Thanks for the confident boost that its not only me that stumbled upon this
However, I fail to see how the files that you have added to the "*_FILES" variables actually get copied to the staging area when builddir != srcdir hold (and cleaned up later on) ? Would you mind clarify? Cheers On 2014-08-05 18:58, Russ Allbery wrote: > Johan Persson <joh...@aditus.nu> writes: > >> The (non) solutions I have thought of (and rejected) are: > >> 1. Copy the resources into the staging path (with an extra recipe in the >> makefile). Not a good idea since this gets complicated since the movement >> should not be made when doing a build in the original tree but only for a >> staged build. Furthermore, "distclean" will give an error if these files are >> not manually removed again with another recipe and only when doing a staged >> build > > For what it's worth, this is what I do for some similar problems around > building Perl and other language bindings. It's annoying, but once you > get the system set up, it's mostly a matter of remembering to add new > files to the right variable in Makefile.am. > > If you want to see a complex and complete example of this, see the > Makefile.am for remctl: > > http://git.eyrie.org/?p=kerberos/remctl.git;a=blob;f=Makefile.am;h=50d8f808295fba2f0d9ae3490fe5611f077777e3;hb=HEAD > [1] > > The important bits are the *_FILES variables at the top (and > PERL_DIRECTORIES), the clean-local, distclean-local, and check-local > rules, and all the rules at the bottom of the file. Note that this is > much more complex than your problem, since it also has to integrate with > the native build systems of Perl, Python, PHP, and Ruby. Links: ------ [1] http://git.eyrie.org/?p=kerberos/remctl.git;a=blob;f=Makefile.am;h=50d8f808295fba2f0d9ae3490fe5611f077777e3;hb=HEAD
