Idiomatically determine all the characters that can be used for symbols
You are encouraged to solve this task according to the task description, using any language you may know.
Idiomatically determine all the characters that can be used for symbols. The word symbols is meant things like names of variables, procedures (i.e., named fragments of programs, functions, subroutines, routines), statement labels, events or conditions, and in general, anything a computer programmer can choose to name, but not being restricted to this list. Identifiers might be another name for symbols.
The method should find the characters regardless of the hardware architecture that is being used (ASCII, EBCDIC, or other).
- Task requirements
Display the set of all the characters that can be used for symbols which can be used (allowed) by the computer program. You may want to mention what hardware architecture is being used, and if applicable, the operating system.
Note that most languages have additional restrictions on what characters can't be used for the first character of a variable or statement label, for instance. These type of restrictions needn't be addressed here (but can be mentioned).
- See also
Ada
with Ada.Strings.Maps; use Ada.Strings.Maps;
with Ada.Text_IO;
procedure Show_Valid_Identifiers is
Valid_First_Character : constant Character_Set := To_Set (Ranges => (1 => ('a', 'z'), 2 => ('A', 'Z')));
Valid_Last_Character : constant Character_Set := Valid_First_Character or To_Set (Ranges => (1 => ('0', '9')));
Valid_Middle_Character : constant Character_Set := Valid_Last_Character or To_Set ('_');
begin
Ada.Text_IO.Put_Line ("Valid for the first character: " & String (To_Sequence (Valid_First_Character)));
Ada.Text_IO.Put_Line ("Valid for any middle character: " & String (To_Sequence (Valid_Middle_Character)));
Ada.Text_IO.Put_Line ("Valid for the last character: " & String (To_Sequence (Valid_Last_Character)));
end Show_Valid_Identifiers;
- Output:
Valid for the first character: ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz Valid for any middle character: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz Valid for the last character: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz
ALGOL 68
...should also work with other Algol 68 implementations that use upper-stropping (reserved words in upper-case).
There are a number of different types of symbols that can be defined in Algol 68 (informally as follows):
- identifiers used for variables, constants, structure members, procedures
- monadic operators
- dyadic operators
- mode indicants - used for type names
Monadic and dyadic operators can be symbolic or have "bold" names. Mode indicants also have "bold" names. When upper-stropping is used, bold words are formed from upper-case letters. Algol 68G also allows underscores in bold words - other implementations of Algol 68 may also allow digits.
In the output, the characters shown for monadic and dyadic operators include the upper-case letters - these can't be mixed with symbols, e.g. +A and B- are not valid operator symbols. Additionally, the only valid two character operator name where ":" is the second character is "=:" - the surlaw operator, perhaps :).
Symbolic operator symbols can be one or two characters, optionally suffixed with := or =:.
The following uses the same approach as the AWK sample, though due to the different symbol types, rather more possible symbols have to be checked.
The sample assumes Windows/Linux is the operating system and the Algol 68 compiler/interpreter can be invoked with "a68". It should be possible to modify it for other operating systems/commands. Only 7-bit ASCII characters > space are considered
BEGIN # determine which characters can be in identifiers, etc. by trying to #
# compile test programs #
STRING source name = "_tmp.a68";
STRING a68 command = "a68 " + source name + " > _tmp.err 2>&1";
# attenpts to compile the code with "%" substituted with id, #
# returns 0 if it compiled OK, non-zero otherwise #
PROC attempt compilation = ( STRING template, id )INT:
BEGIN
STRING code := "";
# replace "%" with the identifier in the template #
FOR t pos FROM LWB template TO UPB template DO
code +:= IF template[ t pos ] /= "%"
THEN template[ t pos ]
ELSE id
FI
OD;
# output the source file and try compiling it #
FILE source file;
BOOL open error := IF open( source file, source name, stand out channel ) = 0
THEN
# opened OK - file already exists and #
# will be overwritten #
FALSE
ELSE
# failed to open the file #
# - try creating a new file #
establish( source file, source name, stand out channel ) /= 0
FI;
IF open error
THEN # failed to open the file #
print( ( "Unable to open ", source name, newline ) );
stop
ELSE # file opened OK #
put( source file, ( code ) ); # write source #
close( source file );
system( a68 command ) # compile it #
FI
END # attempt compilation # ;
# attempt to construct all two-charaacter symbols and determine whether #
# they are valid by attempting to compile a program containing them #
# only 7-bit ASCII characters > space are considered #
PROC try = ( STRING template, legend )VOID:
