HEAD doesn't build (was Re: patch applied (ghc): Add hardwire-lib-paths option and make it default for inplace invocations)

Simon Marlow simonmarhaskell at gmail.com
Wed Aug 1 04:20:57 EDT 2007


Roman Leshchinskiy wrote:
> Hi Clemens,
> 
> Simon Marlow wrote:
>> Tue Jul 17 12:26:22 PDT 2007  Clemens Fruhwirth <clemens at endorphin.org>
>>   * Add hardwire-lib-paths option and make it default for inplace 
>> invocations
> 
> This patch unconditionally adds -rpath to the linker options for 
> in-place invocations. The MacOS linker doesn't understand this option 
> (and has no concept of an rpath) which means that stage2 can't be 
> linked. IIRC, this will also break Solaris as the flag is called -R 
> there. Note also that -rpath won't help you if someone sets 
> LD_LIBRARY_PATH as the latter usually has precedence. Perhaps using 
> libtool might help here?

We had a discussion on glasgow-haskell-users recently about whether to use 
libtool, and the general concensus was not:

http://www.haskell.org/pipermail/glasgow-haskell-users/2007-June/012730.html

> In any case, when you change things which are obviously system-dependent 
> could you please try to make sure that they are portable at least across 
> the major platforms. If you do not have access to some of those please 
> ask on the mailing list. I will gladly check on MacOS.

Partly my fault for sloppy reviewing, sorry about that.  I was under the 
impression that this only affected dynamic linking so it wouldn't break 
anything.

Clemens, shouldn't we only be adding -rpath when -dynamic is on?

Cheers,
	Simon



More information about the Cvs-ghc mailing list