30u8 then # b'0'
+				byte = byte << 1
+			else if c == 0x31u8 then # b'1'
+				byte = byte << 1 | 1u8
+			else
+				continue
+			end
+
+			i -= 1
+			if i < 0 then
+				# Last bit known: store and restart
+				ret.add byte
+				i = 7
+				byte = 0u8
+			end
+		end
+		return ret
 	end
 end
 
 redef class FlatText
 	redef fun append_to_bytes(b) do
 		var from = if self isa FlatString then first_byte else 0
-		b.append_ns_from(items, bytelen, from)
+		b.append_ns_from(items, byte_length, from)
 	end
 end
 
-redef class NativeString
-	# Creates a new `Bytes` object from `self` with `strlen` as length
-	fun to_bytes: Bytes do
-		var len = cstring_length
+redef class CString
+	# Creates a new `Bytes` object from `self` with `len` as length
+	#
+	# If `len` is null, strlen will determine the length of the Bytes
+	fun to_bytes(len: nullable Int): Bytes do
+		if len == null then len = cstring_length
 		return new Bytes(self, len, len)
 	end
+
+	# Creates a new `Bytes` object from a copy of `self` with `len` as length
+	#
+	# If `len` is null, strlen will determine the length of the Bytes
+	fun to_bytes_with_copy(len: nullable Int): Bytes do
+		if len == null then len = cstring_length
+		var nns = new CString(len)
+		copy_to(nns, len, 0, 0)
+		return new Bytes(nns, len, len)
+	end
 end
 
 # Joins an array of bytes `arr` separated by `sep`