X-Git-Url: http://nitlanguage.org diff --git a/lib/standard/string.nit b/lib/standard/string.nit index 07a7e1f..0ffb303 100644 --- a/lib/standard/string.nit +++ b/lib/standard/string.nit @@ -16,6 +16,7 @@ module string import math import collection +intrude import collection::array `{ #include @@ -156,13 +157,6 @@ abstract class 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 # @@ -178,7 +172,7 @@ abstract class Text # 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 @@ -187,8 +181,25 @@ abstract class Text # # 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 @@ -247,8 +258,8 @@ abstract class Text 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 @@ -278,12 +289,10 @@ abstract class Text 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 @@ -300,7 +309,8 @@ abstract class Text # assert "0G".is_hex == false fun is_hex: Bool do - for c in self.chars 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 @@ -316,7 +326,8 @@ abstract class Text # 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 @@ -329,7 +340,8 @@ abstract class Text # 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 @@ -363,7 +375,7 @@ abstract class Text if iter.item.ascii > 32 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 @@ -374,12 +386,40 @@ abstract class Text # assert "\na\nb\tc\t".trim == "a\nb\tc" fun trim: SELFTYPE do return (self.l_trim).r_trim + # 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) + # + # 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): SELFTYPE + do + var diff = length - self.length + if diff <= 0 then return self + 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 fun to_cmangle: String do var res = new FlatBuffer var underscore = false - for c in self.chars do + for i in [0..length[ do + var c = chars[i] if (c >= 'a' and c <= 'z') or (c >='A' and c <= 'Z') then res.add(c) underscore = false @@ -412,7 +452,8 @@ abstract class Text 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 == '\0' then @@ -453,6 +494,50 @@ abstract class Text # 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" @@ -464,7 +549,8 @@ abstract class Text 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 @@ -498,7 +584,8 @@ abstract class Text do var buf = new FlatBuffer - for c in self.chars do + 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 @@ -554,6 +641,29 @@ abstract class Text return buf.to_s end + # Escape the four characters `<`, `>`, `&`, and `"` with their html counterpart + # + # assert "a&b->\"x\"".html_escape == "a&b->"x"" + fun html_escape: SELFTYPE + 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 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. # @@ -602,6 +712,14 @@ abstract class Text 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 @@ -613,7 +731,8 @@ abstract class Text # 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 @@ -628,7 +747,11 @@ 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 @@ -658,11 +781,6 @@ private abstract class StringCharView private var target: SELFTYPE - private init(tgt: SELFTYPE) - do - target = tgt - end - redef fun is_empty do return target.is_empty redef fun length do return target.length @@ -704,6 +822,8 @@ abstract class String 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" @@ -739,7 +859,8 @@ abstract class String 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 @@ -777,7 +898,8 @@ abstract class String 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 @@ -793,6 +915,38 @@ abstract class String 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 @@ -800,8 +954,6 @@ 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) @@ -818,13 +970,22 @@ class FlatString 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) + 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 # ################################################ @@ -931,15 +1092,15 @@ class FlatString 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 + if real_items != null then + return real_items.as(not null) + else var newItems = calloc_string(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) @@ -1021,8 +1182,9 @@ class FlatString 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 @@ -1095,7 +1257,7 @@ private class FlatStringReverseIterator 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] @@ -1160,6 +1322,14 @@ abstract class Buffer # 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 @@ -1222,6 +1392,38 @@ abstract class Buffer # 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: Char = 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 @@ -1246,6 +1448,24 @@ class FlatBuffer 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 @@ -1253,6 +1473,7 @@ class FlatBuffer add(item) return end + if written then reset assert index >= 0 and index < length items[index] = item end @@ -1267,6 +1488,7 @@ class FlatBuffer redef fun clear do is_dirty = true + if written then reset length = 0 end @@ -1277,6 +1499,9 @@ class FlatBuffer var c = capacity if cap <= c then return while c <= cap do c = c * 2 + 2 + # The COW flag can be set at false here, since + # it does a copy of the current `Buffer` + written = false var a = calloc_string(c+1) if length > 0 then items.copy_to(a, length, 0, 0) items = a @@ -1285,7 +1510,9 @@ class FlatBuffer 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 @@ -1314,8 +1541,9 @@ class FlatBuffer 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 @@ -1343,8 +1571,9 @@ class FlatBuffer 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 @@ -1381,6 +1610,7 @@ class FlatBuffer redef fun reverse do + written = false var ns = calloc_string(capacity) var si = length - 1 var ni = 0 @@ -1403,6 +1633,7 @@ class FlatBuffer redef fun upper do + if written then reset var it = items var id = length - 1 while id >= 0 do @@ -1413,6 +1644,7 @@ class FlatBuffer redef fun lower do + if written then reset var it = items var id = length - 1 while id >= 0 do @@ -1626,12 +1858,12 @@ redef class Int 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 @@ -1650,13 +1882,15 @@ redef class Float 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" @@ -1667,25 +1901,34 @@ redef class Float 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 + + if i < 0 then p1 = "-" + p1 + + return p1 + "." + p2 end # `self` representation with `nb` digits after the '.'. @@ -1787,18 +2030,51 @@ redef class Collection[E] end redef class Array[E] + # Fast implementation redef fun to_s do - var s = new FlatBuffer - var i = 0 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 e = self[i] - if e != null then s.append(e.to_s) + 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 - 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 @@ -1843,7 +2119,8 @@ end # 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 fun [](index: Int): Char is intern fun []=(index: Int, item: Char) is intern fun copy_to(dest: NativeString, length: Int, from: Int, to: Int) is intern @@ -1866,7 +2143,8 @@ extern class NativeString `{ char* `} 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 fun to_s_with_copy: FlatString @@ -1874,7 +2152,10 @@ extern class NativeString `{ char* `} var length = cstring_length var new_self = calloc_string(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 @@ -1884,7 +2165,7 @@ interface StringCapable 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] @@ -1932,7 +2213,8 @@ end # # 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] @@ -1950,7 +2232,7 @@ end # 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 @@ -1962,7 +2244,7 @@ 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]