Safe Haskell

Introduction

Safe Haskell is a Haskell language extension. It is described in detail:

In the GHC user manual:

http://www.haskell.org/ghc/docs/latest/html/users_guide/safe-haskell.html

In the Safe Haskell paper:

http://community.haskell.org/~simonmar/papers/safe-haskell.pdf

Further technical discussion of Safe Haskell is on the GHC Wiki:

http://hackage.haskell.org/trac/ghc/wiki/SafeHaskell

For a high-level overview, see this blog post: http://begriffs.com/posts/2015-05-24-safe-haskell.html.

In detail

As the Safe Haskell paper describes, it "hardens" the Haskell language by providing five properties: type safety, referential transparency, strict module encapsulation, modular reasoning and semantic consistency.

Safe Haskell is not magic, and not about catching bugs. It does not ensure that library authors who mark modules as "Trustworthy" are not lying, or incorrect. It does not ensure code inferred safe but in IO cannot perform arbitrary IO. It does ensure that untrusted code inferred to be safe will, assuming its "Trustworthy" imports are indeed "Trustworthy" obey the above five properties. As such, again assuming "Trustworthy" imports are indeed so, Safe Haskell infers that untrusted code inferred safe and not in IO can be run without fear (aside from fear of resource over-utilization/exhaustion).

Most code that most people want to write is going to be Safe Haskell by default. As Simon Marlow has pointed out,

"Normally when you use an unsafe feature, the purpose is to use it to implement a safe API - if that's the case, all you have to do is add Trustworthy to your language pragma, and the API is available to use from Safe code. 99% of Hackage should be either Safe or Trustworthy. We know that 27% is already inferred Safe (see the paper), and a lot of the rest is just waiting for other libraries to add Trustworthy where necessary."

Again, as Simon Marlow argues,

"For typical Haskell programmers, using {-# LANGUAGE Safe #-} will be like -Wall: something that is considered good practice from a hygiene point of view. If you don't need access to unsafe features, then it's better to write in the safe subset, where you have stronger guarantees. Just like -Wall, you get to choose whether to use {-# LANGUAGE Safe #-} or not."