BEGIN
[ 0 : 127 ]BOOL first, second;
FOR i FROM LWB first TO UPB first DO
first[ i ] := second[ i ] := FALSE
OD;
FOR f FROM ABS " " + 1 TO UPB first DO
CHAR fc = REPR f;
IF attempt compilation( template, fc ) = 0
THEN
# this character can be the first character of a symbol #
first[ f ] := TRUE;
FOR s FROM ABS " " + 1 TO UPB second DO
IF NOT second[ s ]
THEN
# haven't found this is a valid second character #
# yet #
IF attempt compilation( template, fc + REPR s ) = 0
THEN
# compiled OK #
second[ s ] := TRUE
FI
FI
OD
FI
OD;
print( ( "Characters valid for ", legend, ":", newline ) );
print( ( " as first: " ) );
FOR c pos FROM LWB first TO UPB first DO
IF first[ c pos ]
THEN print( ( REPR c pos ) )
ELIF second[ c pos ]
THEN print( ( " " ) )
FI
OD;
print( ( newline ) );
print( ( " as other: " ) );
FOR c pos FROM LWB first TO UPB first DO
IF second[ c pos ]
THEN print( ( REPR c pos ) )
ELIF first[ c pos ]
THEN print( ( " " ) )
FI
OD;
print( ( newline ) )
END # try # ;
try( "BEGIN INT %; % := 1 END", "identifiers" );
try( "BEGIN OP % = ( INT a )INT: a; % 1 END", "monadic operators" );
try( "BEGIN PRIO % = 5; OP % = ( INT a, b )INT: a; 1 % 1 END", "dyadic operators" );
try( "BEGIN MODE % = INT; % x; x := 1 END", "mode indicants" )
END- Output:
Characters valid for identifiers:
as first: abcdefghijklmnopqrstuvwxyz
as other: 0123456789_abcdefghijklmnopqrstuvwxyz
Characters valid for monadic operators:
as first: !%& +- ?ABCDEFGHIJKLMNOPQRSTUVWXYZ^ ~
as other: * /<=> ABCDEFGHIJKLMNOPQRSTUVWXYZ _
Characters valid for dyadic operators:
as first: !%&*+-/ <=>?ABCDEFGHIJKLMNOPQRSTUVWXYZ^ ~
as other: * /:<=> ABCDEFGHIJKLMNOPQRSTUVWXYZ _
Characters valid for mode indicants:
as first: ABCDEFGHIJKLMNOPQRSTUVWXYZ
as other: ABCDEFGHIJKLMNOPQRSTUVWXYZ_
AWK
# usage: gawk -f Idiomatically_determine_all_the_characters_that_can_be_used_for_symbols.awk
function is_valid_identifier(id, rc) {
fn = "is_valid_identifier.awk"
printf("function unused(%s) { arr[%s] = 1 }\n", id, id, id) >fn
printf("BEGIN { exit(0) }\n") >>fn
close(fn)
rc = system("gawk -f is_valid_identifier.awk 2>errors")
return rc == 0
}
BEGIN {
for (i = 0; i <= 255; i++) {
c = sprintf("%c", i)
if (is_valid_identifier(c))
good1 = good1 c;
else
bad1 = bad1 c
if (is_valid_identifier("_" c "_"))
good2 = good2 c;
else
bad2 = bad2 c;
}
printf("1st character: %d bad, %d ok: %s\n",
length(bad1), length(good1), good1)
printf("2nd..nth char: %d bad, %d ok: %s\n",
length(bad2), length(good2), good2)
exit(0)
}
output:
1st character: 203 bad, 53 ok: ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz 2nd..nth char: 193 bad, 63 ok: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Delphi
This code test the all printable ASCII characters to see if they are valid in symbols. It uses the the Delphi system call "IsValidIdent" to determine what the compiler will accept. It starts by testing single characters identifiers, which tests the first character of the identifier. Then tests the second character. This tells which characters are valid for the rest of an identifier.
procedure ShowValidSymbols(Memo: TMemo);
{Uses Delphi system tool "IsValidIndent" }
{To identify valid characters in indentifiers}
var I: integer;
var TS: string;
var Good,Bad: string;
begin
{Test first characters in a symbol}
Good:=''; Bad:='';
for I:=$21 to $7F do
begin
TS:=Char(I);
if IsValidIdent(TS) then Good:=Good+TS
else Bad:=Bad+TS;
end;
Memo.Lines.Add('First Characters Allowed');
Memo.Lines.Add('Allowed: '+Good);
Memo.Lines.Add('Not Allowed: '+Bad);
{Test remaining characters in a symbol}
Good:=''; Bad:='';
for I:=$21 to $7F do
begin
TS:='A'+Char(I);
if IsValidIdent(TS) then Good:=Good+TS[2]
else Bad:=Bad+TS[2];
end;
Memo.Lines.Add('');
Memo.Lines.Add('Remaining Characters Allowed');
Memo.Lines.Add('Allowed: '+Good);
Memo.Lines.Add('Not Allowed: '+Bad);
end;
- Output:
First Characters Allowed
Allowed: ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Not Allowed: !"#$%&'()*+,-./0123456789:;<=>?@[\]^`{|}~?
Remaining Characters Allowed
Allowed: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Not Allowed: !"#$%&'()*+,-./:;<=>?@[\]^`{|}~?
Elapsed Time: 9.048 ms.
EasyLang
print "First character:"
write "_"
for i = strcode "A" to strcode "Z"
write strchar i
.
for i = strcode "a" to strcode "z"
write strchar i
.
print ""
print "Not as the first character additionally:"
for i = strcode "0" to strcode "9"
write strchar i
.
- Output:
First character: _ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz Not as the first character additionally: 0123456789
F#
Well, if the purpose of this task is to determine what can be used as an identifier then in F# anything so long as you enclose it in double backticks so:
let ``+`` = 5
printfn "%d" ``+``
- Output:
5
Is this idiotmatically determined?
Factor
USING: parser see ;
\ scan-word-name see
- Output:
: scan-word-name ( -- string )
scan-token dup "\"" = [ t ] [ dup string>number ] if
[ invalid-word-name ] when ;
From this code we can see that any characters may be used in an identifier unless it parses as a string or a number.
