Re: [PATCH 21/43] gnu: ghc-transformers-compat: New variables.

[Paul van der Walt]

On 2015-10-19 at 16:37, quoth Eric Bavier:
>> +  (package
>> +    (name "ghc-exceptions")
>> .....
>> +    (inputs
>> +     `(("ghc-quickcheck" ,ghc-quickcheck)))
>
> Is this still required even with '#:tests? #f'?

Turns out it's not.  Fixed.

>> +    (description
>> +     "Extensible optionally-pure exceptions")
>
> This is a fragment.  In general, I notice that the formatting style
> varies a bit accross packages.  Could you make sure that formatting, at
> least within gnu/packages/haskell.scm is mostly consistent?

Fixed in this package, will look at the rest before pushing.

>> +  (package
>> +    (name "ghc-temporary")
>> .....
>> +    (synopsis "Portable temporary file and directory support for Windows and
>> +Unix, based on code from Cabal")
>> +    (description "The functions for creating temporary files and directories
>> +in the base library are quite limited. The unixutils package contains some
>> +good ones, but they aren't portable to Windows. This library just repackages
>> +the Cabal implementations of its own temporary file and folder functions so
>> +that you can use them without linking against Cabal or depending on it being
>> +installed.")
>> +    (license bsd-3)))
>
> Again, we don't really need to mantion portability aspects.

Fixed.

>> +  (package
>> +    (name "ghc-silently")
>> +    (arguments `(#:tests? #f))  ;; circular dependency with nanospec
>> +    (inputs
>> +     `(("ghc-temporary" ,ghc-temporary)))
>
> Is this required with '#:tests? #f', and if so, does it need to be
> propagated?

You're right, it seems it was only necessary for the tests.  Removed
entirely.

>> +    (description
>> +     "Prevent or capture writing to stdout and other handles.")
>
> Could you expand this so it's not a fragment?

Done.