From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ludovic =?UTF-8?Q?Court=C3=A8s?= Subject: bug#40522: =?UTF-8?Q?=E2=80=98guile3.0-guix=E2=80=99?= sometimes fails to build on i686 (=?UTF-8?Q?=E2=80=9CToo?= many heap =?UTF-8?Q?sections=E2=80=9D?=) Date: Thu, 09 Apr 2020 17:33:02 +0200 Message-ID: <87v9m81xox.fsf@gnu.org> References: <877dyo50ht.fsf@inria.fr> <20200409152928.681fd892@scratchpost.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]:51102) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jMZBq-000671-Vr for bug-guix@gnu.org; Thu, 09 Apr 2020 11:34:03 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1jMZBp-0006gU-SI for bug-guix@gnu.org; Thu, 09 Apr 2020 11:34:02 -0400 Received: from debbugs.gnu.org ([209.51.188.43]:42805) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1jMZBp-0006gD-P6 for bug-guix@gnu.org; Thu, 09 Apr 2020 11:34:01 -0400 Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1jMZBp-00047Z-Kq for bug-guix@gnu.org; Thu, 09 Apr 2020 11:34:01 -0400 Sender: "Debbugs-submit" Resent-Message-ID: In-Reply-To: <20200409152928.681fd892@scratchpost.org> (Danny Milosavljevic's message of "Thu, 9 Apr 2020 15:29:28 +0200") 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-mx.org@gnu.org Sender: "bug-Guix" To: Danny Milosavljevic Cc: 40522@debbugs.gnu.org Hi, Danny Milosavljevic skribis: > https://github.com/Macaulay2/M2/issues/500 says there it's the Boehm GC b= eing > configured for small memory by default. Hmm, weird. =E2=80=98core-updates=E2=80=99 is on libgc 8.x (instead of 7.x= ), so we=E2=80=99ll have to check if the same happens. > They say setting the environment variable GC_INITIAL_HEAP_SIZE (to > like "50G") helps. Not sure why. > > Also, there's an environment variable GC_MAXIMUM_HEAP_SIZE. Looks like an ugly workaround, at best. :-) Thanks for checking! Ludo=E2=80=99.