FreeBASIC
Dim As String*1 C1
Dim As Integer C
Print "First character set: ";
For C = 0 To 255
If (Chr(C) >= "A" And Chr(C) <="Z") Or Chr(C)="_" Then Print Chr(C);
Next
Print !"\nNext characters set: ";
For C = 0 To 255
C1 = Chr(C)
If (C1 >= "A" And C1 <= "Z") Or (C1 >= "0" And C1 <= "9") Or C1 = "_" Or (C1 >= "a" And C1 <= "z") Then Print C1;
Next C
Sleep
- Output:
First character set: ABCDEFGHIJKLMNOPQRSTUVWXYZ_ Next characters set: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Go
This program uses the Go parser to check whether an identifier is indeed valid. It checks separately which Unicode code points may appear at the beginning of an identifier, or in the remaining name. The code assumes that the Go language does not have a keyword or otherwise reserved symbol of length 1, or of length 2 starting with the underscore.
package main
import (
"fmt"
"go/ast"
"go/parser"
"strings"
"unicode"
)
func isValidIdentifier(identifier string) bool {
node, err := parser.ParseExpr(identifier)
if err != nil {
return false
}
ident, ok := node.(*ast.Ident)
return ok && ident.Name == identifier
}
type runeRanges struct {
ranges []string
hasStart bool
start rune
end rune
}
func (r *runeRanges) add(cp rune) {
if !r.hasStart {
r.hasStart = true
r.start = cp
r.end = cp
return
}
if cp == r.end+1 {
r.end = cp
return
}
r.writeTo(&r.ranges)
r.start = cp
r.end = cp
}
func (r *runeRanges) writeTo(ranges *[]string) {
if r.hasStart {
if r.start == r.end {
*ranges = append(*ranges, fmt.Sprintf("%U", r.end))
} else {
*ranges = append(*ranges, fmt.Sprintf("%U-%U", r.start, r.end))
}
}
}
func (r *runeRanges) String() string {
ranges := r.ranges
r.writeTo(&ranges)
return strings.Join(ranges, ", ")
}
func main() {
var validFirst runeRanges
var validFollow runeRanges
var validOnlyFollow runeRanges
for r := rune(0); r <= unicode.MaxRune; r++ {
first := isValidIdentifier(string([]rune{r}))
follow := isValidIdentifier(string([]rune{'_', r}))
if first {
validFirst.add(r)
}
if follow {
validFollow.add(r)
}
if follow && !first {
validOnlyFollow.add(r)
}
}
_, _ = fmt.Println("Valid first:", validFirst.String())
_, _ = fmt.Println("Valid follow:", validFollow.String())
_, _ = fmt.Println("Only follow:", validOnlyFollow.String())
}
- Output:
Valid first: U+0041-U+005A, U+005F, U+0061-U+007A, U+00AA, ..., U+00F8-U+02C1, U+02C6-U+02D1, ... Valid follow: U+0030-U+0039, U+0041-U+005A, U+005F, U+0061-U+007A, U+00AA, ..., U+00F8-U+02C1, ..., U+2CEB0-U+2EBE0, U+2F800-U+2FA1D Only follow: U+0030-U+0039, U+0660-U+0669, U+06F0-U+06F9, U+07C0-U+07C9, ..., U+1D7CE-U+1D7FF, U+1E950-U+1E959
Haskell
Quotation from the Haskell 2010 language specification [1]
varid → (small {small | large | digit | ' }) / reservedid
conid → large {small | large | digit | ' }
reservedid → case | class | data | default | deriving | do | else
| foreign | if | import | in | infix | infixl
| infixr | instance | let | module | newtype | of
| then | type | where | _
small → ascSmall | uniSmall | _
ascSmall → a | b | … | z
uniSmall → any Unicode lowercase letter
large → ascLarge | uniLarge
ascLarge → A | B | … | Z
uniLarge → any uppercase or titlecase Unicode letter
digit → ascDigit | uniDigit
ascDigit → 0 | 1 | … | 9
uniDigit → any Unicode decimal digit
An identifier consists of a letter followed by zero or more letters, digits, underscores, and single quotes. Identifiers are lexically distinguished into two namespaces: those that begin with a lowercase letter (variable identifiers) and those that begin with an upper-case letter (constructor identifiers). Identifiers are case sensitive: name, naMe, and Name are three distinct identifiers (the first two are variable identifiers, the last is a constructor identifier).
Underscore, “_”, is treated as a lowercase letter, and can occur wherever a lowercase letter can. However, “_” all by itself is a reserved identifier, used as wild card in patterns.
According to the specification we may give predicates for valid symbols and identifiers in Haskell:
import Data.Char
-- predicate for valid symbol
isSymbolic ch = isAlphaNum ch || ch `elem` "_'"
-- predicate for valid type construtor
isConId s = and [ not (null s)
, isUpper (head s)
, all isSymbolic (tail s) ]
-- predicate for valid identifier
isVarId s = and [ not (null s)
, isLower (head s)
, all isSymbolic (tail s)
, not (isReserved s) ]
-- predicate for reserved words
isReserved s = elem s ["case", "class", "data", "default", "deriving", "do "
, "else", "foreign", "if", "import", "in", "infix "
, "infixl", "infixr", "instance", "let", "module "
, "newtype", "of", "then", "type", "where", "_"
J
J is defined in terms of ascii, but that would not prevent it from being ported to other environments. But we can still use J's parser to determine if a specific character combination is a single, legal word:
a.#~1=#@;: ::0:"1 'b',.a.,.'c'
0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Here, a. is the set of chararacters we are testing. We prefix each of these with an arbitrary letter, and suffix each with an arbitrary character and then try counting how many parsed tokens are formed by the result. If the token count is 1, then that character was a legal word-forming character.
Of course, we also only need to do this once. Once we have a set of these characters, it's faster and easier to use a set membership test on the characters themselves than on the expression which generates them.
