xcode 5 psa info on haskell platform page

Darin Morrison darinmorrison at gmail.com
Tue Oct 22 12:28:22 BST 2013


Hi All,

I don't really know what to say or do about the platform stuff on
Mavericks. At the moment I'm not able to investigate that but if someone in
charge (with write access to the docs) wants to contact me, I can provide a
blurb or something that might help OS X users on the bleeding edge.

As Carter pointed out, the best bet currently is to install via Homebrew:

———

$ xcode-select --install
$ ruby -e "$(curl -fsSL https://raw.github.com/mxcl/homebrew/go)"

# then, a choice to either use the tap I maintain at darinmorrison/haskell

$ brew tap darinmorrison/haskell
$ brew install apple-gcc42 darinmorrison/haskell/ghc
darinmorrison/haskell/cabal-install

# or just to use the mainline ghc and cabal install formula

$ brew install apple-gcc42 ghc cabal-install

———

Either of those should work. The mainline ghc formula does include the GHC
patches for compiling on 10.9 (which are now also upstream in HEAD).

The advantage to using the branch I maintain is that it's likely to be
updated with Haskell-specific fixes/releases sooner than the mainline. And
I also try to come down on the side of reasonable Haskell conventions
rather than Homebrew conventions if there is a conflict. I'm also willing
to try and help troubleshoot if there's a Haskell specific problem and
someone contacts me or creates an issue on github.

The advantage to using mainline is that they are using pre-built binaries,
so it's faster. I haven't gotten around to that yet (there's also some
question about whether to try and use their bottling system or the GHC
bindists).

The back-ported clang patches for GHC 7.6.3 will probably show up in my
branch a little earlier (although they'll need to be vetted by Austin and
co) too, although any sort of fixes or improvements like that will be
submitted to the mainline also.

I realize this doesn't help folks who don't want or are unable to use
Homebrew. And it doesn't directly address the Platform issue either.
(There's a formula for that too but I think there are problems with it
currently—in any case, I haven't tested it). So we still need to get some
updated installers and docs and stuff for those folks in the standard way.

I'd like to help with that too if I can. But like I said, I've already got
enough on my plate for the moment, so it'd need to be initiated by someone
else since I don't have time to investigate just now. Feel free to email me
or ping me on #ghc (actually, please try any means of contact _other_ than
email).

In the meantime, here's what I'm focusing on:

1. Making a Mavericks development machine available (to Austin and co.)
(which has already taken me down several unrelated rabbit holes…)

2. Finishing up the back-ported clang compatibility patches for 7.6.3
(these will show up at github.com/darinmorrison/ghc/tree/ghc-{clang,
mavericks, clang+mavericks} hopefully in the next day or so)

3. Updating the GHC and cabal-install homebrew formulae

4. Creating and hosting updated bindists (and also providing updated
binaries for older OS X releases)

I _should_ have a lot more free time this week than the last few, so
hopefully this stuff will finally get sorted.

Cheers,
Darin

On Mon, Oct 21, 2013 at 4:01 PM, Carter Schonwald <
carter.schonwald at gmail.com> wrote:

> that said, this doesn't handle mavericks.
>
> darin morrison's brew tap  is currently the only good way to get mavericks
> support working well (without lots of end user intervention), though he's
> collabing with austin seipp on getting a ghc 7.6.4 bugfix sorted out. His
> brew formulae should work though in the mean time
>
>
> On Mon, Oct 21, 2013 at 5:07 PM, Carter Schonwald <
> carter.schonwald at gmail.com> wrote:
>
>> Hey all
>>
>> http://www.haskell.org/platform/mac.html
>>
>> currently lacks any prominent explanation / hint / reference to the xcode
>> 5 issue and the (simple) possible work arounds.
>> http://www.haskell.org/pipermail/haskell-cafe/2013-September/110320.html
>>
>>
>> I really really really really think we need to fix that. Who's got the
>> right acls etc to add such a notice?
>>
>>
>> considering IRC regulars are a small fraction of haskellers (however
>> large that may be overall), not having this info in a discoverable place
>> for all the folks who just use haskell platform (possibly for the first
>> time) is a HUGE problem.  :(
>>
>>
>> thanks
>>
>>
>> -Carter
>>
>
>


-- 
Darin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://projects.haskell.org/pipermail/haskell-platform/attachments/20131022/6097fd1c/attachment.htm>


More information about the Haskell-platform mailing list