[Dune-devel] First impressions with new website infrastructure

[Christian Engwer]

Hi Steffen,

> > * All images and PDFs are in LFS and not in the common Git
> >  repo. I ended up downloading the whole .5 GiB LFS stuff.
> >  That will improve once I learned Git LFS. I am not sure
> >  whether I want to have small binary files in Git and only
> >  large stuff in LFS. Let's see how this develops.
> 
> I very much want it that way because otherwise the repository will just keep growing (remember, if you
> commit a file and later delete it, it will still be contained in your cloned repo). The correct way of avoiding
> the auto-download of all the big files is to configure Git LFS (see git lfs help fetch). We should add the
> relevant instructions to the Howto suggested by you, but basically it goes like this:
> 
> git clone --config lfs.fetchexclude="static/download,static/pdf" https://gitlab.dune-project.org/infrastructure/dune-website.git
> 
> This will skip the download and PDF directories. If you later decide that you do not the directories after all,
> just do
> 
> git config lfs.fetchexclude "“ # or add whatever you still want to ignore in between the quotes
> git pull # will fetch the additional objects

I second Christophs comment.

I think git-lfs should really be the replacement for our previous
"large-file-storage", i.e. the webdav drive. The images are small and
change very seldom. I don't see the big issue having them in the git
repository... particularly if you use a hook to reject large files.

Christian