Java
import java.util.function.IntPredicate;
import java.util.stream.IntStream;
public class Test {
public static void main(String[] args) throws Exception {
print("Java Identifier start: ", 0, 0x10FFFF, 72,
Character::isJavaIdentifierStart, "%c");
print("Java Identifier part: ", 0, 0x10FFFF, 25,
Character::isJavaIdentifierPart, "[%d]");
print("Identifier ignorable: ", 0, 0x10FFFF, 25,
Character::isIdentifierIgnorable, "[%d]");
print("Unicode Identifier start: ", 0, 0x10FFFF, 72,
Character::isUnicodeIdentifierStart, "%c");
print("Unicode Identifier part : ", 0, 0x10FFFF, 25,
Character::isUnicodeIdentifierPart, "[%d]");
}
static void print(String msg, int start, int end, int limit,
IntPredicate p, String fmt) {
System.out.print(msg);
IntStream.rangeClosed(start, end)
.filter(p)
.limit(limit)
.forEach(cp -> System.out.printf(fmt, cp));
System.out.println("...");
}
}
Java Identifier start: $ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz¢£¤¥ªµºÀÁÂÃÄÅÆÇÈÉÊ... Java Identifier part: [0][1][2][3][4][5][6][7][8][14][15][16][17][18][19][20][21][22][23][24][25][26][27][36][48]... Java Identifier ignorable: [0][1][2][3][4][5][6][7][8][14][15][16][17][18][19][20][21][22][23][24][25][26][27][127][128]... Unicode Identifier start: ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyzªµºÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐ... Unicode Identifier part: [0][1][2][3][4][5][6][7][8][14][15][16][17][18][19][20][21][22][23][24][25][26][27][48][49]...
jq
jq identifiers
Excluding key names from consideration, in jq 1.4 the set of characters that can be used in jq identifiers corresponds to the regex: [A-Za-z0-9$_]. Thus, assuming the availability of test/1 as a builtin, the test in jq for a valid identifier character is: test("[A-Za-z0-9$_]").
To generate a string of such characters idiomatically:
[range(0;128) | [.] | implode | select(test("[A-Za-z0-9$_]"))] | addjq 1.5 also allows ":" as a joining character in the form "module::name".
JSON key names
Any JSON string can be used as a key. Accordingly, some characters must be entered as escaped character sequences, e.g. \u0000 for NUL, \\ for backslash, etc. Thus any Unicode character except for the control characters can appear in a jq key. Therefore, assuming the availability in jq of the test/1 builtin, the test in jq for whether a character can appear literally in a jq identifier or key is:
test("[^\u0000-\u0007F]")Symbols
The following function screens for characters by "\p" class:
def is_character(class):
test( "\\p{" + class + "}" );For example, to test whether a character is a Unicode letter, symbol or numeric character:
is_character("L") or is_character("S") or is_character("N")An efficient way to count the number of Unicode characters within a character class is to use the technique illustrated by the following function:
def count(class; m; n):
reduce (range(m;n) | [.] | implode | select( test( "\\p{" + class + "}" ))) as $i
(0; . + 1);For example the number of Unicode "symbol" characters can be obtained by evaluating:
count("S"; 0; 1114112)The result is 3958.
Julia
Julia allows almost any Unicode character as a symbol if it is not a reserved operator symbol and, if numeric, if it is not the first letter of a symbol name.
For example, x2 is a valid identifier, but 2x is not-- it is interpreted as 2 times the identifier x. In Julia, the Symbol() function turns a string into a symbolic token. So, for example:
for i in 1:0x200000 - 1
Symbol("x" * Char(i))
end
When run, this loop runs without error up to 0x200000 but not at Unicode symbol numbered 0x200000.
Kotlin
According to the Kotlin grammar, the rules regarding which characters can appear in symbols (or identifiers as we usually call them) are the same as in Java, namely:
1. An identifier is a sequence of any number of unicode letters or digits, other than a reserved word.
2. Identifiers are case sensitive.
3. The first character must be a letter, an underscore or a $ sign. Subsequent characters can include digits and certain control characters as well though the latter are ignored for identifier matching purposes.
However, in practice, identifiers which include a $ symbol or control characters don't compile unless (in the case of $) the entire identifier is enclosed in back-ticks. The use of this device also allows one to use a reserved word or many otherwise prohibited unicode characters in an identifier including spaces and dashes.
A Kotlin label name is a valid identifier followed by an @ symbol and an annotation name is an identifier preceded by an @ symbol.
// version 1.1.4-3
typealias CharPredicate = (Char) -> Boolean
fun printChars(msg: String, start: Int, end: Int, limit: Int, p: CharPredicate, asInt: Boolean) {
print(msg)
(start until end).map { it.toChar() }
.filter { p(it) }
.take(limit)
.forEach { print(if (asInt) "[${it.toInt()}]" else it) }
println("...")
}
fun main(args: Array<String>) {
printChars("Kotlin Identifier start: ", 0, 0x10FFFF, 72,
Char::isJavaIdentifierStart, false)
printChars("Kotlin Identifier part: ", 0, 0x10FFFF, 25,
Character::isJavaIdentifierPart, true)
printChars("Kotlin Identifier ignorable: ", 0, 0x10FFFF, 25,
Character::isIdentifierIgnorable, true)
}
- Output:
Kotlin Identifier start: $ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz¢£¤¥ªµºÀÁÂÃÄÅÆÇÈÉÊ... Kotlin Identifier part: [0][1][2][3][4][5][6][7][8][14][15][16][17][18][19][20][21][22][23][24][25][26][27][36][48]... Kotlin Identifier ignorable: [0][1][2][3][4][5][6][7][8][14][15][16][17][18][19][20][21][22][23][24][25][26][27][127][128]...
