From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:470:142:3::10]:40700) by lists.gnu.org with esmtp (Exim 4.86_2) (envelope-from ) id 1himjC-0003vB-T2 for gwl-devel@gnu.org; Wed, 03 Jul 2019 17:23:49 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1himjB-00010M-LS for gwl-devel@gnu.org; Wed, 03 Jul 2019 17:23:46 -0400 Received: from pelzflorian.de ([5.45.111.108]:42222 helo=mail.pelzflorian.de) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1himjA-0000z9-Uh for gwl-devel@gnu.org; Wed, 03 Jul 2019 17:23:45 -0400 Date: Wed, 3 Jul 2019 23:23:42 +0200 From: "pelzflorian (Florian Pelz)" Message-ID: <20190703212342.lwa5cq6oepr2nf53@pelzflorian.localdomain> References: <20190703132504.4vxrdbscwwz3vpnj@pelzflorian.localdomain> <87sgrntbjh.fsf@elephly.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <87sgrntbjh.fsf@elephly.net> Content-Transfer-Encoding: quoted-printable Subject: Re: [Bug?] Cannot use Guix Workflow Language on Guix System List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: gwl-devel-bounces+kyle=kyleam.com@gnu.org Sender: "gwl-devel" To: Ricardo Wurmus , zimoun Cc: gwl-devel@gnu.org Hi Ricardo & Simon, On Wed, Jul 03, 2019 at 04:53:06PM +0200, Ricardo Wurmus wrote: > Oh, it crashed=E2=80=A6 I fixed the bug causing the crash and the site = is up > again. >=20 The website looks nicer than before, thank you! (This Wisp style of Scheme seems easier to use with text editors that do not support indenting parentheses, but I would need to learn Wisp before I can fully understand it. It is probably easy like Scheme proper. I see the parenthetical Scheme examples still exist in the Git repo.) > On Wed, 3 Jul 2019 at 15:48, pelzflorian (Florian Pelz) > wrote: > > On Debian I had written a Guix workflow to deploy a not-yet-public > > Haunt website to a shared hosting provider and tried to use it on Gui= x > > now. (I hope that is a legitimate use case; I used gwl because I can > > specify package inputs.) >=20 > Interesting. While the primary audience is HPC users with scientific > workflows where parallel execution of processes on a cluster is > desirable, I guess building a website automatically with GWL processes > would work fine. Personally, I=E2=80=99d probably directly use a Guix = build > script because this workflow does not seem to benefit much from paralle= l > execution. >=20 I see. I will rewrite it then using: (run-with-store (open-connection) (gexp->derivation =E2=80=A6)) because it currently seems easier to explain and simpler for my limited needs. Maybe you could explain on the website and in the Guix package description which uses gwl is good for and why. I am not really familiar with scientific workflows and what they are used for in practice. > [=E2=80=A6] > FWIW =E2=80=9Cguix process=E2=80=9D no longer exists in the latest vers= ion, and > =E2=80=9CGUIX_WORKFLOW_PATH=E2=80=9D has been removed =E2=80=94 workflo= ws are now executed > directly from files and are no longer arranged in Guile modules. >=20 So my old file cannot work. On Wed, Jul 03, 2019 at 05:00:38PM +0200, zimoun wrote: > Myself, I `git clone` the GWL repo then enter in `guix environment` > with some --ad-hoc packages. Then from there I use ./pre-env-inst. > [=E2=80=A6] It works (with the example workflow)! (I omitted aclocal, autoconf, automake.) > > On Debian I had written a Guix workflow to deploy a not-yet-public > > Haunt website to a shared hosting provider and tried to use it on Gui= x > > now. (I hope that is a legitimate use case; I used gwl because I can > > specify package inputs.) >=20 > Interresting use case. > Could you elaborate? Or let me know when you publicly release your work= flow. >=20 I attach it. It is not the most beautiful code, but maybe you get the idea. I am using ssh to set up a git repository on the remote server and add a git hook to rebuild the website with each push like kind of a poor man=E2=80=99s continuous integration. Regards, Florian