This module exports a pure interface to the MPFR library functions. Functions return new MPFR structures instead of modifying existing ones and so all functions which produce a new MPFR structure take one more parameter than their original C counterparts. This parameter, Precision, is the precision of the resulting MPFR.

This is naturally slower than modifying in-place, especially when dealing with lower precisions, so a "mutable" interface is provided in Data.Number.MPFR.Mutable module.

Naming conventions

• functions ending with _ (underscore) usually return a pair (MPFR, Int), where Int is a return value of a corresponding mpfr_ function. See the MPFR manual for a description of return values.
• the same functions without the _ return just the MPFR.
• mpfr_ prefix in functions is removed
• _ui and ui_ in function becomes w (stands for Word). For example mpfr_sub_ui becomes subw and mpfr_ui_sub becomes wsub.
• si_ and _si in functions becomes i (stands for Int). For example mpfr_sub_si becomes subi and mpfr_si_sub becomes isub.
• comparison functions which have _p appended loose it. For example mpfr_less_p becomes less.

Instances

Eq

• NaN /= NaN,
• Infinity = Infinity,
• -Infinity = -Infinity
• otherwise normal comparison

Ord

• compare NaN _ = GT
• compare _ NaN = GT
• infinity < _ = False
• -infinity > _ = False
• NaN [<,>,>=,<=] _ = False

This mimics the behaviour of built in Haskell Float and Double.

If you need instances of numeric typeclasses import one of the Data.Number.MPFR.Instances.* modules.