On Thu, Nov 06, 2014 at 03:01:51PM +0000, Garth N. Wells wrote: > On Thu, 6 Nov, 2014 at 10:42 AM, Johannes Ring <[email protected]> > wrote: > >On Thu, Nov 6, 2014 at 8:13 AM, Anders Logg <[email protected]> wrote: > >> Hi, > >> > >> One thing I think is missing on our web page is a collection of good > >> installation recipes and, possibly, other FEniCS-related recipes. > >> > >> We do have good binary packages, but instructions for building from > >> source could be better. Here are some issues I think we need to > >>solve: > >> > >> 1. How do we want to continue to support Dorsal? Is anyone actively > >> maintaining it? Should it be the officially supported way to install > >> FEniCS from source on all platforms, including Mac? > > > >No, Dorsal is not actively maintained. Many package files and platform > >files are not particular up to date. > > Agree, and I don't think that supporting Dorsal is feasible. The > variety and variability of platforms is too great.
Agree. So what do we do? Retire Dorsal entirely? > >I have been working for a while on HashDist [1] and I have > >successfully used it to install FEniCS on Debian, Ubuntu, RHEL6/7, > >Mac, Windows (Cygwin) and on two different clusters (Abel and SciNet). > >HashDist is better than Dorsal in many ways: it has caching of builds, > >it handles dependencies between packages, it is reproducible, it is > >customizable, it will get support for binary builds, and it is > >actively maintained/developed. Currently, HashDist does not detect > >your platform like Dorsal does, or create a conf file that can be > >sourced to set environment variables, so it might not be as easy as > >Dorsal for FEniCS newcomers. However, this will likely improve in the > >future. > > > >It would be great if someone else than me could try to build FEniCS > >using HashDist. Please don't hesitate to send comments, suggestions or > >report issues. For those interested, there is also an introduction to > >HashDist at [2]. > > Nice - I look forward to trying it out. > > > > >[1] https://hashdist.github.io/ > >[2] http://www.youtube.com/watch?v=wviHkzk0AkY > > > >Johannes > > > >> 2. The installation instructions on the web page are very static and > >> difficult to keep up-to-date, since it requires making a pull > >>request > >> for the website Sphinx code, getting someone to apply the pull > >>request > >> etc. > >> > >> 3. Is there a good alternative in the form of a collection of > >>"FEniCS > >> recipes", to which anyone can contribute (perhaps wiki style), > >>ideally > >> also in combination with some voting mechanism (thumbs up = > >>works for > >> me) so that we may organically keep track of, say, the best way of > >> the day to build FEniCS on OS X. > >> > >> Ideally, we would host this on fenicsproject.org, but perhaps there > >> are already existing web services that could be used? > >> > >> 4. Could those installation recipes be made scriptable? And then how > >> would this be different from Dorsal? I like Dorsal, but it seems not > >> everyone is on board with using Dorsal as the main/official way of > >> FEniCS installation. > > > I suggest that we support 'official' scripted installations for the > latest version of Ubuntu. There can be (i) an apt-get version, (ii) > a stable version build, and (iii) a dev version. I suggest these > three because this is what we'll support in the Docker/Vagrant > approach (https://bitbucket.org/garth-wells/fenics-virtual). Via > Docker/Vagrant, the scripts can be tested by anyone independently of > their host OS. Agree, but to this I think we should add official build instructions for the latest OS X. Then there's the issue of scripted installations vs explained instructions. I think we need both and they need to match up. Perhaps it would be possible to reuse the demo documentation system where in place of the usual Python demo we have an installation bash script, and in place of the usual demo documentation (documentation.rst) we have a page explaining the steps in the installation. Then we could test nightly that the two match as we do for the demos, and we could make the installation instructions part of the DOLFIN source tree. How about that? > An advantage of a limited number of 'official' scripts is that they > can be a template (to read) for those installing on other platforms. > Simple scripted installation for HPC systems is not viable because > users don't have root access, and the machines are too varied and > are usually very different beasts from the nominal installed > distribution. > > For OSX, I'd like to see FEniCS in Homebrew Science > (https://github.com/Homebrew/homebrew-science). A downside of this > is that some of the Homebrew packages on which DOLFIN depends are > rather limited, e.g. Homebrew PETSc isn't configured with a decent > LU solver. > > More generally, we could ease the burden of installing Python > packages by having good pip support. I have never used pip (other than to install packages myself). What is needed to get all the FEniCS packages installable by pip? And how does pip differentiate between latest stable and latest development version? -- Anders _______________________________________________ fenics mailing list [email protected] http://fenicsproject.org/mailman/listinfo/fenics
