On Fri, 2006-04-07 at 14:03 -0500, Michael Hennebry wrote: > On Fri, 7 Apr 2006, Paul Howarth wrote: > > > Michael Hennebry wrote: > > > On Fri, 7 Apr 2006, Ralf Corsepius wrote: > > > > > >> On Thu, 2006-04-06 at 21:22 -0500, Michael Hennebry wrote: > > >>> One does not have to become root. > > >>> If one does want the software > > >>> installed as system software, > > >>> do some testing first. > > >>> Edit Makefile, > > >>> removing any .SILENT targets. > > >>> As an ordinary user, > > >>> make -n prefix=/opt 'CP=cp -i' install 2>&1 | tee makeno.out > > >> Same as above. This won't work in most cases. > > > > > > In this case, I'm not sure what "won't work" means. > > > Do you mean that it won't give one a good idea whether > > > something one doesn't want clobbered will get clobbered? > > > > Possibly. A large proportion of projects use "install" rather than "cp" > > in their Makefile "install" targets. > > Is there a reason that the install commands wouldn't show up in > the output or that they would be performed despite the -n flag? > The 'CP=cp -i' would obviously be useless. It was the 'CP=cp -i' that I was referring to. The "-n" should indeed prevent any commands from actually being run. However, the volume of output you get from an autotools-generated Makefile with "-n" would make it easy to miss something important, particularly given that some of the output could include complex if/then type arrangements. Paul.
