Hi all,
The world keeps on turning, and maybe it is time to revisit mimimum version
requirements for Flow dependencies. One of the more pressing issues is Dune. We
now have an implementation of CPR (constrained pressure residual) which
requires Dune to be version 2.6.0 or newer. Today, this mea
Another recurring issue is Boost. If we jump to Debian Buster and Ubuntu
> 19.04, this entails that we can bump minimum version of GCC to 8.3.0. Is
> this sufficient to consider removing Boost as a dependency?
>
The parser uses boost spirit::qi which at some stage proved to give a
performance gain
In addition to the Boost.Spirit component that Joakim mentions there's also the
practical issue that almost all of our unit tests target Boost.Test as a
framework. For instance, the header is included 193
times across the OPM code base. If we're going to replace that with another
framework--
Hi!
I support increasing the Dune dependency to 2.6.
With a minimum GCC version of 8.3 we can adopt C++17 and drop several boost
libraries, but not yet all.
Atgeirr
Frå: Opm på vegne av Alf Birger Rustad
Sendt: tysdag 28. januar 2020 09:13
Til: opm@opm-projec
Personally I don't think Boost.Test is worth replacing, but getting rid of
filesystem and regex is certainly worthwhile.
As for dune, the ppa provides 2.6 for 18.04 and 16.04 already so it does not
have a big impact for ubuntu builds.
Fra: Opm på vegne av Bård S
Despite what I initially wrote (that I want to increase the requirements), if
we turn away potential users by this, that would be a good reason to NOT bump
the requirements.
Arne Morten's message implies that we can still provide binary packages for new
releases (i.e. 2020.04 and 2020.10 etc.)
On Tue, Jan 28, 2020 at 09:36:47AM +0100, Joakim Hove wrote:
> Another recurring issue is Boost. If we jump to Debian Buster and Ubuntu
> > 19.04, this entails that we can bump minimum version of GCC to 8.3.0. Is
> > this sufficient to consider removing Boost as a dependency?
> >
>
> The parser us
Good point, but do we really want to support three generations of Ubuntu LTS?
Personally, I think that is too much.
Cheers,
Alf
-Original Message-
From: Opm On Behalf Of Atgeirr Rasmussen
Sent: tirsdag 28. januar 2020 10:00
To: opm@opm-project.org
Subject: Re: [Opm] Dependencies revisit
Hi,
On Tue, Jan 28, 2020 at 09:00:26AM +, Atgeirr Rasmussen wrote:
> Despite what I initially wrote (that I want to increase the requirements), if
> we turn away potential users by this, that would be a good reason to NOT bump
> the requirements.
>
> Arne Morten's message implies that we ca
hi,
yes, following my 'normal' schedule I drop after 4 years. Thus, 16.04 would not
be published for opm release 20.04, only 18.04 and 20.04.
My point was that even if people would want to stick to 16.04, we would still
provide Dune packages for them.
Fra: Opm
___
Opm mailing list
Opm@opm-project.org
https://opm-project.org/cgi-bin/mailman/listinfo/opm
My only problem with the new requirements is that I manage many Ubuntu
servers, some for clients, some for myself. I will only upgrade to LTS
versions of Ubuntu, ever.
That means I'm on Ubuntu 18.04 LTS right now, and won't be upgrading
until the next LTS release, whichever that is (I think 20
12 matches
Mail list logo