Lua
From the 5.4 reference manual: "Names (also called identifiers) in Lua can be any string of Latin letters, Arabic-Indic digits, and underscores, not beginning with a digit and not being a reserved word."
function isValidIdentifier(id)
local reserved = {
["and"]=true, ["break"]=true, ["do"]=true, ["end"]=true, ["else"]=true, ["elseif"]=true, ["end"]=true,
["false"]=true, ["for"]=true, ["function"]=true, ["goto"]=true, ["if"]=true, ["in"]=true,
["local"]=true, ["nil"]=true, ["not"]=true, ["or"]=true, ["repeat"]=true, ["return"]=true,
["then"]=true, ["true"]=true, ["until"]=true, ["while"]=true }
return id:find("^[a-zA-Z_][a-zA-Z0-9_]*$") ~= nil and not reserved[id]
end
vfc, vsc = {}, {}
for i = 0, 255 do
local c = string.char(i)
if isValidIdentifier(c) then vfc[#vfc+1]=c end
if isValidIdentifier("_"..c) then vsc[#vsc+1]=c end
end
print("Valid First Characters: " .. table.concat(vfc))
print("Valid Subsequent Characters: " .. table.concat(vsc))
- Output:
Valid First Characters: ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz Valid Subsequent Characters: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Mathematica /Wolfram Language
chars = Characters[FromCharacterCode[Range[0, 1114111]]];
out = Reap[Do[
If[Quiet[Length[Symbol[c]] == 0],
Sow[c]
]
,
{c, chars}
]][[2, 1]];
Print["Possible 1st characters: ", out // Length]
out = Reap[Do[
If[Quiet[Length[Symbol["a" <> c]] == 0],
Sow[c]
]
,
{c, chars}
]][[2, 1]];
Print["Possible 2nd-nth characters: ", out // Length]
- Output:
In Wolfram Language almost all characters (there are 1114112 characters defined) can be used in variable/function names. I can't show all valid characters as there are over a million that are allowed. I do not show the list of characters 'out' but rather their length for practical purposes:
Possible 1st characters: 1113704 Possible 2nd-nth characters: 1113726
Nim
As regards identifiers, there exists a general rule which describes how they can be formed. For this rule, the following program prints the allowed starting characters and the allowed characters:
import sequtils, strutils
echo "Allowed starting characters for identifiers:"
echo toSeq(IdentStartChars).join()
echo ""
echo "Allowed characters in identifiers:"
echo toSeq(IdentChars).join()
- Output:
Allowed starting characters for identifiers: ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz Allowed characters in identifiers: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
But Nim is a lot more flexible and allows using Unicode symbols in identifiers provided these are letters and digits. Thus, the following program is valid:
var à⁷ = 3
echo à⁷
Using escape character `, it is possible to override the rules and to include any character in an identifier and even to use a keyword as identifier. Here is an example of the possibilities:
var `const`= 3
echo `const`
proc `<`(a, b: int): bool =
echo a, " ", b
system.`<`(a, b)
echo 4 < 7
proc `Π`(a: varargs[int]): int =
result = 1
for n in a: result *= n
echo Π(4, 5, 7)
var `1` = 2
echo `1`
Ol
Absolutely any Unicode or ANSI character can be used as part of symbol name. There only some limitations in form of symbol declaration.
1. Direct symbol declaration (in form of quote or ') must not be started from control codes (first 32 characters), numbers and @. Next characters in symbol must not be control code neither @.
2. Direct symbol declaration (in form of ||) must not contain character |.
3. Functional symbol creation (in form of string->symbol) have no any limitations.
ooRexx
Although this program does not use any feature that is not in Classic Rexx, it is included here to show what characters are valid for symbols in ooRexx.
/*REXX program determines what characters are valid for REXX symbols.*/
/* copied from REXX version 2 */
Parse Version v
Say v
symbol_characters='' /* start with no chars */
do j=0 To 255 /* loop through all the chars.*/
c=d2c(j) /* convert number to character*/
if datatype(c,'S') then /* Symbol char */
symbol_characters=symbol_characters || c /* add to list. */
end
say 'symbol characters:' symbol_characters /*display all */
- Output:
REXX-ooRexx_4.2.0(MT)_32-bit 6.04 22 Feb 2014 symbol characters: !.0123456789?ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
PARI/GP
The only symbols that can be used in variable names (including function names as a special case) are a-z, A-Z, 0-9, and the underscore. Additionally, the first character must be a letter. (That is, they must match this regex: [a-zA-Z][a-zA-Z0-9_]*.)
v=concat(concat([48..57],[65..90]),concat([97..122],95));
apply(Strchr,v)- Output:
%1 = ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M", "N", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z", "a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "_"]
Perl
# When not using the <code>use utf8</code> pragma, any word character in the ASCII range is allowed.
# the loop below returns: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
for $i (0..0x7f) {
$c = chr($i);
print $c if $c =~ /\w/;
