Successful build of GHC HEAD on OS X 10.9 Mavericks

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Successful build of GHC HEAD on OS X 10.9 Mavericks

Luke Iannini
Hi guys,

Just wanted to report on the steps/hacks to a successful build of GHC HEAD
on OS X 10.9 thanks to help from Carter Schonwald and Austin Seipp.

1) I installed LLVM 3.4
brew install llvm --with-clang --HEAD

2) I used this build.mk:
HADDOCK_DOCS       = NO
BUILD_DOCBOOK_HTML = NO
BUILD_DOCBOOK_PS   = NO
BUILD_DOCBOOK_PDF  = NO
SPLIT_OBJS         = NO
INTEGER_LIBRARY    = integer-simple

3) Austin let me know I needed to pass "-x assembler-with-cpp" to clang
when in preprocessor mode, so I wrote a wrapper clang.hs, compiled it, put
it in my path, and changed /usr/local/lib/ghc-7.6.2/settings to use it. (it
also passes some flags Carter suggested, along with extra include flags,
which were necessary as of a couple weeks ago but might have been fixed in
the newer DPs)
https://gist.github.com/lukexi/e11668ced04524564a8b

4) I used
./configure --with-gcc=clang
(where clang is the wrapper script above)

5) "inplace/bin/mkdirhier" bootstrapping/. failed mysteriously, so I
manually created bootstrapping/ in the ghc build dir.

6) DTrace failed with an "unsupported compiler" error during compilation,
so I killed that section in configure.ac
# if test -n "$DtraceCmd"; then
#   if test "x$TargetOS_CPP-$TargetVendor_CPP" = "xdarwin-apple" -o
"x$TargetOS_CPP-$TargetVendor_CPP" = "xsolaris2-unknown"; then
#     HaveDtrace=YES
#   fi
# fi

7) I got the error:
libraries/integer-simple/GHC/Integer/Type.hs:836:50:
    Not in scope: '.'

libraries/integer-simple/GHC/Integer/Type.hs:847:67:
    Not in scope: '.'

I hacked through this by manually find-and-replacing WORD_SIZE_IN_BITS with
64 in Type.hs.

make && make install then ran to completion and some simple test programs
seem to compile and run fine.

Cheers
Luke
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130809/1c73953c/attachment.htm>

Reply | Threaded
Open this post in threaded view
|

Successful build of GHC HEAD on OS X 10.9 Mavericks

Dr. ERDI Gergo
On Fri, 9 Aug 2013, Luke Iannini wrote:

> Hi guys,
> Just wanted to report on the steps/hacks to a successful build of GHC HEAD on OS X
> 10.9 thanks to help from Carter Schonwald and Austin Seipp.

HEAD as of 3e598fe0b? I wasn't able to build that one because the Core
lint pass fails with the new Role checker.

--

   .--= ULLA! =-----------------.
    \     http://gergo.erdi.hu   \
     `---= gergo at erdi.hu =-------'
Don't anthropomorphize computers. They hate that.



Reply | Threaded
Open this post in threaded view
|

Successful build of GHC HEAD on OS X 10.9 Mavericks

Luke Iannini
Yes, I'm at 3e598fe0b, couldn't tell you why it worked for me though : ).


On Fri, Aug 9, 2013 at 3:33 AM, Dr. ERDI Gergo <gergo at erdi.hu> wrote:

> On Fri, 9 Aug 2013, Luke Iannini wrote:
>
>  Hi guys,
>> Just wanted to report on the steps/hacks to a successful build of GHC
>> HEAD on OS X
>> 10.9 thanks to help from Carter Schonwald and Austin Seipp.
>>
>
> HEAD as of 3e598fe0b? I wasn't able to build that one because the Core
> lint pass fails with the new Role checker.
>
> --
>
>   .--= ULLA! =-----------------.
>    \     http://gergo.erdi.hu   \
>     `---= gergo at erdi.hu =-------'
> Don't anthropomorphize computers. They hate that.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130809/6c085033/attachment.htm>

Reply | Threaded
Open this post in threaded view
|

Successful build of GHC HEAD on OS X 10.9 Mavericks

Richard Eisenberg-2
In reply to this post by Dr. ERDI Gergo
I'm not aware of any outstanding bugs in the role checker that would cause a core lint failure (or any other problems), so if you're seeing one, that sounds like new news. How did you run into this?

Thanks,
Richard

On Aug 9, 2013, at 3:33 AM, Dr. ERDI Gergo wrote:

> On Fri, 9 Aug 2013, Luke Iannini wrote:
>
>> Hi guys,
>> Just wanted to report on the steps/hacks to a successful build of GHC HEAD on OS X
>> 10.9 thanks to help from Carter Schonwald and Austin Seipp.
>
> HEAD as of 3e598fe0b? I wasn't able to build that one because the Core lint pass fails with the new Role checker.
>
> --
>
>  .--= ULLA! =-----------------.
>   \     http://gergo.erdi.hu   \
>    `---= gergo at erdi.hu =-------'
> Don't anthropomorphize computers. They hate that.
>
> _______________________________________________
> ghc-devs mailing list
> ghc-devs at haskell.org
> http://www.haskell.org/mailman/listinfo/ghc-devs




Reply | Threaded
Open this post in threaded view
|

Successful build of GHC HEAD on OS X 10.9 Mavericks

Dr. ERDI Gergo
Hi,

I've posted the details here:
http://www.haskell.org/pipermail/ghc-devs/2013-August/001907.html

It fails in phase 2 for me.

Bye,
Gergo
On Aug 12, 2013 9:47 AM, "Richard Eisenberg" <eir at cis.upenn.edu> wrote:

> I'm not aware of any outstanding bugs in the role checker that would cause
> a core lint failure (or any other problems), so if you're seeing one, that
> sounds like new news. How did you run into this?
>
> Thanks,
> Richard
>
> On Aug 9, 2013, at 3:33 AM, Dr. ERDI Gergo wrote:
>
> > On Fri, 9 Aug 2013, Luke Iannini wrote:
> >
> >> Hi guys,
> >> Just wanted to report on the steps/hacks to a successful build of GHC
> HEAD on OS X
> >> 10.9 thanks to help from Carter Schonwald and Austin Seipp.
> >
> > HEAD as of 3e598fe0b? I wasn't able to build that one because the Core
> lint pass fails with the new Role checker.
> >
> > --
> >
> >  .--= ULLA! =-----------------.
> >   \     http://gergo.erdi.hu   \
> >    `---= gergo at erdi.hu =-------'
> > Don't anthropomorphize computers. They hate that.
> >
> > _______________________________________________
> > ghc-devs mailing list
> > ghc-devs at haskell.org
> > http://www.haskell.org/mailman/listinfo/ghc-devs
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.haskell.org/pipermail/ghc-devs/attachments/20130812/9a1e8b14/attachment.htm>