From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ludovic =?UTF-8?Q?Court=C3=A8s?= Subject: bug#38336: =?UTF-8?Q?=E2=80=98wine64-staging=E2=80=99?= is very expensive to build Date: Sat, 23 Nov 2019 00:12:07 +0100 Message-ID: <87r21z7bg8.fsf@gnu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: Received: from eggs.gnu.org ([2001:470:142:3::10]:48418) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1iYI6t-0000Wd-78 for bug-guix@gnu.org; Fri, 22 Nov 2019 18:13:08 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1iYI6s-0000HQ-4n for bug-guix@gnu.org; Fri, 22 Nov 2019 18:13:07 -0500 Received: from debbugs.gnu.org ([209.51.188.43]:46879) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1iYI6s-0000HM-1a for bug-guix@gnu.org; Fri, 22 Nov 2019 18:13:06 -0500 Sender: "Debbugs-submit" Resent-Message-ID: Received: from eggs.gnu.org ([2001:470:142:3::10]:48244) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1iYI5x-0000T0-DG for bug-Guix@gnu.org; Fri, 22 Nov 2019 18:12:10 -0500 Received: from fencepost.gnu.org ([2001:470:142:3::e]:47973) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1iYI5x-0008PA-9x for bug-Guix@gnu.org; Fri, 22 Nov 2019 18:12:09 -0500 Received: from [2a01:e0a:1d:7270:af76:b9b:ca24:c465] (port=60130 helo=ribbon) by fencepost.gnu.org with esmtpsa (TLS1.2:RSA_AES_256_CBC_SHA1:256) (Exim 4.82) (envelope-from ) id 1iYI5w-0001vH-R6 for bug-Guix@gnu.org; Fri, 22 Nov 2019 18:12:09 -0500 List-Id: Bug reports for GNU Guix List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: bug-guix-bounces+gcggb-bug-guix=m.gmane.org@gnu.org Sender: "bug-Guix" To: bug-Guix@gnu.org Cc: Rutger Helling Hello, I noticed that =E2=80=98wine64-staging=E2=80=99 is our most expensive packa= ge to build. Initially I was surprised because it only has a couple of nodes more in its object than =E2=80=98wine-staging=E2=80=99: --8<---------------cut here---------------start------------->8--- ludo@ribbon ~/src/guix$ guix graph wine-staging |grep 'label =3D ' |wc -l 509 ludo@ribbon ~/src/guix$ guix graph wine64-staging |grep 'label =3D ' |wc -l 511 --8<---------------cut here---------------end--------------->8--- However, that single additional node leads to the duplication of the whole derivation graph on x86_64-linux: --8<---------------cut here---------------start------------->8--- ludo@ribbon ~/src/guix$ guix graph -t derivation wine-staging |grep 'label = =3D ' |wc -l 2738 ludo@ribbon ~/src/guix$ guix graph -t derivation wine64-staging |grep 'labe= l =3D ' |wc -l 4598 --8<---------------cut here---------------end--------------->8--- This is because =E2=80=98wine-staging=E2=80=99 has a hard-coded =E2=80=98#:= system "i686-linux"=E2=80=99, whereas =E2=80=98wine64-staging=E2=80=99 is (unsurprisingly :-)) built on x= 86_64-linux. (The same problem happens with =E2=80=98wine=E2=80=99 vs. =E2=80=98wine64= =E2=80=99.) Likewise, =E2=80=98guix size wine64=E2=80=99 shows that every dependency ap= pears twice (one 32-bit, one 64-bit), and thus the total size is twice that of =E2=80=98wine=E2=80=99. Rutger, is there something that can be done to avoid this? Apparently only JSON files are copied from =E2=80=98wine-staging=E2=80=99 into =E2=80= =98wine64-staging=E2=80=99, but maybe they refer to 32-bit shared libraries or something? Thanks, Ludo=E2=80=99.