}
# When 'use utf8' is in force, the same holds true, but the Unicode-aware version of the 'word-character' test is used.
# not supplying output, too much of it
use utf8;
binmode STDOUT, ":utf8";
for (0..0x1ffff) {
$c = chr($_);
print $c if $c =~ /\p{Word}/;
}
Phix
Translation of AWK, extended with separation of ansi and utf8 handling
without js -- file i/o, system_exec, \t and \r chars function run(string ident) integer fn = open("test.exw","w") printf(fn,"object %s",ident) close(fn) return system_exec("p -batch test.exw") end function function check(integer lo, hi) string ok1 = "", ok2 = "" integer ng1 = 0, ng2 = 0 for ch=lo to hi do printf(1,"%d/%d...\r",{ch,hi}) if find(ch,"\t\r\n \0\x1A;") then ng1 += 1 ng2 += 1 else string c = sprintf("%c",ch) if run(c)==0 then ok1 &= c else ng1 += 1 end if if run("_"&c)==0 then ok2 &= c else ng2 += 1 end if end if end for return {{ng1,length(ok1),ok1}, {ng2,length(ok2),ok2}} end function sequence r = check(0,127) printf(1,"ansi characters:\n===============\n") printf(1,"1st character: %d bad, %d OK %s\n",r[1]) printf(1,"2nd..nth char: %d bad, %d OK %s\n\n",r[2]) r = check(128,255) integer ok8 = 0, ng8 = 0 for i=#80 to #10FFFF do if i<#D800 or i>#DFFF then printf(1,"#%x/#10FFFF...\r",i) string utf8 = utf32_to_utf8({i}) bool ok = true if not find(utf8[1],r[1][3]) then ok = false else for j=2 to length(utf8) do if not find(utf8[j],r[2][3]) then ok = false exit end if end for end if if ok then ok8 += 1 else ng8 += 1 end if end if end for printf(1,"utf8 characters: \n===============\n") printf(1,"bad:%,d, good:%,d\n",{ng8,ok8})
- Output:
ansi characters: =============== 1st character: 75 bad, 53 OK ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz 2nd..nth char: 65 bad, 63 OK 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz utf8 characters: =============== bad:0, good:1,111,936
Note that versions prior to 0.8.1 only permit a mere 48 utf8 characters, running the same code on 0.7.9 gave me
utf8 characters: =============== bad:1,111,888, good:48
Python
See String class isidentifier.
Quackery
[ $ "0123456789AaBbCcDdEeFfGgHhIiJjKkLlMmNnOoPpQqRrS"
$ QsTtUuVvWwXxYyZz()[]{}<>~=+-*/^\|_.,:;?!'"`%@&#$Q
join ] constant is tokenchars ( --> $ )
( The first non-whitespace character after the word $
(pronounced "string") is deemed to be the delimiter
for the string that follows it. In the first string
the conventional character " is used, so cannot
appear as a character in that string. In the second
string all the reasonable delimiters are used, so Q
is used as the delimiter.
As it is not possible to make a string that uses all
the characters, two strings are concatenated (join)
to make the string during compilation. (Which is why
$ "0...S" $ Qs...$Q join is nested (inside [ ... ])
and followed by the word constant, which causes the
nest to be evaluated during compilation.)
Regardless of operating system, Quackery only knows
the characters in the string tokenchars, plus space
and carriage return.
The characters in tokenchars are in QACSFOT order
(the Quackery Arbitrary Character Sequence For
Ordered Text) which it uses for string comparison,
but the valid tokens (which is all of them) will
be printed by alltokens in the order native to the
operating system. (In this instance, Unicode.) )
[ tokenchars find
tokenchars found ] is validtoken ( c --> b )
[ 256 times
[ i^ validtoken if [ i^ emit ] ] ] is alltokens ( --> )
alltokensOutput:
!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~
Racket
Symbols in the Racket Guide states that:
Any string (i.e., any character sequence) can be supplied to
string->symbolto obtain the corresponding symbol.
Reading Symbols defines what symbols can be "read" without needing quoting.
The docuementation for
integer->char
says that a character must lie in the ranges: 0 to 55295, and 57344 to 1114111.
That's too much to be printing out here... call (main) yourself, at home.
#lang racket
;; Symbols that don't need to be specially quoted:
(printf "~s~%" '(a a-z 3rd ...---... .hidden-files-look-like-this))
;; Symbols that do need to be specially quoted:
(define bar-sym-list
`(|3|
|i have a space|
|i've got a quote in me|
|i'm not a "dot on my own", but my neighbour is!|
|.|
,(string->symbol "\u03bb")
,(string->symbol "my characters aren't even mapped in unicode \U10e443")))
(printf "~s~%" bar-sym-list)
(printf "~a~%" bar-sym-list)
(define (main)
(for
((c (sequence-map
integer->char
(in-sequences (in-range 0 (add1 55295))
(in-range 57344 (add1 1114111)))))
(i (in-naturals 1)))
(when (zero? (modulo i 80)) (newline))
(display (list->string (list c)))))
- Output:
(a a-z 3rd ...---... .hidden-files-look-like-this) (|3| |i have a space| |i've got a quote in me| |i'm not a "dot on my own", but my neighbour is!| |.| λ |my characters aren't even mapped in unicode |) (3 i have a space i've got a quote in me i'm not a "dot on my own", but my neighbour is! . λ my characters aren't even mapped in unicode )
The output to (main) is massive, and probably not dissimilar to Tcl's (anyone want to compare?)
Raku
(formerly Perl 6) Any Unicode character or combination of characters can be used for symbols in Raku. Here's some counting rods and some cuneiform:
sub postfix:<𒋦>($n) { say "$n trilobites" }
sub term:<𝍧> { unival('𝍧') }
𝍧𒋦
- Output:
8 trilobites
And here is a Zalgo-text symbol:
sub Z̧̔ͩ͌͑̉̎A̢̲̙̮̹̮͍̎L̔ͧ́͆G̰̬͎͔̱̅ͣͫO͙̔ͣ̈́̈̽̎ͣ ($n) { say "$n COMES" }
Z̧̔ͩ͌͑̉̎A̢̲̙̮̹̮͍̎L̔ͧ́͆G̰̬͎͔̱̅ͣͫO͙̔ͣ̈́̈̽̎ͣ 'HE'
- Output:
HE COMES
Of course, as in other languages, most of the characters you'll typically see in names are going to be alphanumerics from ASCII (or maybe Unicode), but that's a convention, not a limitation, due to the syntactic category notation demonstrated above, which can introduce any sequence of characters as a term or operator.
