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

[Stefan O'Rear]

On Thu, Aug 02, 2007 at 09:08:31AM +0100, Simon Marlow wrote:
> So I wasn't aware that LD_LIBRARY_PATH takes precedence over rpath on some 
> platforms, that does present a slight problem, but maybe it's not a killer. 
>  I don't expect that most people will be using LD_LIBRARY_PATH, and the 
> build system could always check for it.  Perhaps we should have more safety 
> nets - a runtime check for the right libraries, for example.
>
> If you have a better proposal, please tell us!

If you're building stuff with an in-place GHC, you're probably
developing GHC.

If you're developing GHC, you're probably rebuilding GHC frequently.

If you're rebuilding GHC frequently, the symbols in your libraries are
probably changing frequently.

If the symbols in the libraries are changing meanings frequently, using
shared libraries is probably a Bad Idea.


If all those probablies are still probably when multiplied (and I
suspect this is the case), then the best thing to do - even if you
ignore implementation complexity! - is to disable shared library usage
in the ghc-inplaces.

As a nice side effect, this means that we don't have to worry about
RPATH-foo at all under any circumstance.

Stefan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www.haskell.org/pipermail/cvs-ghc/attachments/20070802/793751e0/attachment.bin