patch applied (cabal): Also respect the max. command line size in Makefile driven builds

[Duncan Coutts]

On Wed, 2008-10-08 at 19:26 +0200, Clemens Fruhwirth wrote:
> At Wed, 08 Oct 2008 18:34:41 +0200,
> Christian Maeder wrote:
> > 
> > Clemens Fruhwirth wrote:
> > > Wed Oct  8 08:44:10 PDT 2008  Clemens Fruhwirth <clemens at endorphin.org>
> > >   * Also respect the max. command line size in Makefile driven builds
> > > 
> > >     M ./Distribution/Simple/GHC.hs -7 +13
> > >     M ./Distribution/Simple/GHC/Makefile.hs -1 +1
> > >     M ./Distribution/Simple/GHC/Makefile.in -1 +1
> > > 
> > > View patch online:
> > > http://darcs.haskell.org/cabal/_darcs/patches/20081008154410-ed0c4-33d531c4ead53107e5447792adc7f251dab605d2.gz
> > > 
> > 
> > I think passing -s to xargs in Distribution/Simple/GHC/Makefile.in is a
> > bad idea (as long as max-cmd-line-size is not determined) and was
> > removed some time ago!
> 
> What was the rational? It seems to break any reasonably large package
> built with split-objs under Windows.
> 
> And why isn't the correct solution to this to determine the max
> command line size?

I thought that xargs knew how big the command line args should be, even
on windows. You're saying that without specifying -s, xargs on windows
still fails with too long command line args?

Duncan