Actually, the above is a slight prevarication. The syntactic category notation does not allow you to use whitespace in the definition of a new symbol. But that leaves many more characters allowed than not allowed. Hence, it is much easier to enumerate the characters that cannot be used in symbols:
say .fmt("%4x"),"\t", uniname($_)
if uniprop($_,'Z')
for 0..0x1ffff;
- Output:
20 SPACE a0 NO-BREAK SPACE 1680 OGHAM SPACE MARK 2000 EN QUAD 2001 EM QUAD 2002 EN SPACE 2003 EM SPACE 2004 THREE-PER-EM SPACE 2005 FOUR-PER-EM SPACE 2006 SIX-PER-EM SPACE 2007 FIGURE SPACE 2008 PUNCTUATION SPACE 2009 THIN SPACE 200a HAIR SPACE 2028 LINE SEPARATOR 2029 PARAGRAPH SEPARATOR 202f NARROW NO-BREAK SPACE 205f MEDIUM MATHEMATICAL SPACE 3000 IDEOGRAPHIC SPACE
We enforce the whitespace restriction to prevent insanity in the readers of programs. That being said, even the whitespace restriction is arbitrary, and can be bypassed by deriving a new grammar and switching to it. We view all other languages as dialects of Raku, even the insane ones. :-)
REXX
version 1
/*REXX program determines what characters are valid for REXX symbols. */
@= /*set symbol characters " " */
do j=0 for 2**8 /*traipse through all the chars. */
_=d2c(j) /*convert decimal number to char.*/
if datatype(_,'S') then @=@ || _ /*Symbol char? Then add to list.*/
end /*j*/ /* [↑] put some chars into a list*/
say ' symbol characters: ' @ /*display all symbol characters.*/
/*stick a fork in it, we're done.*/
Programming note: REXX allows any symbol to begin a (statement) label, but variables can't begin with a period (.) or a numeric digit.
All examples below were executed on a (ASCII) PC using Windows/XP and Windows/7 with code page 437 in a DOS window.
Using PC/REXX
Using Personal REXX
Using Regina (versions 3.2 ───► 3.82)
output
symbol characters: !#$.0123456789?@ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
Using R4
output
symbol characters: !#$.0123456789?@ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyzÇüéâäàåçêëèïîìÄÅÉæÆôöòûùÿÖÜ¢£áíóúñÑ╡╢╖─╞╟╨╤╥╙╘╒╓╫╪▐αßΓπΣσµτΦΘΩδ∞φ
Using ROO
output
symbol characters: !#$.0123456789?@ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyzÇüéâäàåçêëèïîìÄÅÉæÆôöòûùÿÖÜ¢£áíóúñÑ╡╢╖╞╟╨╤╥╙╘╒╓╫╪▐αßΓπΣσµτΦΘΩδ∞φ
version 2 ooRexx compatible
Because version 1 does not work correctly with ooRexx - showing this error message:
2 *-* @
Error 13 running D:\v1.rex line 2: Invalid character in program
Error 13.1: Incorrect character in program "@" ('40'X)
I've added version 2 which should work correctly for all Rexx interpreters and compilers
/*REXX program determines what characters are valid for REXX symbols.*/
/* version 1 adapted for general acceptance */
Parse Version v
Say v
symbol_characters='' /* start with no chars */
do j=0 To 255 /* loop through all the chars.*/
c=d2c(j) /* convert number to character*/
if datatype(c,'S') then /* Symbol char */
symbol_characters=symbol_characters || c /* add to list. */
end
say 'symbol characters:' symbol_characters /*display all */
- Output:
for some interpreters
Note that $#@ are not valid symbol characters for ooRexx.
REXX-ooRexx_4.2.0(MT)_32-bit 6.04 22 Feb 2014 symbol characters: !.0123456789?ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz REXX-Regina_3.8.2(MT) 5.00 22 Jun 2014 symbol characters: !#$.0123456789?@ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
RPL
The RPL character set is an 8-bit character set, sometimes referred to simply as "ECMA-94" in documentation, although it is for the most part a variant of ISO/IEC 8859-1 / ECMA-94. See the related Wikipedia entry for more details.
≪ "" "'A '"
1 255 FOR c
3 c CHR REPL
IFERR DUP STR→ THEN DROP
ELSE
IF 'A' SAME NOT THEN SWAP c CHR + SWAP END
END
NEXT DROP
≫ 'SYMBOLS' STO
- Output:
1: "!$%&.0123456789?ABCDEFGHIJKLMNOPQRSTUVWXYZ\abcdefghijklmnopqrstuvwxyz~∇∑▶πα→←↓↑γδεηθλρστωΔΠΩ▬∞ ¡¢£¤¥¦§¨©ª¬®¯°±²³´µ¶·¸¹º¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ"
Scala
- Output:
Best seen running in your browser either by ScalaFiddle (ES aka JavaScript, non JVM) or Scastie (remote JVM).
