[GHC] #11527: Pattern match translation suboptimal
GHC
ghc-devs at haskell.org
Tue Feb 2 13:10:44 UTC 2016
#11527: Pattern match translation suboptimal
-------------------------------------+-------------------------------------
Reporter: augustss | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 7.8.3
Resolution: | Keywords:
Operating System: Unknown/Multiple | Architecture:
| Unknown/Multiple
Type of failure: None/Unknown | Test Case:
Blocked By: | Blocking:
Related Tickets: | Differential Rev(s):
Wiki Page: |
-------------------------------------+-------------------------------------
Comment (by simonpj):
For `Integer` I agree we could and should do better. But what about
{{{
baz :: (Num a) => a -> Int -> Int
}}}
which is what you'll get if you don't have a type signature.
The report specifies something like this:
{{{
baz n x
| n == fromInteger 10, x == 1 = 1
| n == fromInteger 20, x == 1 = 2
| n == fromInteger 10, x == 2 = 2
...etc...
}}}
Now you might think that we can common up all those tests for `n ==
fromInteger 10` to get
{{{
baz n x
| n == fromInteger 10
= case x of
1 -> 1
2 -> 2
... etc ...
| n == fromInteger 20
= ...etc..
}}}
But what if both `n == fromInteger 10` and `n == fromInteger 20` are
''both true''? Now it's not OK to group together all the matches against
10. How annoying!
The best we can do is to do CSE on those `fromInteger` calls.
It's all very irritating but I don't see how to do better. Except for
`Integer` when we know that the `Num` instance has good properties. Is
that particular case worth improving? Perhaps, and it would not be hard
-- I can offer guidance.
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11527#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list