From mboxrd@z Thu Jan 1 00:00:00 1970 From: ludo@gnu.org (Ludovic =?utf-8?Q?Court=C3=A8s?=) Subject: Re: Overriding PostGIS install directories Date: Wed, 24 Jan 2018 15:28:54 +0100 Message-ID: <87wp07z5y1.fsf@gnu.org> References: <87r2qm5t7o.fsf@sturm.com.au> <8737325679.fsf@gnu.org> <87po65nzdi.fsf@sturm.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:4830:134:3::10]:52698) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1eeM2r-0003pi-Jq for guix-devel@gnu.org; Wed, 24 Jan 2018 09:28:58 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1eeM2q-0002oX-OB for guix-devel@gnu.org; Wed, 24 Jan 2018 09:28:57 -0500 Received: from hera.aquilenet.fr ([185.233.100.1]:36856) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1eeM2q-0002mf-IS for guix-devel@gnu.org; Wed, 24 Jan 2018 09:28:56 -0500 In-Reply-To: <87po65nzdi.fsf@sturm.com.au> (Ben Sturmfels's message of "Sat, 20 Jan 2018 11:32:41 +1100") List-Id: "Development of GNU Guix and the GNU System distribution." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org Sender: "Guix-devel" To: Ben Sturmfels Cc: Guix-devel Hi, Ben Sturmfels skribis: > On Fri, 19 Jan 2018, Ludovic Court=C3=A8s wrote: > >> However, I=E2=80=99d recommend simply patching =E2=80=98configure=E2=80= =99 itself in a snippet, >> using =E2=80=98substitute*=E2=80=99. That way, you won=E2=80=99t have t= o add these dependencies >> and extra phase. > > Thanks, substitute* is much neater and changing configure directly is > working. > > I've just discovered though that the build includes makefile templates > from the input postgres package, eg > ...-postgresql-10.1/lib/pgxs/src/makefiles/pgxs.mk and parents. These > makefiles are created from a separate run of `pg_config --sharedir` when > building postgresql. > > I wonder whether it would be simpler to mock the "pg_config" program to > return the paths I want, rather than patch out all the calls to it. Is > that possible to do for a package and its inputs? What do you think? I guess it all depends on the number of =E2=80=98pg_config=E2=80=99 invocat= ions that need to be patch, and whether each one needs different treatment. We should minimize complexity. Ludo=E2=80=99.