Symbol Names

Since Haskell allows many symbols in constructor and variable names that C compilers or assembly might not allow (e.g. :, %, #) these have to be encoded using z-encoding. The encoding is as follows. See compiler/GHC/Utils/Encoding.hs and $1535 for encoding/decoding implementations.

Tuples

Decoded	Encoded	Comment
()	Z0T	Unit / 0-tuple
		There is no Z1T
(,)	Z2T	2-tuple
(,,)	Z3T	3-tuple
...		And so on

Unboxed Tuples

Decoded	Encoded	Comment
		There is no Z0H
(# #)	Z1H	unboxed 1-tuple (note the space)
(#,#)	Z2H	unboxed 2-tuple
(#,,#)	Z3H	unboxed 3-tuple
...		And so on

Alphanumeric Characters

Decoded	Encoded	Comment
a-y, A-Y, 0-9	a-y, A-Y, 0-9	Regular letters don't need escape sequences
z, Z	zz, ZZ	'Z' and 'z' must be escaped

Constructor Characters

Decoded	Encoded	Comment
(	ZL	Left
)	ZR	Right
[	ZM	'M' before 'N' in []
]	ZN
:	ZC	Colon

Variable Characters

Decoded	Encoded	Mnemonic
&	za	Ampersand
|	zb	Bar
^	zc	Caret
$	zd	Dollar
=	ze	Equals
>	zg	Greater than
#	zh	Hash
.	zi	The dot of the 'i'
<	zl	Less than
-	zm	Minus
!	zn	Not
+	zp	Plus
'	zq	Quote
\	zr	Reverse slash
/	zs	Slash
*	zt	Times sign
_	zu	Underscore
%	zv	(TODO I don't know what the mnemonic for this one is. Perhaps relatiVe or diVide?)

Other

Any other character is encoded as a 'z' followed by its hex code (lower case, variable length) followed by 'U'. If the hex code starts with 'a', 'b, 'c', 'd', 'e' or 'f', then an extra '0' is placed before the hex code to avoid conflicts with the other escape characters.

Examples

Before	After
Trak	Trak
foo_wib	foozuwib
>	zg
>1	zg1
foo#	foozh
foo##	foozhzh
foo##h	foozhzhh
fooZ	fooZZ
:+	ZCzp
()	Z0T
(,,,,)	Z5T
(# #)	Z1H
(#,,,,#)	Z5H