I've just used haskell as calculator expecting to get some cryptic error message about ambiguity of the type instance of the numeric literal. But got none.

For example in ghci

> :t 1
1 :: Num p => p
> print 1
1

It seems crazy that I expect code like print 1 to break xD. But here ghci needs to use some implementation of Show class. Also the 1 is represented in some way in runtime. What representation is used and why doesn't it ask me to specify the type?

Simmilar example:

> 7*15 2
107

The expression has general Num a => a type but is calculated and for that some type needs to be chosen. Is it Integer? This behaviour of course makes sense but how does it work?

Typing e.g. print $ 5 ^ 5 ^ 5 (outside of Int range) prints the correct number so I guess that it is Integer rather than Int.

I haven't found any general Num a => Show a instance, rather the implementations for types like Int are specific to them.

I've tested with ghc in versions 8.8.4, 7.10.3 and 9.0.2. Thanks in advance!

CodePudding user response：

Haskell has a type defaulting mechanism to avoid having to specify the type of every numeric literal you type. The default type for Num is Integer, while for Fractional (literals with a decimal point) it is Double. You can change these defaults using a top level default declaration. This article explains the defaulting rules

CodePudding user response：

The expression has general Num a => a type but is calculated and for that some type needs to be chosen. Is it Integer? This behaviour of course makes sense but how does it work?

Haskell works with type defaulting and the default for a Num a => a type is indeed Integer. Since Integer is also a member of the Show typeclass, it will thus still use Integer and print $ 5 ^ 5 ^ 5 will thus indeed use the Integer instance of Show to print the value.

CodePudding user response：

The default default is (Integer, Double). More info in the Report. You may also want to read about ExtendedDefaultRules; notably this is on by default in ghci.