object IdiomaticallyDetermineSymbols extends App {
private def print(msg: String, limit: Int, p: Int => Boolean, fmt: String) =
println(msg + (0 to 0x10FFFF).filter(p).take(limit).map(fmt.format(_)).mkString + "...")
print("Java Identifier start: ", 72, cp => Character.isJavaIdentifierStart(cp), "%c")
print("Java Identifier part: ", 25, cp => Character.isJavaIdentifierPart(cp), "[%d]")
print("Identifier ignorable: ", 25, cp => Character.isIdentifierIgnorable(cp), "[%d]")
print("Unicode Identifier start: ", 72, cp => Character.isUnicodeIdentifierStart(cp), "%c")
print("Unicode Identifier part : ", 25, cp => Character.isUnicodeIdentifierPart(cp), "[%d]")
}
Tcl
Tcl permits any character to be used in a variable or command name (subject to the restriction that :: is a namespace separator and, for variables only, a (…) sequence is an array reference). The set of characters that can be used after $ is more restricted, excluding many non-letter-like symbols, but still large. It is recommended practice to only use ASCII characters for variable names as this makes scripts more resistant to the majority of encoding problems when transporting them between systems, but the language does not itself impose such a restriction.
for {set c 0;set printed 0;set special {}} {$c <= 0xffff} {incr c} {
set ch [format "%c" $c]
set v "_${ch}_"
#puts "testing variable named $v"
if {[catch {set $v $c; set $v} msg] || $msg ne $c} {
puts [format "\\u%04x illegal in names" $c]
incr printed
} elseif {[catch {subst $$v} msg] == 0 && $msg eq $c} {
lappend special $ch
}
}
if {$printed == 0} {
puts "All Unicode characters legal in names"
}
puts "Characters legal after \$: $special"
- Output:
Only the first 256 characters are displayed:
All Unicode characters legal in names Characters legal after $: 0 1 2 3 4 5 6 7 8 9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z _ a b c d e f g h i j k l m n o p q r s t u v w x y z ª µ º À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ø ù ú û ü ý þ ÿ İ ı IJ ij Ĵ ĵ Ķ ķ ĸ Ĺ Ł ł Ń ń Ņ ņ Ň ň ʼn Ŋ ŋ Ō ō Ŏ ŏ Ő ő Œ œ Ŕ ŕ Ŗ ŗ Ř ř Ś ş š Ţ ţ Ť ť Ŧ ŧ Ũ ũ Ū ū Ŭ ŭ Ů ů Ű ű Ų ų Ŵ ŵ Ŷ ŷ Ÿ Ź ź ƪ Ƶ ƺ ǀ ǁ ǂ ǃ DŽ Dž dž LJ Lj lj NJ Nj nj Ǎ ǎ Ǐ ǐ Ǒ ǒ Ǔ ǔ Ǖ ǖ ǘ Ǚ ǚ Ǜ ǜ ǝ Ǟ ǟ Ǡ ǡ Ǣ ǣ Ǥ ǥ Ǧ ǧ Ǩ ǩ Ǫ ǫ Ǭ ǭ Ǯ ǯ ǰ DZ Dz dz Ǵ ǵ Ƕ Ǹ ǹ Ǻ ǻ Ǽ ǽ Ǿ ǿ ...
Wren
In Wren identifiers consist of upper and lower case letters, digits or underscores and must begin with either a letter or an underscore. Only ASCII letters or digits can be used, though Wren supports Unicode for other purposes.
Identifiers which begin with underscores can only be used as instance field names (one underscore) or static field names (two or more underscores).
for (i in 97..122) System.write(String.fromByte(i))
for (i in 65..90) System.write(String.fromByte(i))
for (i in 48..57) System.write(String.fromByte(i))
System.print("_")
- Output:
abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_
XPL0
Paraphrasing code from the compiler's parser:
char C, C1;
[Text(0, "First character set: ");
for C:= 0 to 255 do
if C>=^A then if C<=^Z ! C=^_ then
ChOut(0, C);
CrLf(0);
Text(0, "Next characters set: ");
for C:= 0 to 255 do
[if C>=^a & C<=^z then C1:= C & $DF \to uppercase
else C1:= C;
case of
C1>=^A & C1<=^Z, C1>=^0 & C1<=^9, C1=^_ :
ChOut(0, C)
other [];
];
CrLf(0);
]- Output:
First character set: ABCDEFGHIJKLMNOPQRSTUVWXYZ_ Next characters set: 0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz
zkl
zkl only supports ASCII, although other character sets might be finessed.
[0..255].filter(fcn(n){
try{ Compiler.Compiler.compileText("var "+n.text) }
catch{ False }
}).apply("text").concat()- Output:
<compiler noise> ;ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz
This code works by compiling "var <char>". Since "var ;" is valid syntax (dead code), ";" is a false positive. We could also use "fcn <char>{}" but "fcn {}" is lambda syntax, so space would be a false positive. "_" is excluded because it is not valid variable name although it can be anywhere in a multi-character name.
- Programming Tasks
- Solutions by Programming Task
- 6502 Assembly/Omit
- Pascal/Omit
- Ada
- ALGOL 68
- AWK
- Delphi
- SysUtils,StdCtrls
- EasyLang
- F Sharp
- Factor
- Examples needing attention
- FreeBASIC
- Go
- Haskell
- J
- Java
- Jq
- Julia
- Kotlin
- Lua
- Mathematica
- Wolfram Language
- Nim
- Ol
- OoRexx
- PARI/GP
- Perl
- Phix
- Python
- Quackery
- Racket
- Raku
- REXX
- RPL
- Scala
- Tcl
- Wren
- XPL0
- Zkl