BSD-2-Clause licensed by Christiaan Baaij
This version can be pinned in stack with:ghc-typelits-extra-0.4.3@sha256:5d14a15d075406e80b008a0fad24d3f56a55d7534f471d1d7627a14d9b3e2678,4901
Module documentation for 0.4.3
ghc-typelits-extra
Extra type-level operations on GHC.TypeLits.Nat and a custom solver implemented
as a GHC type-checker plugin:
GHC.TypeLits.Extra.Max
: type-level max
GHC.TypeLits.Extra.Min
: type-level min
GHC.TypeLits.Extra.Div
: type-level div
GHC.TypeLits.Extra.Mod
: type-level mod
GHC.TypeLits.Extra.FLog
: type-level equivalent of integerLogBase#
.i.e. the exact integer equivalent to “floor (logBase x y)
”
GHC.TypeLits.Extra.CLog
: type-level equivalent of the ceiling of integerLogBase#
.i.e. the exact integer equivalent to “ceiling (logBase x y)
”
- ‘GHC.TypeLits.Extra.Log’: type-level equivalent of <https://hackage.haskell.org/package/integer-gmp/docs/GHC-Integer-Logarithms.html#v:integerLogBase-35- integerLogBase#>
where the operation only reduces when “
floor (logBase b x) ~ ceiling (logBase b x)
”
GHC.TypeLits.Extra.GCD
: a type-level gcd
GHC.TypeLits.Extra.LCM
: a type-level lcm
0.4.3 June 18th 2021
- Add support for GHC 9.2.0.20210422
0.4.2 January 1st 2021
- Add support for GHC 9.0.1-rc1
0.4.1 November 10 2020
- Reduce
n <=? Max (n + p) p
to True
0.4 March 9 2020
Max
short-circuits on zero, but is stuckness preserving. i.e. Max (0-1) 0
reduces to (0-1)
- Reduce inside arithmetic equations. e.g.
1 + a ~ Max 0 a + CLog 2 2
0.3.3 February 6th 2020
- Add support for GHC 8.10.1-alpha2
0.3.2 January 18th 2020
0.3.1 August 26th 2019
- Reduce
a <=? Max a b
to True
- Reduce
n ~ (Max a b) => a <=? n
to True
- Prove
Max (1 + n) 1 ~ (n+1)
0.3 September 14th 2018
- Move
KnownNat2
instances for GHC 8.4’s Div
and Mod
from ghc-typelits-extra
to ghc-typelits-knownnat
0.2.6 Julty 10th 2018
- Add support for GHC-8.6.1-alpha1
0.2.5 May 9th 2018
- Add support for ghc-typelits-natnormalise-0.6
0.2.4 January 4th 2018
- Add support for GHC-8.4.1-alpha1
0.2.3 May 15th 2017
- Support GHC 8.2
Max
, Min
, GCD
, and LCM
now have a commutativity property #9
- Reduce
GCD 0 x
to x
#9
- Reduce
GCD 1 x
to 1
#9
- Reduce
GCD x x
to x
#9
- Reduce
LCM 0 x
to 0
#9
- Reduce
LCM 1 x
to x
#9
- Reduce
LCM x x
to x
#9
- Reduce
Max (0-1) 0
to 0
#10
- Reduce
Min (0-1) 0
to 0 - 1
#10
- Fixes bugs:
- Solver turns LCM into GCD #8
- Solver turns Max into Min
0.2.2 January 15th 2017
- Reduce
Min n (n+1)
to n
- Reduce
Max n (n+1)
to n+1
- Reduce cases like
1 <=? Div 18 6
to True
- Add a type-level division that rounds up:
type DivRU n d = Div (n + (d - 1)) d
- Add a type-level
divMod
: DivMod :: Nat -> Nat -> '(Nat, Nat)
0.2.1 September 29th 2016
- Reduce
Max n n
to n
- Reduce
Min n n
to n
0.2 August 19th 2016
- New type-level operations:
Max
: type-level max
Min
: type-level min
Div
: type-level div
Mod
: type-level mod
FLog
: floor of logBase
Log
: exact integer logBase (i.e. where floor (logBase b x) ~ ceiling (logBase b x)
holds)
LCM
: type-level lcm
- Fixes bugs:
CLog b 1
doesn’t reduce to 0
0.1.3 July 19th 2016
- Fixes bugs:
- Rounding error in
CLog
calculation
0.1.2 July 8th 2016
- Solve KnownNat constraints over CLog and GCD, i.e., KnownNat (CLog 2 4)
0.1.1 January 20th 2016
0.1 October 21st 2015