tfausak/witch
1
1
Fork 0
mirror of https://github.com/tfausak/witch.git synced 2024-11-22 14:58:13 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #44 from tfausak/gh-43-laws

Browse Source 
Add some laws
This commit is contained in:
Taylor Fausak 2021-12-09 22:30:22 -05:00 committed by GitHub
commit 9d5e13cec3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 25 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
src/lib/Witch.hs
View File

@ -184,12 +184,37 @@ module Witch
-- - You should not have both a @From@ instance and a @TryFrom@ instance for
-- the same pair of types.
--
-- - If you have a @From@ or @TryFrom@ instance for a pair of types, then
-- you should probably have a @From@ or @TryFrom@ instance for the same
-- pair of types but in the opposite direction. In other words if you have
-- @From a b@ then you should have @From b a@ or @TryFrom b a@.
--
-- In general if @s@ /is/ a @t@, then you should add a 'Witch.From.From'
-- instance for it. But if @s@ merely /can be/ a @t@, then you could add a
-- 'Witch.TryFrom.TryFrom' instance for it. And if it is technically
-- possible to convert from @s@ to @t@ but there are a lot of caveats, you
-- probably should not write any instances at all.
-- ** Laws
-- | As the previous section notes, there aren't any cut and dried laws for
-- the @From@ and @TryFrom@ type classes. However it can be useful to
-- consider the following equations for guiding instances:
--
-- > -- same strictness
-- > seq (from @a @b x) y = seq x y
-- > seq (tryFrom @a @b x) y = seq x y
--
-- > -- round trip
-- > from @b @a (from @a @b x) = x
--
-- > -- transitive
-- > from @b @c (from @a @b x) = from @a @c x
-- > tryFrom @b @a (from @a @b x) = Right x
-- > if isRight (tryFrom @a @b x) then
-- > fmap (from @b @a) (tryFrom @a @b x) = Right x
-- > if isRight (tryFrom @a @b x) then do
-- > fmap (tryFrom @b @a) (tryFrom @a @b x) = Right (Right x)
-- ** Integral types
-- | There are a lot of types that represent various different ranges of
-- integers, and Witch may not provide the instances you want. In particular