New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Advertise candi as default installer? #1818
Comments
Yes! |
In general: yes. The more people use candi, the more robust it will become. Manual installation of packages is just too tedious. If you specify the packages manually (as you do above), do you still get the packages that the platform requires? Regarding an Would it be that much harder to say: "now get/configure/build aspect by
"? |
You are right the suggestion above overwrites config files. Lets postpone that idea then, and simply use candi to install deal.II. I have rewritten the install instructions in #1821, and will close this issue for now, as it seems we generally agree on using candi. |
Our local installation section in the manual is out of date (e.g. it recommends deal.II 8.4, which is not longer supported), and while I am willing to rewrite that section I thought it is also time to discuss if we should advertise candi (https://github.com/dealii/candi) as the new default installer.
As far as I know @tjhei , @egpuckett and @hlokavarapu use candi already for different purposes, and I have started to use it as well (also for cluster installations). While it requires one more configuration step (and therefore one more level of possible problems) it avoids a lot of the usual installation problems like inconsistent compilers between different packages. Additionally, it would allow us to ship different 'platform' files with ASPECT that contain configurations that are known to work on some clusters (e.g. stampede).
I am imagining that we could use candi to make a full ASPECT installation for most usual desktop systems as simple as:
Is this something worth pursuing? I have spent far too much time helping with installation problems, and it would be nice to at least only solve these problems once, incorporate the improvements in candi, and then never think about them again.
The text was updated successfully, but these errors were encountered: