From mboxrd@z Thu Jan 1 00:00:00 1970 From: ludo@gnu.org (Ludovic =?utf-8?Q?Court=C3=A8s?=) Subject: Graft slowness fixed Date: Mon, 16 Jan 2017 22:18:55 +0100 Message-ID: <87tw8y66v4.fsf@gnu.org> References: <1484159585-12289-1-git-send-email-thomas.danckaert@gmail.com> 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]:52013) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cTEg9-00025D-SR for guix-devel@gnu.org; Mon, 16 Jan 2017 16:19:02 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cTEg6-0004Ed-LV for guix-devel@gnu.org; Mon, 16 Jan 2017 16:19:01 -0500 In-Reply-To: <1484159585-12289-1-git-send-email-thomas.danckaert@gmail.com> (Thomas Danckaert's message of "Wed, 11 Jan 2017 19:33:01 +0100") 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: Thomas Danckaert Cc: guix-devel@gnu.org Hi Thomas, Thomas Danckaert skribis: > There is one issue: when I apply this patch series on current master (aft= er > commit e10872c gnu: gpgme: Patch CMake files), `guix build kdevelop` seem= s to > get stuck after completing the build (at least it doesn't return after ru= nning > at ~100%cpu for over 40minutes, with no output). When I apply the patch > series on commit 4d0a3d8 (system: grub: Load relevant EFI modules on EFI > systems), I don't have this problem. I'm not sure if this is an issue ju= st on > my system, or not (I upgraded guix-daemon to lastest master, but the prob= lem > remains). Hints on how to investigate this are welcome! You found a terrrrrible issue that somehow went unnoticed for two weeks! This cute commit: http://git.savannah.gnu.org/cgit/guix.git/commit/?id=3Dd38bc9a9f6feefc465= 964531520fee5663a12f48 had the effect of disabling caching in the DAG traversal performed by =E2=80=98graft-derivation=E2=80=99. For big DAGs like that of kdevelop, yo= u could really feel it. :-) This is fixed by this one-liner: http://git.savannah.gnu.org/cgit/guix.git/commit/?id=3D90ad5c8836138b7fd4= d1bd0243dfa8b30ae0cf21 This other one fixes a related issue that had always been there apparently: http://git.savannah.gnu.org/cgit/guix.git/commit/?id=3D0aeed5e310504a9ef2= cf6a2b2a7e76086eb8c2fc Thanks for the heads-up! Ludo=E2=80=99.