[GHC] #15646: ghci takes super long time to find the type of large fractional number
GHC
ghc-devs at haskell.org
Tue Sep 18 08:03:50 UTC 2018
#15646: ghci takes super long time to find the type of large fractional number
-------------------------------------+-------------------------------------
Reporter: Johannkokos | Owner:
| JulianLeviston
Type: bug | Status: new
Priority: normal | Milestone: 8.6.1
Component: GHCi | Version: 8.4.3
Resolution: | Keywords: newcomer
Operating System: Unknown/Multiple | Architecture:
Type of failure: Compile-time | Unknown/Multiple
performance bug | Test Case:
Blocked By: | Blocking:
Related Tickets: | Differential Rev(s):
Wiki Page: |
-------------------------------------+-------------------------------------
Comment (by simonpj):
> Wait, are you saying that typing 1.7e1234111111111111111111111 to take
minutes and use up more than 20G of memory (residence, not allocation!)
not surprising?
I agree: that's absurd!
HOwever, unlike Python, we can't just compute a suitable `Float`, because
literals are overloaded. The
[https://www.haskell.org/onlinereport/haskell2010/haskellch3.html#x8-240003.2
Report] does say that a floating point literal like `1e100` means
`fromRational (n % d)` where "the integers n and d are chosen so that n/d
= f".
I suspect that in computing `17 % 10000000000000000000000` we try to find
the GCD of the two before we even start with `fromRational`, and you can
see this isn't going to end well.
What to do? Probably we need a special case for `fromRational :: Rational
-> Float` (and similarly `Double`); and maybe even a special literal
representation inside GHC for `Rational`s of form `N / 10000000000000`
for some number of zeros.
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15646#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list