Safe Haskell and Haskell Platform: near-term tactics

[Simon Marlow]

On 16/07/12 18:32, Gregory Collins wrote:

> This slightly underestimates the amount of work required. Each package's
> api must be carefully audited for unsafe functions, you can't just slap
> a "trustworthy" on everything and call it a day. If any legitimately
> unsafe functions are found, the APIs need to be separated out into safe
> and unsafe modules, and the old modules must go through a deprecation cycle.
>
> Also, because I figured this might be your answer, I want to make it
> clear that "emailing package authors and demanding that they do the
> legwork" is the unfair scenario that I thought we should avoid. If
> advocates of Safe Haskell want to insist that all of the platform
> packages are Safe Haskell-clean, then I think the onus is on them to
> provide patches. Not to mention, a large percentage of the platform
> packages are maintained by the libraries community at large, there's no
> specific maintainer to harass.

I agree, we should send patches to maintainers.

In fact, a lot of the packages are maintained by GHC now, following the 
last upheaval of the library submissions process.

http://www.haskell.org/haskellwiki/Library_submissions

> A quick audit of platform packages:
>
>   * Data.Text.Array is marked as "Safe-Inferred", but exports the
>     following function:
>
>     unsafeIndex :: Array -> Int -> Word16

I think you're seeing the Haddock bug there.  It should be Unsafe.

Cheers,
	Simon