Source   Edit  

simple types

TLDR

• Conversion between int and int32 or int64 must be explicit except for string literals.
• stay away from blah% operators in practice
  • % are mainly for backwards compatibility with previous nim versions
• generally procs that work for strings work for chars
• generally strings can use any seq proc for manipulation
• similar to js isNull, use blah.isNil, or blah != nil
  • however you cannot if blah not nil: ...

links

• wide strings
• dollars

string

• value semantics
• are really just seqchar|byte except for the terminating nullbyte 0
• ^0 terminated so nim strings can be converted to a cstring without a copy
• compared using lexicographical order
• to intrepret unicode, you need to import the unicode module
• addQuoted escapes and quotes y then appends to x in place
• addEscapedChar escapes y (backslashes everything) then appends to x

char

• 8 bit char (unsigned, basically an alias for uint8)
• always 1 byte so cant represent most UTF-8 chars
• single ASCII characters
• enclosed in single quotes

byte

• alias for uint8 that is 8 bits wide
• if dealing with binary blobs, prefer seqbyte > string,
• if dealing with binary data, prefer seqchar|uint8
• ByteAddress alias for int, used for converting pointers to itneger addresses

bool

• only true & false (i.e. not 1/0) evaluate to bool
  • but its an enum, so 0=false, 1=true
  • off/on are aliases for true/false
• if and while conditions must be of type bool

boolean procs

• != == etc
• is
• isnot
• not
• notin
• and
• or
• in
• xor
• of

comparisons

• cmp two numbers/strings

string/char procs

• len(s) Return the length of a string
• chr(i) Convert an int in the range 0..255 to a character
• ord(c) Return int value of a character
• a & b Concatenate two strings
• s.add(c) Add character to the string
• $ Convert various types to string
• substr
• find returns index of char in string
• contains true/false

numbers

number procs

• and Bitwise and &
• ashr Arithmetic shift right
• abs value of some number
• div Integer division
• high(NumberType)
• Inf == high(int)
• low(NumberType)
• max(x,y)
• min(x,y)
• mod Integer modulo (remainder) %
• not Bitwise not (complement) ~
• or Bitwise or |
• Positive range1..high(int)
• shl Shift left <<
• shr Shift right >>
• toBiggestFloat
• toBiggestInt
• toFloat Convert an integer into a float
• toInt Convert floating-point number into an int
• xor Bitwise xor ^

number types

• BackwardsIndex alias for distinct int, see range docs
• BiggestInt alias for int64
• BiggestUInt alias uinty64
• float (alias for float64|BiggestFloat) === processors fastest type,
• float32,64
• Inf float64 infinity
• int (signed), 32bit/64bit os arch, platform word size, bitwidth depends on os arch
• int8,16,32,64
• NaN not a number
• Natural alias for range[0, ..high(int): useful for documentation/debugging/guards
• NegInf !inf
• PFloat32 alias for ptr float32
• PFloat64 alias for ptr float64
• PInt32 alias for ptr int32
• Pint64 alias for ptrint64
• Positive alias for range[1, ..high(int): useful for documentation/debugging/guards
• SomeFloat matches all float types
• SomeInteger matches SomeSigned|UnsignedInt
• SomeNumber matches SomeInteger|Float
• SomeSignedInt matches all signed integer types
• SomeUnsignedInt matches all unsigned integer types
• uint: unsigned integers, 32/64 bit depending on system,
• uint8,16,32,64

reference types

• ptr generic untraced pointer
• ref generic traced pointer

reference procs

• addr Take the address of a memory location
• owned
• unown
• disarm

misc types

• typedesc meta type denoting a type description, e.g. in a proc signature to pass in a type
• nil
• void
• auto for type inference